Datasets:

OrganizedProgrammers
/

3GPPSpecContent

Dataset card Data Studio Files Files and versions Community

Split (1)

train · 466k rows

hash stringlengths 32 32	doc_id stringlengths 5 12	section stringlengths 4 1.47k	content stringlengths 0 6.67M
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.1.5.1 General	Based on roaming agreements (e.g. SLA), analytics may be exchanged between PLMNs (i.e. HPLMN and VPLMN of a UE served by the NWDAF analytics consumer). In this case, an NWDAF with roaming exchange capability (RE-NWDAF) is used as entry point in a PLMN to exchange analytics in roaming scenario with other PLMNs. It authorizes the analytics request according to roaming agreements, and filters the information exposed to other PLMNs. The roaming architecture is defined in clause 4.3. The H-RE-NWDAF may provide analytics to the V-RE-NWDAF. The V-RE-NWDAF may provide analytics to the H-RE-NWDAF. NOTE 1: The roaming agreements (e.g. SLA) between HPLMN and VPLMN on network analytics and data exchange takes into consideration all relevant regulatory requirements. NOTE 2: It depends on the deployment whether analytics exchange between PLMNs is supported, and thus whether H-RE-NWDAF and V-RE-NWDAF are provided in a PLMN. The H-RE-NWDAF or V-RE-NWDAF provides the Nnwdaf_RoamingAnalytics service for that purpose. An RE-NWDAF is the only consumer of these services, i.e. both NWDAF in HPLMN and NWDAF in VPLMN need to have the roaming exchange capability (in other words, be an H-RE-NWDAF or V-RE-NWDAF, respectively) when used as entry point or exit point to exchange analytics in roaming scenario. NOTE 3: The access to the Nnwdaf_AnalyticsSubscription service and the Nnwdaf_AnalyticsInfo service is expected to be restricted by the NRF to NF service consumers within the same PLMN to prevent bypassing checks based on user consent and operator policy NOTE 4: See clause X.7 and Annex V of TS 33.501 [49] for details of the user consent check procedures. See clause X.8 of TS 33.501 [49] for protection of analytics exchange in roaming case. V-RE-NWDAF may request or subscribe to HPLMN analytics from the H-RE-NWDAF as described in clause 6.1.5.2, and then the analytics can be leveraged by the 5GC NF in the VPLMN, for example: - In home routed roaming scenarios, HPLMN analytics (i.e. slice load level analytics, NF load analytics, etc.) can be leveraged by the AMF in the VPLMN for network slice selection and SMF selection for PDU Session management. - UE-related analytics (e.g.. service experience analytics, etc.) can include statistics or predictions for outbound roaming UEs. NOTE 5: Analytics that rely on input data from the VPLMN are preferably not provided from H-RE-NWDAF to V-RE-NWDAF, but generated by a NWDAF in the VPLMN. H-RE-NWDAF may request or subscribe to VPLMN analytics from the V-RE-NWDAF as described in clause 6.1.5.3, and then the analytics can be leveraged by the 5GC NF in the HPLMN, for example: - In home routed roaming scenarios, analytics information with statistics or predictions for outbound roaming UEs can be leveraged by the H-PCF for QoS control of the PDU Session. - Analytics (i.e. service experience analytics, slice load level analytics, etc.) can be leveraged by the H-PCF for decision of NSSP in URSP rules provisioned to the UE roaming in the VPLMN. NOTE 6: Analytics that rely on input data from the HPLMN are preferably not provided from V-RE-NWDAF to H-RE-NWDAF, but generated by a NWDAF in the HPLMN.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.1.5.2 Analytics Exposure from HPLMN to VPLMN	Figure 6.1.5.2-1 shows the procedure where a V-RE-NWDAF requests network analytics (i.e. slice load level analytics, NF load analytics, service experience analytics, etc.) in HPLMN from a H-RE-NWDAF, upon receiving an analytics information request/subscribe from a service consumer NF (e.g. AMF) in the VPLMN. Figure 6.1.5.2-1: Procedure for analytics exposure from HPLMN to VPLMN 1a. The Consumer NF in VPLMN (e.g. AMF) discovers a V-RE-NWDAF as described in clause 5.2 and sends an Analytics request/subscribe (Analytics ID, Analytics Filter Information, Target of Analytics Reporting) to the V-RE-NWDAF by invoking a Nnwdaf_AnalyticsInfo_Request service operation or a Nnwdaf_AnalyticsSubscription_Subscribe service operation. 1b. For the inbound roaming UE(s) indicated in the Target of Analytics Reporting: - The V-RE-NWDAF determines based on operator configuration and the requested analytics whether analytics or input data from the HPLMN are required, or the analytics can be derived locally. The subsequent steps only apply if analytics from the HPLMN are required. If input data from the HPLMN are required, the procedures in clause 6.2.11 apply. NOTE 1: It is possible that the Target of Analytics Reporting sent by the Consumer NF to the V-RE-NWDAF includes both inbound roaming UE(s) and non-roaming UE(s). -. V-RE-NWDAF checks the roaming agreements related to analytics from the HPLMN to determine if the roaming analytics request/subscribe can be accepted or must be rejected with a proper cause in the response to the Consumer NF. If the V-RE-NWDAF determined the roaming analytics request/subscribe is rejected, the following steps are skipped. 2. The V-RE-NWDAF discovers a H-RE-NWDAF as described in clause 5.2 if the V-RE-NWDAF determines the roaming analytics request/subscribe can be accepted. The V-RE-NWDAF sends a roaming analytics request/subscribe (Analytics ID, Analytics Filter Information, Target of Analytics Reporting) to H-RE-NWDAF by invoking a Nnwdaf_RoamingAnalytics_Request service operation or a Nnwdaf_RoamingAnalytics_Subscribe service operation, based on the Analytics request/subscribe received from the Consumer NF in the VPLMN. The Target of Analytics Reporting sent by the V-RE-NWDAF to the H-RE-NWDAF only contains the inbound roaming UE(s). NOTE 2: The inbound roaming UE(s) are distinguished by the V-RE-NWDAF according to the UE ID(s) (i.e. SUPI(s)). 3a. The H-RE-NWDAF checks the roaming agreements between the HPLMN and the VPLMN, and user consent for analytics as defined in clause 6.2.9 if needed, to determine if the roaming analytics request/subscribe can be accepted or must be rejected with a proper cause in response to the V-RE-NWDAF (which then relays the response to the Consumer NF). If the roaming analytics request/subscribe is rejected, the following steps are skipped. If the H-RE-NWDAF supports to generate the requested analytics, it collects data from the NF(s) and/or OAM in the HPLMN and derives the requested analytics; otherwise steps 3b and 3c are executed. NOTE 3: See clause X.7 and Annex V of TS 33.501 [49] for details of the user consent check procedures. See clause X.8 of TS 33.501 [49] for protection of analytics exchange in roaming case. 3b-3c. [Optional] If the H-RE-NWDAF does not support to generate the requested analytics, it may request/subscribe to other NWDAF(s) in the HPLMN (if available) for the analytics and get corresponding response/notification. 4. The H-RE-NWDAF sends the HPLMN analytics information to the V-RE-NWDAF using either Nnwdaf_RoamingAnalytics_Request response or Nnwdaf_RoamingAnalytics_Notify service operation, depending on the service used in step 2. The H-RE-NWDAF may restrict the exposed analytics information based on HPLMN operator polices. 5. [Optional] If the Consumer NF also indicates request or subscription of analytics information available in the VPLMN (e.g. via Target of Analytics Reporting) in step 1, the V-RE-NWDAF collects data from the NF(s) and/or OAM in VPLMN and derives the requested analytics. These steps can be executed in parallel with steps 3-4. The V-RE-NWDAF may perform analytics aggregation with the analytics information received from the H-RE-NWDAF and analytics information generated by itself, based on the analytics request or subscription. 6. The V-RE-NWDAF sends the HPLMN analytics information received in step 4, or the aggregated analytics information if step 5 are performed, to the Consumer NF in the VPLMN using either Nnwdaf_AnalyticsInfo_Request response or Nnwdaf_AnalyticsSubscription_Notify service operation, depending on the service used in step 1. NOTE 4: The present document describes that the RE-NWDAF can perform analytics aggregation for roaming scenario, but whether and how the RE-NWDAF performs analytics aggregation for roaming scenario are up to implementation.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.1.5.3 Analytics Exposure from VPLMN to HPLMN	Figure 6.1.5.3-1 shows the procedure that the H-RE-NWDAF requests network analytics (i.e. service experience analytics, slice load level analytics, etc.) in the VPLMN from the V-RE-NWDAF, upon receiving an analytics information request/subscription from the service consumer NF (e.g. PCF) in HPLMN. Figure 6.1.5.3-1: Procedure for analytics exposure from VPLMN to HPLMN 1a. If the Consumer NF in the HPLMN (e.g. H-PCF) is aware that the UE(s) indicated in Target of Analytics Reporting is/are outbound roaming UE(s), the Consumer NF discovers a H-RE-NWDAF as described in clause 5.2 and sends an Analytics request/subscribe (Analytics ID, Target of Analytics Reporting, Analytics Filter Information) to the H-RE-NWDAF by invoking a Nnwdaf_AnalyticsInfo_Request service operation or a Nnwdaf_AnalyticsSubscription_Subscribe service operation. 1b. For the outbound roaming UE(s) indicated in the Target of Analytics Reporting: - The H-RE-NWDAF determines based on operator configuration and the requested analytics whether analytics or input data from the VPLMN are required, or the analytics can be derived locally. The subsequent steps only apply if analytics from the VPLMN are required. If input data from the VPLMN are required, the procedures in clause 6.2.10 apply. - If the Consumer NF is unaware that the Target of Analytics Reporting includes roaming UE(s) and sends the Analytics request/subscribe to a H-NWDAF which does not support roaming exchange capability, then the H-NWDAF may perform either of the following after determining that the Target of Analytics Reporting includes roaming UE(s) (e.g. by inquiring the AMF(s) serving the UE(s) at the UDM with the Nudm_UECM service): - forward the Analytics request/subscribe to a H-RE-NWDAF after discovering a H-RE-NWDAF as described in clause 5.2. The H-NWDAF may include the VPLMN ID in the Analytics request/subscribe; or - reject the Analytics request/subscribe with a proper cause value in the response to the Consumer NF. And the following steps will not be performed. NOTE 1: It is possible that the Target of Analytics Reporting sent by the Consumer NF to the H-RE-NWDAF includes both outbound roaming UE(s) and non-roaming UE(s). NOTE 2: The H-NWDAF is not depicted in the flow. - If PLMN ID of the VPLMN is not included in the Analytics request/subscribe, the H-NWDAF inquires it at the UDM for the UE(s) indicated as Target of Analytics Reporting. The H-RE-NWDAF checks user consent for analytics as defined in clause 6.2.9. The H-RE-NWDAF checks the roaming agreements between the HPLMN and the VPLMN to determine if the roaming analytics request/subscription can be accepted or must be rejected with a proper cause in the response to the Consumer NF. If the H-RE-NWDAF determined the roaming analytics request/subscribe is rejected, the following steps are skipped. NOTE 3: See clause X.7 and Annex V of TS 33.501 [49] for details of the user consent check procedures. See clause X.8 of TS 33.501 [49] for protection of analytics exchange in roaming case. 2. The H-RE-NWDAF discovers the V-RE-NWDAF as described in clause 5.2 if the H-RE-NWDAF determined the roaming analytics request/subscribe can be accepted. The H-RE-NWDAF sends a roaming analytics request/subscribe (Analytics ID, Analytics Filter Information, Target of Analytics Reporting, [NF ID(s)]) to the V-RE-NWDAF by invoking a Nnwdaf_RoamingAnalytics_Request service operation or a Nnwdaf_RoamingAnalytics_Subscribe service operation, based on the Analytics request/subscribe received from the Consumer NF in HPLMN. The Target of Analytics Reporting sent by the H-RE-NWDAF to the V-RE-NWDAF only contains the outbound roaming user(s). The H-RE-NWDAF may obtain NF ID(s) of the NF(s) serving the roaming UE(s) in the VPLMN, e.g. AMF ID(s), SMF ID(s), from the UDM and include the NF ID(s) in the analytics request/subscribe. 3a. The V-RE-NWDAF checks the roaming agreements between the HPLMN and the VPLMN, to determine if the roaming analytics request/subscribe can be accepted or must be rejected with a proper cause in response to the H-RE-NWDAF (which then relays the response to the Consumer NF). If the roaming analytics request/subscribe is rejected, the following steps are skipped. If the V-RE-NWDAF supports to generate the requested analytics, it collects data from the NF(s) serving the roaming UE(s) and/or OAM in VPLMN and derives the analytics; otherwise step 3b and step 3c are executed. The NF(s) serving the roaming UE(s), e.g. AMF(s) or SMF(s), if indicated in step 2, can be used as data source. 3b-3c. [Optional] If the V-RE-NWDAF does not support to generate the requested analytics, it may request/subscribe to other NWDAF(s) in the VPLMN (if available) for the analytics and get corresponding response/notification(s). 4. The V-RE-NWDAF sends the VPLMN analytics information to the H-RE-NWDAF using either Nnwdaf_RoamingAnalytics_Request response or Nnwdaf_RoamingAnalytics_Notify service operation, depending on the service used in step 2. The V-RE-NWDAF may restrict the exposed analytics information based on VPLMN operator polices. 5. [Optional] If the Consumer NF also indicates request or subscription of analytics information available in the HPLMN (e.g. via Target of Analytics Reporting) in step 1, the H-RE-NWDAF collects data from the NF(s) and/or OAM in HPLMN and derives the analytics. These steps can be executed in parallel with steps 3-4. The H-RE-NWDAF may perform analytics aggregation with the analytics information received from the V-RE-NWDAF and analytics information generated by itself, based on the analytics request or subscription. 6. The H-RE-NWDAF sends the VPLMN analytics information received in step 4, or the aggregated analytics information if step 5 are performed, to the Consumer NF in HPLMN using either Nnwdaf_AnalyticsInfo_Request response or Nnwdaf_AnalyticsSubscription_Notify, depending on the service used in step 1. If the Analytics request/subscribe is forwarded to the H-RE-NWDAF by a H-NWDAF as described in step 1, the H-NWDAF forwards the analytics information received from the H-RE-NWDAF to the Consumer NF using either Nnwdaf_AnalyticsInfo_Request response or Nnwdaf_AnalyticsSubscription_Notify, depending on the service used in step 1. NOTE 4: The present document describes that the RE-NWDAF may perform analytics aggregation for roaming scenario, but whether and how the RE-NWDAF performs analytics aggregation for roaming scenario are up to implementation.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.1.5.4 Contents of Analytics Exposure in roaming case	When requesting or subscribing to analytics involving one or more roaming UEs, the Consumer NF shall send the request or subscription to an RE-NWDAF belonging to the same PLMN as the Consumer NF. The Consumer NF may indicate the following parameters in the Nnwdaf_AnalyticsInfo_Request service operation or the Nnwdaf_AnalyticsSubscription_Subscribe service operation: - The parameters as defined in clause 6.1.3, with the following differences: - Parameters related to analytics aggregation, analytics transfer or ML Model selection should not be included; - If Target of Analytics Reporting is included and indicates specific UE(s) or a group of UEs, these UEs shall belong to the same HPLMN (if HPLMN analytics are subscribed/requested) or be served by the same VPLMN (if VPLMN analytics are subscribed/requested) - Additionally, the following parameters: - If the NWDAF service consumer in a VPLMN requests/subscribes to HPLMN analytics, in Analytics Filter Information: - [OPTIONAL] PLMN ID of the HPLMN; NOTE 1: If PLMN ID of the HPLMN is not provided by the NWDAF service consumer, the V-RE-NWDAF derives it from the SUPIs of UEs indicated as Target of Analytics Reporting. - [OPTIONAL] mapped S-NSSAI of the HPLMN. - [OPTIONAL] AOI (in the form of geographical area) in the HPLMN - If the NWDAF service consumer in a HPLMN requests/subscribes to VPLMN analytics, in Analytics Filter Information: - [OPTIONAL] PLMN ID of the VPLMN - [OPTIONAL] AOI (in the form of geographical area) in the VPLMN NOTE 2: If PLMN ID of the VPLMN is not provided by the NWDAF service consumer, the H-RE-NWDAF inquires it at the UDM for the UEs indicated as Target of Analytics Reporting. NOTE 3: In this release of the specification, only one PLMN ID (of HPLMN or VPLMN) is supported in the Nnwdaf_AnalyticsInfo_Request service operation or the Nnwdaf_AnalyticsSubscription_Subscribe service operation. Based on the analytics request or subscription from the NWDAF service consumer in the HPLMN, and local configuration, the H-RE-NWDAF may indicate the following parameters in the Nnwdaf_RoamingAnalytics_Request service operation or the Nnwdaf_RoamingAnalytics_Subscribe service operation to the V-RE-NWDAF, for requesting/subscribing to analytics in the VPLMN: - Analytics ID; - PLMN ID of the HPLMN; NOTE 4: Security aspects for analytics exchange are covered in TS 33.501 [49]. - Analytics Filter Information: - [OPTIONAL] HPLMN S-NSSAI; NOTE 5: V-NWDAF maps that S-NSSAI to an S-NSSAI of the VPLMN which will be used in the Analytics Filter Information. - other Analytics Filter Information (e.g. AOI in the form of geographical area in the VPLMN) as provided by NWDAF service consumer in the HPLMN, if applicable in the VPLMN. - [OPTIONAL] NF ID(s) of the NF(s) (e.g. AMF(s), SMF(s)) serving the roaming UE(s) in the VPLMN; - other parameters as provided by NWDAF service consumer in the HPLMN, if applicable in the VPLMN. Based on the analytics request or subscription from the NWDAF service consumer in the VPLMN, and local configuration, the V-RE-NWDAF may indicate the following parameters in the Nnwdaf_RoamingAnalytics_Request service operation or the Nnwdaf_RoamingAnalytics_Subscribe service operation to the H-RE-NWDAF, for requesting/subscribing to analytics in the HPLMN: - Analytics ID; - PLMN ID of the VPLMN; NOTE 6: Security aspects for analytics exchange are covered in TS 33.501 [49]. - Analytics Filter Information: [OPTIONAL] HPLMN S-NSSAI; NOTE 7: If an S-NSSAI but no mapped S-NSSAI is provided by the NWDAF service consumer in the Analytics Filter Information, the V-NWDAF maps the S-NSSAI to the S-NSSAI of the HPLMN and provides that mapped S-NSSAI of the HPLM as the Analytics Filter Information. - other Analytics Filter Information (e.g. AOI in the form of geographical area in the HPLMN) as provided by NWDAF service consumer in the VPLMN, if applicable in the HPLMN. - other parameters as provided by NWDAF service consumer in the VPLMN, if applicable in the HPLMN. The RE-NWDAF provides the following output information to the consumer RE-NWDAF of the Nnwdaf_RoamingAnalytics_Request service operation or the Nnwdaf_RoamingAnalytics_Notify service operation: - The output information as defined in clause 6.1.3, with the following differences: - Information related to analytics aggregation (i.e. analytics metadata information) should not be included. NOTE 8: The output of the analytics depends on the roaming agreements. NOTE 9: UE specific data and analytics exchange between HPLMN and VPLMN and the possible storage is agreed between the operators bilaterally via roaming agreements (e.g. SLA) and takes into consideration all relevant regulatory requirements. 6.1A Analytics aggregation from multiple NWDAFs 6.1A.1 General In a multiple NWDAF deployment scenario, an NWDAF instance may be specialized to provide Analytics for one or more Analytics IDs. Each of the NWDAF instances may serve a certain Area of Interest or TAI(s). Multiple NWDAFs may collectively serve the particular Analytics ID. An NWDAF may have the capability to support the aggregation of Analytics (per Analytics ID) received from other NWDAFs, possibly with Analytics generated by itself. The procedure for analytics aggregation from multiple NWDAFs is as defined in clause 6.1A.3. 6.1A.2 Analytics Aggregation The analytics aggregation from multiple NWDAFs is used to address cases where an NWDAF service consumer requests Analytics ID(s) that requires multiple NWDAFs to collectively serve the request. Analytic aggregation applies to scenarios where NWDAF service consumer requests or subscribes to analytics information with or without provisioning Area of Interest. Aggregator NWDAF or aggregation point: - Is an NWDAF instance with additional capabilities to aggregate output analytics provided by other NWDAFs. This is in addition to regular NWDAF behaviour such as collecting data from other data sources to be able to generate its own output analytics. - Is able to divide the area of interest, if received from the consumer, into sub area of interest based on the serving area of each NWDAF to be requested for analytics and then send analytics requests including the sub area of interest as an Analytics Filter to corresponding NWDAFs. The Aggregator NWDAF may maintain information on the discovered NWDAFs, including their supported Analytics IDs, NWDAF Serving Areas, etc. - Has "analytics aggregation capability" registered in its NF Profile within the NRF. - Supports the requesting and exchange of "Analytics Metadata Information" between NWDAFs when required for the aggregation of output analytics. "Analytics Metadata Information" is additional information associated with the requested Analytics ID(s) as defined in clause 6.1.3. - Supports dataset statistical properties, output strategy and data time window parameters per type of analytics (i.e. Analytics ID) as defined in clause 6.1.3. NRF: - Stores the NF Profile of the NWDAF instances, including "analytics aggregation capability" for Aggregator NWDAFs and "analytics metadata provisioning capability" when supported by the NWDAF. - Returns the NWDAF(s) matching the attributes provided in the Nnrf_NFDiscovery_Request, as specified in clause 5.2.7.3 of TS 23.502 [3]. NWDAF service consumer: - Requests or subscribes to receive analytics for one or more Analytics IDs, as specified in clause 6.1 of the present document. - Does discovery and selection as defined in clause 6.3.13 of TS 23.501 [2] to identify NWDAFs with analytics aggregation capability and other capabilities (e.g. providing data/analytics for specific TAI(s)). - Can differentiate and select the preferred NWDAF in case multiple NWDAFs are returned in the NWDAF discovery response based on its internal selection criteria (considering the registered NWDAF capabilities and information in NRF or UDM). 6.1A.3 Procedure for analytics aggregation 6.1A.3.1 Procedure for analytics aggregation with Provision of Area of Interest The procedure depicted in figure 6.1A.3-1 is used to address cases where an NWDAF service consumer requests Analytics ID(s) for an Area of Interest that requires multiple NWDAFs that collectively serve the request. Figure 6.1A.3.1-1: Procedure for analytics aggregation 1a-b. NWDAF service consumer discovers the NWDAF as specified in clause 5.2. When NRF is used, NRF may return multiple NWDAF candidates matching the requested capabilities, area of interest and supported Analytics ID(s). NWDAF service consumer selects an NWDAF (e.g. NWDAF1) with analytics aggregation capability (i.e. Aggregator NWDAF), based on its internal selection criteria, considering registered NWDAF capabilities and information in NRF including the Supported Analytics Delay per Analytics ID (if available). 2. NWDAF service consumer invokes Nnwdaf_AnalyticsInfo_Request or Nnwdaf_AnalyticsSubscription_Subscribe service operation from the selected Aggregator NWDAF (e.g. NWDAF1). In the request, NWDAF service consumer provides Analytics ID(s) (e.g. Analytics ID 1) Analytics Filter Information (area of interest, e.g. TAI-1, TAI-2, TAI-n, if known to the NWDAF service consumer), Target of Analytics Reporting as defined in clause 6.1.3. It may also provide "time when analytics information is needed" (e.g. T1). It is expected that T1 is equal or greater than the Supported Analytics Delay per Analytics ID (if available) of the Aggregator NWDAF. Otherwise, the aggregator NWDAF may reject the analytics request or analytics subscription. 3. On receiving the request in step 2, Aggregator NWDAF (e.g. NWDAF1), based on e.g. configuration, queries to NRF including the Real-Time Communication Indication per Analytics ID and queries to UDM for checking which NWDAF(s) is serving the Target of Analytics Reporting. Considering the request from the NWDAF service consumer (e.g. Analytics Filter Information, T1, etc.) and Supported Delay per Analytics ID per NWDAF instance (when Real-Time Communication Indication was included), Aggregator NWDAF determines the other NWDAF instances that collectively can cover the area of interest indicated in the request (e.g. TAI-1, TAI-2, TAI-n). NOTE 1: In the discovery request sent to NRF, Aggregator NWDAF might indicate "analytics metadata provisioning capability" (e.g. as query parameter), thus, requesting to NRF to reply back with, if available, those NWDAF instance(s) which also supports "analytics metadata provisioning capability" functionality as indicated during particular NWDAF instance registration procedure. 4-5. Aggregator NWDAF (e.g. NWDAF1) invokes Nnwdaf_AnalyticsInfo_Request or Nnwdaf_AnalyticsSubscription_Subscribe service operation from each of the NWDAFs discovered/determined in step 3 (e.g. NWDAF2 and NWDAF3). The request may optionally indicate "analytics metadata request" parameter to the determined NWDAFs (e.g. NWDAF-2 and/or NWDAF3), when analytics metadata is supported by these NWDAFs. The request or subscription to the determined NWDAFs (e.g. NWDAF2 and/or NWDAF3) may also include the dataset statistical properties, output strategy and data time window. This indicates to the determined NWDAFs that the Analytics ID output shall be generated based on such parameters when requested. If "time when analytics information is needed" (T1) was provided in step 2, the Aggregator NWDAF shall also provide a "time when analytics information is needed" to the determined NWDAFs, with a smaller value compared with T1 (e.g. T2). NOTE 2: T2 in step 4-5 is smaller than T1 accounting for the analytics delay and processing time within the Aggregator NWDAF itself. 6-7a-b. The determined NWDAFs (e.g. NWDAF-2 and/or NWDAF3) reply or notify with the requested output analytics. If "analytics metadata request" was included in the request received by such NWDAF (in steps 4-5), the NWDAF additionally returns the "analytics metadata information" used for generating the analytics output as defined in clause 6.1.3. If the determined NWDAFs (e.g. NWDAF 2 and/or NWDAF 3) cannot reply or notify with the requested output analytics before the expiry of T2, they may send an error response or error notification to the Aggregator NWDAF including a "revised waiting time". 8. Aggregator NWDAF (e.g. NWDAF1) aggregates received Analytics information, i.e. generates a single output analytics based on the multiple analytics outputs and, optionally, the "analytics metadata information" received from the determined NWDAFs (e.g. NWDAF2 and NWDAF3). The Aggregator NWDAF (e.g. NWDAF1) may also take its own analytics for TAI-n into account for the analytics aggregation. 9a-b. Aggregator NWDAF (e.g. NWDAF1) sends a response or notifies to the NWDAF service consumer the aggregated output analytics for the requested Analytics ID. If the Aggregator NWDAF (e.g. NWDAF 1) cannot reply or notify with the requested output analytics before the expiry of T1 or anticipates that it cannot reply or notify with the requested output analytics before the expiry of T1 (e.g. due to error notification in step 6-7a-b), it may send an error response or error notification to the NWDAF service consumer including a "revised waiting time". The NWDAF service consumer may optionally use the "revised waiting time" to update the "time when analytics information is needed" parameter (i.e. T1) for future analytics requests/subscriptions to the same Aggregator NWDAF as defined in clause 6.2.5.2. 6.1A.3.2 Procedure for Analytics Aggregation without Provision of Area of Interest The procedure depicted in Figure 6.1A.3.2-1 is used to address cases where an NWDAF service consumer requests Analytics ID(s) without providing an Area of Interest, but requires multiple NWDAFs to collectively serve the request. Figure 6.1A.3.2-1: Procedure for analytics aggregation without provision of Area of Interest 1. This step is a NWDAF discovery procedure without providing any area of interest. The service consumer discovers an aggregation NWDAF (e.g. NWDAF1) as specified in clause 5.2. When NRF is used, NRF may return multiple NWDAF candidates matching the requested capabilities and supported Analytics ID(s). Depending on the requested Analytics ID, the NWDAF service consumer, e.g. based on internal logic, can be able to determine which NWDAF should be selected for providing the required data analytics. If not, the NWDAF service consumer should select a NWDAF with large serving area from the candidate NWDAFs which supports analytics aggregation, e.g. NWDAF1. 2. NWDAF service consumer sends Analytics information or analytics subscription request to the aggregator NWDAF, i.e. NWDAF1 in the Figure 6.1A.3.2-1. In the request, NWDAF service consumer provides the requested Analytics ID(s), e.g. Analytics ID 1. The NWDAF service consumer may also provide "time when analytics information is needed" (e.g. T1). It is expected that T1 is equal or greater than the Supported Analytics Delay per Analytics ID of the Aggregator NWDAF (if available). Otherwise, the aggregator NWDAF may reject the analytics request or analytics subscription. Once receiving the request, the Aggregator NWDAF1 may decide to subscribe data analytics from other NWDAF instances which can provide the requested data analytics. Based on the Analytics ID, there are two cases for the Aggregator NWDAF1 to subscribe data analytics from other NWDAF instances. 3a. If the data analytics requires UE location information, e.g. for the Analytics IDs "UE Mobility", "Abnormal behaviour", or "User Data Congestion", then: - 3a-1: (optional) The Aggregator NWDAF1 queries UDM to discover the NWDAF serving the UE, if it is supported. - 3a-2: If step 3a-1 is not supported, was not executed, or did not return a suitable NWDAF serving the UE, the Aggregator NWDAF1 determines the AMF serving the UE as specified in the clause 6.2.2.1, then requests UE location information from the AMF to be used in the query to NRF in step 4. NOTE: If an Aggregator NWDAF receives an Analytics request for a group of UEs, i.e. the Target of Analytics Reporting set to a list of Internal Group IDs, it performs NWDAF discovery based on location information of all UEs in each of the Internal Group Ids in the list and then requests all discovered NWDAFs to provide the required analytics. 3b. If the data analytics does not require to collect UE location information, e.g. for the Analytics IDs "Service Experience", "NF load information", or "UE Communication", the Aggregator NWDAF1 can determine the NFs to be contacted for data collection as specified in clause 6.2.2.1 and then it can retrieve NF service area for each of the data source NFs from NRF. 4. (conditional) With the data obtained in step 3, the Aggregator NWDAF1 queries the NRF for discovering the required NWDAF, by sending an NF discovery request which may include UE location (e.g. TAI-1) or NF serving area (e.g. TAI list-1) as a filter to NRF and obtains candidates target NWDAF(s) that can provide the required analytics. This step is skipped if a suitable NWDAF was discovered in step 3a-1. Additionally the Aggregator NWDAF1 may include in the NF discovery request the Real-Time Communication Indication per Analytics ID to request Supported Delay per Analytics ID per NWDAF instance. Depending on the discovered NWDAF instance(s), there can be two cases: 5a. If a single target NWDAF (e.g. NWDAF2) can provide the requested analytics data, the Aggregator NWDAF (e.g. NWDAF1) can redirect the Nnwdaf_AnalyticsInfo_Request to that target NWDAF or request an analytics subscription transfer to that target NWDAF, depending on the type of the analytics request/subscription received by the NWDAF Service Consumer. 5b. If the Aggregator NWDAF decides to request data analytics from one or more target NWDAFs, the steps 4-9 of the analytics aggregation procedure in clause 6.1A.3.1 are executed. 6.1B Transfer of analytics context and analytics subscription 6.1B.1 General In a multiple NWDAFs deployment scenario, procedures for transfer of analytics context and analytics subscription can be used to support the target NWDAF to produce the needed analytics. When the analytics consumer provides the target NWDAF with information on the subscription that could be transferred from the source NWDAF, the target NWDAF may initiate the transfer of analytics context. The analytics consumer provides the information via Nnwdaf_AnalyticsSubscription_Subscribe service operation. When the analytics consumer is an AMF, the old subscription information (if related to a UE) may be provided by the source AMF to the target AMF using a UE context transfer procedure as specified in TS 23.502 [3]. An analytics subscription transfer to the target NWDAF may be initiated by the source NWDAF, followed by an analytics context transfer initiated by the target NWDAF. An NWDAF may transfer one or more of its analytics subscriptions to another NWDAF instance due to internal (e.g. load balancing, graceful shutdown) or external triggers (e.g. UE mobility). For external triggers, the NWDAF may subscribe to NF(s) to be notified about the corresponding events. As for UE mobility, upon the UE location change event notified by the AMF subscribed by the NWDAF, the NWDAF determines whether it can continue to provide the analytics service. If the NWDAF cannot continue to serve the consumer, it should either select a target NWDAF and initiate analytics subscription transfer, or notify the analytics consumer that it cannot provide the service anymore, so that the analytics consumer can select a new NWDAF. Procedures for analytics subscription transfer allow one NWDAF instance to transfer its ongoing analytics subscriptions to another NWDAF instance. The transfer can be done for all subscriptions or just a selected subset of subscriptions related to specific area(s), specific Analytics ID(s), specific NF(s) and/or specific UE(s). The procedure for prepared analytics subscription transfer can be used if the source NWDAF instance anticipates that it will soon not be able to continue its current analytics tasks. 6.1B.2 Analytics Transfer Procedures 6.1B.2.1 Analytics context transfer initiated by target NWDAF selected by the NWDAF service consumer The procedure in Figure 6.1B.2.1-1 is used when an NF decides to select a new NWDAF instance due to internal or external triggers, e.g. the NF starts serving a UE with analytics subscription information received upon UE context transfer procedure as described in TS 23.502 [3], or the NF starts to request NF related analytics, or the NF receives a "Termination Request" for an existing analytics subscription from an NWDAF. The NF sends to the target NWDAF information about the NWDAF previously used for analytics subscription, if available, in Nnwdaf_AnalyticsSubscription_Subscribe service operation. The target NWDAF may initiate the transfer of the analytics context, using the Nnwdaf_AnalyticsInfo_ContextTransfer or Nnf_DataManagement_Subscribe service operation. The procedure in Figure 6.1B.2.1-1 is also used when an Aggregator NWDAF decides to select a new NWDAF to request output analytics for analytics aggregation. For example, upon receiving a Termination Request from one of the NWDAFs that are collectively serving a request for analytics subscription as specified in clause 6.1A, the Aggregator NWDAF queries the NRF or UDM to select a target NWDAF as specified in clause 6.1A.3 using information e.g. the UE location, the 5GC NFs (identified by their NF Set IDs or NF types) serving the UE or to be contacted for data collection (if Area of Interest is not provisioned for the requested analytics), or the subset of AoI (if Area of Interest is provisioned for the requested analytics). Then, the Aggregator NWDAF sends information about the NWDAF previously used for analytics subscription, if available, in Nnwdaf_AnalyticsSubscription_Subscribe service operation towards the selected target NWDAF. Figure 6.1B.2.1-1: Analytics context transfer initiated by target NWDAF selected by the NWDAF service consumer 1. The NWDAF service consumer determines to select an NWDAF instance. The consumer discovers and selects the target NWDAF as specified in clause 5.2. 2. The consumer sends a request for analytics subscription to the target NWDAF using Nnwdaf_AnalyticsSubscription_Subscribe service operation, including information on the previous analytics subscription (i.e. NWDAF ID, Analytics ID(s), SUPIs, Analytics Filter Information for UE-related Analytics, Subscription Correlation ID, the Analytics Accuracy Request information (as defined in clause 6.1.3) when the Target NWDAF supports accuracy checking capability) which relates to the requested analytics subscription, if available. If the target NWDAF accepts the analytics subscription request, it sends Nnwdaf_AnalyticsSubscription_Subscribe response with a Subscription Correlation ID. If the target NWDAF does not receive information of previous analytics subscription in step 2, for UE related Analytics, the target NWDAF may discover previously used NWDAF in UDM as specified in clause 5.2. NOTE 1: If the selected target NWDAF instance is the same as the source NWDAF instance (as received from the other consumer in step 0), the target NWDAF will update the existing analytics subscription to the new analytics consumer. Following steps are skipped. NOTE 2: The consumer can provide information on the previous analytics subscription when, e.g. the consumer is an AMF and it received information from the old AMF, see clause 5.2.2.2.2 of TS 23.502 [3]. 3a. [Option 1] If the target NWDAF decides to request an analytics context transfer from the previously used NWDAF, it may make use of information sent in step 2 (e.g. the provided Subscription Correlation ID) and use the analytics context transfer procedure as specified in clause 6.1B.3. The target NWDAF may receive an ADRF ID or DCCF ID for collecting the historical data and/or analytics. If the target NWDAF supports the accuracy checking capability and the request received in step 2 contains the Analytics Accuracy Request information, the target NWDAF may include in the Nnwdaf_AnalyticsInfo_ContextTransfer request the field Requested Analytics Context Type with value set to Analytics accuracy related information and ML Model accuracy related information (both defined in clause 6.1B.4) in order to retrieve the necessary information for starting, respectively, the Analytics Accuracy Information generation as well as the registration as provider of ML Model Accuracy Information for the ML Model. 3b-c. [Option 2] If the target NWDAF decides to only request historical data and/or analytics, then it may collect the data and/or analytics via Nnf_DataManagement_Subscribe service, where the NFs may be either the ADRF, NWDAF or DCCF, as described in clauses 10.2.6, 7.4.2 and 8.2.2 respectively. Target NWDAF is now ready to generate analytics information and if applicable, Analytics Accuracy Information, taking into account the information received in step 3. The target NWDAF is also able to perform the registration as a new provider for an existing ML Model Accuracy Information process as defined in clause 6.2E.3.2. 4. [Optional] Source NWDAF may purge analytics context after completion of step 3a, if performed and if not already done, unsubscribes from the data source(s) and/or model source(s) that are no longer needed for the remaining analytics subscriptions. 5. [Optional] Target NWDAF may subscribe to relevant data source(s) and/or model source(s), if it is not yet subscribed to the data source(s) and/or model source(s). 6.1B.2.2 Analytics Subscription Transfer initiated by source NWDAF The procedure in Figure 6.1B.2.2-1 is used by an NWDAF instance to request the transfer of analytics subscription(s) to another NWDAF instance, using the Nnwdaf_AnalyticsSubscription_Transfer service operation defined in clause 7.2.5. If the source NWDAF discovers that the analytics consumer may change concurrently to this procedure, the source NWDAF should not perform the procedure. In such a case, the source NWDAF may send a message to indicate to the analytics consumer that it will not serve this subscription anymore. NOTE 1: To discover the possible change of analytics consumer, if the Analytics ID is UE related, the source NWDAF takes actions responding to external trigger (such as UE mobility), for example, checking if the Target of Analytics Reporting is still within the serving area of the analytics consumer, if the serving area information is available. NOTE 2: Handling of overload situation or preparation for a graceful shutdown are preferably executed inside an NWDAF Set, when available, therefore, not requiring an analytics subscription transfer as described in this clause. The procedure in Figure 6.1B.2.2-1 is applicable for analytics subscription transfer across NF Sets or if the NWDAF is not deployed in a Set. Figure 6.1B.2.2-1: Analytics subscription transfer initiated by source NWDAF 0. The analytics consumer subscribes to analytics from source NWDAF. The analytics consumer may send its NF ID or serving area, enabling NWDAF to determine whether the following analytics subscription transfer procedure is applicable. Optionally the source NWDAF subscribes to UE mobility events. 1. [Optional] Source NWDAF determines, e.g. triggered by a UE mobility event notification, to prepare an analytics subscription transfer to target NWDAF(s), as specified in the procedure illustrated in clause 6.1B.2.3. 2. Source NWDAF determines, e.g. based on the UE location information received and the analytics consumer's serving area either directly received in step 0 or indirectly received via NRF, to perform an analytics subscription transfer to target NWDAF(s). Therefore, the source NWDAF determines the analytics subscription(s) to be transferred to a target NWDAF. 3. Source NWDAF performs an NWDAF discovery and selects the target NWDAF. NWDAF discovery may be skipped if the target NWDAF had already been discovered as part of a prepared analytics subscription transfer. In the case of aggregated analytics from multiple NWDAFs, the source NWDAF may use the set of NWDAF identifiers related to aggregated analytics (see clause 6.1.3) to preferably select a target NWDAF that is already serving the consumer. If the analytics subscription to be relocated to a target NWDAF also includes the Analytics Accuracy Request information, the source NWDAF selects, if possible, a target NWDAF also with accuracy checking capability. If the source NWDAF does not discover a target NWDAF with accuracy checking capability, the source NWDAF notifies the analytics consumer with Accuracy Information Termination. Therefore, the consumer based on local policy may decide to unsubscribe the analytics ID or to keep using the analytics ID even without receiving the accuracy information. 4. Source NWDAF requests, using Nnwdaf_AnalyticsSubscription_Transfer Request service operation, a transfer of the analytics subscription(s) determined in step 2 to the target NWDAF. The request contains a callback URI of the analytics consumer. The request may also contain active data source ID(s) and ML Model related information, which are related to the analytics subscriptions requested to be transferred, if not already provided as part of the prepared analytics subscription transfer in the preparation procedure (see step 1). The ML Model related information contains the ID(s) of NWDAF(s) containing MTLF that provided the trained models and may contain the file address(es) of the trained ML Model(s), where the file address(es) of the trained ML Model(s) is included only when the source NWDAF itself provides the trained ML Model(s) for the analytics subscription(s) being transferred. The request message may also include "analytics context identifier(s)" indicating the availability of analytics context for particular Analytics ID(s). 5. Target NWDAF accepts the analytics subscription transfer and takes over the analytics generation and if applicable, the Analytics Accuracy Information generation, based on the information received from the source NWDAF. The target NWDAF may use analytics accuracy request information included in the analytics subscription transfer received in step 4 to start the process of checking and generating Analytics Accuracy Information for the consumer of the transfer analytics subscription. Target NWDAF may use the ML Model related information, if provided in the Nnwdaf_AnalyticsSubscription_Transfer request. If the ID(s) of NWDAF(s) containing MTLF is provided in the Nnwdaf_AnalyticsSubscription_Transfer request and the NWDAF(s) containing MTLF is part of the locally configured set of NWDAFs containing MTLF, target NWDAF may request or subscribe to the ML Model(s) from the indicated NWDAF(s) containing MTLF as specified in clause 6.2A and use the ML Model(s) for the transferred analytics subscription. If the file address(es) of the trained ML Model(s) is provided and if the NWDAF containing MTLF is part of the locally configured set of NWDAFs containing MTLF, the target NWDAF may retrieve the ML Model using the file address of the trained ML Model. If the provided ID(s) of NWDAF(s) containing MTLF are not part of the locally configured set of ID(s) of NWDAFs containing MTLF, the target NWDAF discovers the NWDAF(s) supporting MTLF that can provide trained ML Model(s) for the Analytics ID(s) as described in clause 5.2. NOTE 3: If not yet done during a prepared analytics subscription transfer, the target NWDAF allocates a new Subscription Correlation ID to the received analytics subscriptions. NOTE 4: The target NWDAF might already have received information on some/all of the analytics subscriptions as part of the prepared analytics subscription transfer request received in step 1 and, thus, might already have started to prepare for the analytics generation, e.g. by having already subscribed to relevant event notifications. 6. Target NWDAF informs the analytics consumer about the successful analytics subscription transfer using a Nnwdaf_AnalyticsSubscription_Notify message. A new Subscription Correlation ID, which was assigned by the target NWDAF, is provided in the Subscription Correlation ID and the old Subscription Correlation Id, which was allocated by the source NWDAF, is provided in the Subscription Change Notification Correlation ID parameter of this message as specified in clause 7.2.4. NOTE 5: Notification correlation information in the Nnwdaf_AnalyticsSubscription_Notify message allows the analytics consumer to correlate the notifications (of analytics output and if applicable of Analytics Accuracy Information) to the initial subscription request made with the source NWDAF in step 0. NOTE 6: The existing Analytics context in the source NWDAF is not deleted directly but will be purged first when it was collected by the target NWDAF. NOTE 7: If this subscription is used as input for analytics aggregation by the analytics consumer, the analytics consumer might inform the other NWDAFs instance participating in this analytics aggregation that the Set of NWDAF identifiers of NWDAF instances used by the NWDAF service consumer for this analytics aggregation (see clause 6.1.3) has changed using the Nnwdaf_AnalyticsSubscription_Subscribe service operation. 7. [Conditional] If "analytics context identifier(s)" had been included in the Nnwdaf_AnalyticsSubscription_Transfer Request received in step 4, the target NWDAF requests the "analytics context". The analytics context transfer procedure is specified in clause 6.1B.3. If the transfer request received by the target NWDAF also includes the Analytics Accuracy Request information, the target NWDAF will include in the Nnwdaf_AnalyticsInfo_ContextTransfer request the field Requested Analytics Context Type with value set to Analytics accuracy related information (as defined in clause 6.1B.4) in order to retrieve the necessary information for generating the Analytics Accuracy Information. The target NWDAF may also retrieve from source NWDAF containing AnLF the ML Model Accuracy Information for the ML Model when ML Model accuracy related information context type is included in the "analytics context identifier(s)" in the transfer request. Based on the retrieved ML Model accuracy related information, the target NWDAF containing AnLF registers as provider of ML Model Accuracy Information for the ML Model as defined in clause 6.2E.3.2. 8. [Optional] Target NWDAF subscribes to relevant data source(s), if it is not yet subscribed to the data source(s) for the data required for the Analytics. 9. Target NWDAF confirms the analytics subscription transfer to the source NWDAF. 10. [Optional] Source NWDAF unsubscribes with the data source(s) that are no longer needed for the remaining analytics subscriptions. In addition, Source NWDAF unsubscribes with the NWDAF(s) containing MTLF, if exist, which are no longer needed for the remaining analytics subscriptions. NOTE 8: At this point, the analytics subscription transfer is deemed completed, i.e. the source NWDAF can delete all information related to the successfully transferred analytics subscription. 11-12. Target NWDAF at some point derives new output analytics and Analytics Accuracy Information (if applicable) based on new input data and notifies the analytics consumer about the new analytics and new Analytics Accuracy Information (if applicable) using a Nnwdaf_AnalyticsSubscription_Notify message as specified in clause 6.1.1. 6.1B.2.3 Prepared analytics subscription transfer The procedure in Figure 6.1B.2.3-1 is used by an NWDAF instance to request another NWDAF instance for a prepared analytics subscription transfer from the source NWDAF instance, using the Nnwdaf_AnalyticsSubscription_Transfer service operation defined in clause 7.2.5. NOTE 1: The source NWDAF might determine that it needs to prepare to transfer analytics to another NWDAF instance, e.g. when the source NWDAF estimates for UE related analytics subscription that the UE might enter an area which is not covered by the source NWDAF (e.g. by subscribing to AMF event exposure service for UE mobility event notifications, by performing UE mobility analytics, or by subscribing to another NWDAF providing UE mobility analytics). If the source NWDAF discovers that the analytics consumer may change concurrently to this procedure, the source NWDAF does not perform the procedure. If the procedure makes use of predictions to determine the candidate NWDAFs, care must be taken with regards to load and signalling cost when sending data to an NWDAF that will not eventually start serving the UE. NOTE 2: The source NWDAF might also determine that it needs to prepare to transfer analytics subscriptions to another NWDAF instance, as the source NWDAF wants to resolve an internal load situation or prepare for a graceful shutdown. NOTE 3: Handling of overload situation or preparation for a graceful shutdown are preferably executed inside an NWDAF Set, when available. Figure 6.1B.2.3-1: Prepared analytics subscription transfer 0. Analytics consumer subscribes to the source NWDAF for certain analytics as specified in clause 6.1.1. 0a. Source NWDAF starts data collection from relevant data source(s) (e.g. NFs or OAM) as specified in clause 6.2. Source NWDAF starts generating requested analytics. 0b. [Conditional] (Only if the source NWDAF does not serve the whole PLMN and the requested analytics involves UE related data) The source NWDAF subscribes, using Namf_EventExposure_Subscribe service operation, to receive notifications on UE mobility events from AMF. 1. The source NWDAF determines that it needs to prepare to transfer analytics to another NWDAF instance. 2. The source NWDAF discovers candidate target NWDAF instances (e.g. NWDAFy and NWDAFz) supporting the requested analytics information for the predicted target area(s). NWDAF discovery and selection is specified in clause 6.3.13 of TS 23.501 [2]. In the case of aggregated analytics from multiple NWDAFs, the source NWDAF may use the set of NWDAF identifiers related to aggregated analytics (see clause 6.1.3) to preferably select a target NWDAF that is already serving the consumer. In the case of the analytics subscription to be transferred also includes an Analytics Accuracy Request information as defined in clause 6.1.3, the source NWDAF takes into consideration in the selection process whether target NWDAF have the accuracy checking capability. NOTE 4: In this procedure, NWDAFy and NWDAFz are examples for target NWDAF instances that are candidates to take over those analytic subscriptions. 3-4. In the case of a prepared analytics subscription transfer, the source NWDAF requests, using Nnwdaf_AnalyticsSubscription_Transfer Request, to the candidate target NWDAFs (e.g. NWDAFy and NWDAFz) to prepare for an analytics subscription transfer by including a "prepared analytics subscription transfer indication" in the request message. The request message includes information on the analytics subscriptions to be transferred. The request message may also include "analytics context identifier(s)" indicating the availability of analytics context for particular Analytics ID(s). The candidate target NWDAFs (e.g. NWDAFy and NWDAFz) respond to the request from the source NWDAF using a Nnwdaf_AnalyticsSubscription_Transfer Response message. 5-6. [Conditional] If "analytics context identifier(s)" had been included in the Nnwdaf_AnalyticsSubscription_Transfer Request received in step 4, the determined target NWDAFs (e.g. NWDAFy and NWDAFz) may request the "analytics context" from the source NWDAF by invoking the "Nnwdaf_AnalyticsInfo_ContextTransferservice" operation. The analytics context transfer procedure is specified in clause 6.1B.3. If the target NWDAF supports the accuracy checking capability and the request received in steps 3-4 contains the Analytics Accuracy Request information, the target NWDAF includes in the Nnwdaf_AnalyticsInfo_ContextTransfer request the field Requested Analytics Context Type with value set to Analytics accuracy related information (as defined in clause 6.1B.4) in order to retrieve the necessary information for starting the Analytics Accuracy Information. The target NWDAF may also include in the request the Requested Analytics Context Type with value set to ML Model accuracy related information in order obtain the proper information to register as a new provider of ML Model Accuracy Information for the ML Model as defined in clause 6.2E.3.2. NOTE 5: The target NWDAFs (e.g. NWDAFy and NWDAFz) can allocate a new Subscription Correlation ID to the received analytics subscriptions. 7. [Optional] Based on the information received from the source NWDAF, the target NWDAFs (e.g. NWDAFy and NWDAFz) start data collection from NFs or OAM (as specified in clause 6.2), analytics generation for the indicated analytics subscriptions and if applicable the analytics accuracy checking and generation (as specified in clause 6.2D). NOTE 6: After step 7, the source NWDAF initiates the analytics subscription transfer to the target NWDAF as specified in steps 4 to 12 of the analytics subscription transfer procedure illustrated in Figure 6.1B.2.2-1. 8. The source NWDAF cancels the prepared analytics subscription transfer to a candidate target NWDAF (e.g. NWDAFz), using Nnwdaf_AnalyticsSubscription_Transfer Request include an "analytics subscription transfer cancel indication". The target NWDAF (e.g. NWDAFz) confirms the cancelation to the source NWDAF and, if applicable, deletes any analytics data that is no longer needed. If the target NWDAF (e.g. NWDAFz), as part of the analytics subscription preparation, had already subscribed to entities to collect data or acquire ML Model, it unsubscribes to those entities if the subscriptions are not needed for other active analytics subscriptions with the target NWDAF. If the candidate NWDAF (e.g. NWDAFz), as part of the analytics subscription preparation, had already started the analytics accuracy generation (as specified in clause 6.2D), it cancels the generation and if any extra data collection has been started, this data collection is also stopped, if not used by any other process in the target NWDAF. Step 8 may take place any time after step 4 if the NWDAF determines that the candidate target NWDAF (e.g. NWDAFz) does no longer need to prepare for the analytics subscription transfer. In particular, the source NWDAF shall cancel the prepared analytics subscription transfer to all remaining candidate target NWDAFs after one target NWDAF has accepted the analytics subscription transfer (see NOTE 6). If the source NWDAF is not able to discover a target NWDAF with accuracy checking capability, the source NWDAF notifies the analytics consumer with Accuracy Information Termination. Therefore, the consumer may decide to unsubscribe the analytics ID or to keep using the analytics ID even without receiving the accuracy information. 6.1B.3 Analytics Context Transfer The procedure depicted in Figure 6.1B.3-1 is used by an NWDAF instance to request analytics context from another NWDAF instance, using the Nnwdaf_AnalyticsInfo_ContextTransfer service operation as defined in clause 7.3.3. This procedure, for example, can be invoked in the procedures described in clause 6.1B.2 to request the transfer of relevant analytics context. Figure 6.1B.3-1: Analytics Context Transfer The procedure of analytics context information transfer comprises the following steps: 1. The consumer NWDAF requests analytics context by invoking Nnwdaf_AnalyticsInfo_ContextTransfer request service operation. The parameters that can be provided in the request are listed in clause 6.1B.4. 2. The provider NWDAF responds with analytics context to the consumer NWDAF. The analytics context that can be provided in the response is listed in clause 6.1B.4. If the provider NWDAF stores analytics context (i.e. Historical output Analytics and/or Data related to Analytics) in ADRF, the provider NWDAF may include in the response the ADRF ID together with an indication of the Analytics Context Type stored in the ADRF (i.e. Historical output Analytics and/or Data related to Analytics). Upon receiving the analytics context, the consumer NWDAF may: - provide the pending output analytics or historical analytics information to the analytics consumer per the subscription/request; - use the historical data and analytics metadata in the analytics context to generate analytics; - use the analytics accuracy related information in the analytics context to activate the checking of Analytics Accuracy Information for the transferred analytics ID, generate and provide the Analytics Accuracy Information for the consumer. NOTE: The consumer NWDAF can analyse the timestamps of the historical data included in the analytics context in order to obtain the inference configuration used at the source NWDAF for data collection and may decide to use the same inference configuration for the analytics accuracy generation. - use the ML Model accuracy related information in the analytics context to determine the need for registration at the NWDAF containing MTLF with the information to enable the NWDAF containing MTLF to reassociated the data of the existing subscription for ML Model Accuracy Information to a new ML Model Accuracy Monitoring process at the target NWDAF containing AnLF, reusing the existing data (as further detailed in clause 6.2E.3). - subscribe to data collected for analytics with the data sources indicated in the analytics context; - if the ID(s) of the NWDAF(s) containing MTLF indicated in the analytics context is part of the locally configured (set of) IDs of NWDAFs containing MTLF, retrieve trained ML Model(s) from the indicated NWDAF(s) containing MTLF or based on the file address(es) of the trained ML Model(s) and use for analytics; and/or - subscribe to output analytics from the indicated NWDAFs that collectively serve the transferred analytics subscription and perform analytics aggregation on the output analytics using the analytics metadata information, based on the analytics subscription aggregation information. 6.1B.4 Contents of Analytics Context The Nnwdaf_AnalyticsInfo_ContextTransfer service operation is used to transfer analytics context from a source NWDAF instance to a target NWDAF instance, whereby the target NWDAF consumes the Nnwdaf_AnalyticsInfo_ContextTransfer service operation produced by the source NWDAF instance to request the analytics context as depicted in Figure 6.1B.3-1. The consumers of the Nnwdaf_AnalyticsInfo_ContextTransfer service operation (as specified in clause 7.3.3) provide the following input parameters: - A list of analytics context identifier(s): identify a set of analytics contexts that are available at the NWDAF instance providing this service and that are requested to be transferred to the consumer NWDAF instance. The analytics context identifier is provided as the following: - Subscription Correlation ID: identifies the analytics subscription for which the related analytics context is requested; or - A set of SUPI and associated Analytics ID for UE related Analytics; or - An Analytics ID for NF related Analytics. - [OPTIONAL] Requested Analytics Context Type per analytics context identifier: indicates which part of the analytics context the consumer wishes to receive. Following values are specified: - Pending output Analytics; - Historical output Analytics; - Analytics subscription aggregation information; - Data related to Analytics; - Aggregation related information; - ML Model related information; - Analytics accuracy related information. - ML Model accuracy related information. NOTE: A list of "analytics context identifiers" can be provided by the source NWDAF to the target NWDAF in an analytics subscriptions transfer request as described in clause 6.1B.2.2. Information allowing to identify an analytics context can also be provided by the NWDAF consumer to the target NWDAF in the Nnwdaf_AnalyticsSubscription_Subscribe request and based on this information the target NWDAF derives the "analytics context identifier", as defined in clause 6.1B.2.1. The producer NWDAF provides to the consumer of the Nnwdaf_AnalyticsInfo_ContextTransfer service operation (as specified in clause 7.3.3), the output information listed below: - (Set of) Analytics context matching the input parameters of the Nnwdaf_AnalyticsInfo_ContextTransfer request. If no Requested Analytics Context type parameters are available in the request, all available analytics context types are sent. Analytics context includes the following information parts, if available: - Analytics related: - Pending output analytics (i.e. not yet notified to the consumer). - Historical output analytics information. The content of the output analytics is specified in clause 6.1.3 as output information of the Nnwdaf_AnalyticsSubscription_Notify or Nnwdaf_AnalyticsInfo_Request service operations. - Timestamp(s) of the last output analytics provided to the analytics consumer(s). Value is set to 0 if no output analytics had been sent yet. - Analytics subscription aggregation information (only provided when analytics context is related to analytics aggregation): information about the analytics subscriptions that the source NWDAF has with the NWDAFs that collectively serve the transferred analytics subscription, which includes IDs and analytics metadata information of these NWDAFs for specific Analytics ID(s) and optionally input parameters of analytics exposure as defined in clause 6.1.3. - Data related to Analytics: - Historical data that is available at the source NWDAF and that is related to the analytics to be handed over to the target NWDAF. If available, the time period of the collected data, NF ID(s) of the data source(s) and information (e.g. filter and event reporting parameters) on the subscriptions with those data sources which were used to generate this historical data. - Aggregation related information: Related to analytic consumers that aggregate analytics from multiple NWDAF subscriptions: - (Set of) NWDAF identifiers of NWDAF instances used by the NWDAF service consumer when aggregating multiple analytic subscriptions. - ML Model related information: - ID(s) of NWDAF(s) containing MTLF: Instance/Set ID(s) of the NWDAF(s) containing MTLF from which the source NWDAF currently subscribes to the ML Model Information used for the analytics. - Optionally, file address(es) of the trained ML Model(s), which is included only when the source NWDAF itself provides the trained ML Model(s) for the analytics subscription(s) for which the related analytics context is requested. - Analytics Accuracy related information: The information is related to the parameters of the Analytics Accuracy Request Information and status of the analytics subscription due to NF consumer and NWDAF interactions pausing or resuming the analytics output. It includes: - Timestamp(s) of the last Analytics Accuracy Information provided to the analytics consumer(s). Value is set to 0 if no Analytics Accuracy Information had been sent yet; - Indication whether analytics subscription is paused; - Remaining time window of paused analytics subscription; - Ground truth information: data types and data sources of the ground truth per analytics used for the accuracy information computation. - ML Model accuracy related information: The information is related to the parameters of the ML Model Accuracy (i.e. Analytics accuracy for an ML model as described in clause 6.2E.3.3) Subscription Information requested by a NWDAF containing MTLF. It includes: - original Subscription Correlation ID for the ML Model Accuracy Information associated with the ML Model and/or analytics ID at the source NWDAF containing AnLF; - Analytics ID, Target of Analytics Reporting and the corresponding Analytics filter associated with the original ML Model Monitoring subscription at the source NWDAF containing AnLF; - NWDAF containing AnLF NF ID of source NWDAF; - The parameters used for the subscription for ML Model Accuracy Information for the given ML Model at the source NWDAF containing AnLF. 6.1C NWDAF Registration/Deregistration in UDM 6.1C.1 General The procedures in this clause are applicable to UE-related analytics (e.g. UE mobility analytics) for some network deployments, e.g. such with an NWDAF co-located to an AMF or SMF, where the NWDAF is configured to register in UDM for the UEs that it is serving or collecting data for and for the related Analytics ID(s). The procedures in this clause are also applicable to analytics that are not UE-related, when the NWDAF collects UE-related data. This enables NWDAF service consumers to discover the NWDAF instance that is already serving the UE for one or more Analytics ID(s). 6.1C.2 NWDAF Registration in UDM Figure 6.1C.2-1 shows the procedures for registration of the NWDAF in UDM for UE-related analytics or UE-related data collection. Figure 6.1C.2-1: NWDAF registration in UDM 1. NWDAF triggers a registration in UDM, e.g. based on local configuration in the NWDAF, the reception of a new Analytics subscription request, start of collection of UE related data or an OAM configuration action. 2. The NWDAF registers into UDM for the served UE, by sending Nudm_UECM_Registration request (UE ID, NWDAF ID, Analytics ID(s)). 3. UDM sends a response to NWDAF. 6.1C.3 NWDAF De-registration from UDM Figure 6.1C.3-1 shows the procedures for deregistration of the NWDAF in UDM. Figure 6.1C.3-1: NWDAF de-registration from UDM 1. NWDAF triggers a de-registration from a previous registration in UDM. This trigger may be that, e.g. the NWDAF has purged the analytics context for the UE (see clause 6.1B.4) for related Analytics ID(s), the NWDAF is no longer collecting data related to the UE, or an administrative action. 2. NWDAF sends Nudm_UECM_Deregistration request (UE ID, NWDAF ID, Analytics ID(s)). 3. UDM sends a response to NWDAF.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2 Procedures for Data Collection
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.1 General	The Data Collection feature permits NWDAF to retrieve data from various sources (e.g. NF such as AMF, SMF, PCF, NSACF, GMLC and AF; OAM), as a basis of the computation of network analytics. All available data encompass: - OAM global NF data, - Data available in NFs, e.g. behaviour data related to individual UEs or UE groups (e.g. UE reachability) and pre-computed metrics covering UE populations (e.g. number of UEs present in a geographical area), per spatial and temporal dimensions (e.g. per region for a period of time), - NF data available in the 5GC (e.g. NRF), - Data available in AF. When DCCF, ADRF, MFAF or NWDAF hosting DCCF or ADRF are present in the network, the data collection also follows the principles described in clause 6.2.6. The NWDAF shall use at least one of the following services: - the Generic management services as defined in TS 28.532 [6], the Performance Management services as defined in TS 28.550 [7] or the Fault Supervision services as defined in TS 28.545 [9], offered by OAM in order to collect OAM global NF data. - the Exposure services offered by NFs in order to retrieve data and other non-OAM pre-computed metrics available in the NFs. - Other NF services in order to collect NF data (e.g. NRF) - DCCF data management service to retrieve data using DCCF. The NWDAF shall obtain the proper information to perform data collection for a UE identified by a SUPI, a group of UEs identified by an Internal-Group-Id or any UE: - For an Analytics ID, NWDAF is configured with the corresponding NF Type(s) and/or event ID(s) and/or OAM measurement types. - NWDAF shall determine which NF instance(s) of the relevant NF type(s) are serving the UE, the group of UEs identified by an Internal-Group-Id or any UE, taking into account the S-NSSAI(s) and area of interest as defined in clause 7.1.3 of TS 23.501 [2]. - NWDAF invokes Nnf_EventExposure_Subscribe services to collect data from the determined NF instance(s) and/or triggers the procedure in clause 6.2.3.2 to subscribe to OAM services to collect the OAM measurement. The NWDAF performs data collection from an AF directly as defined in clause 6.2.2.2 or via NEF as defined in clause 6.2.2.3. According to the data collection request, the AF may further perform data collection from UE (see clause 6.4.2 and clauses 6.5.2-6.5.4) as defined in clause 6.2.8. The NWDAF shall be able to discover the events supported by a NF. Data collection procedures enables the NWDAF to efficiently obtain the appropriate data with the appropriate granularity. When a request or subscription for statistics or predictions is received, the NWDAF may not possess the necessary data to perform the service, including: - Data on the monitoring period in the past, which is necessary for the provision of statistics and predictions matching the Analytics target period. - Data on longer monitoring periods in the past, which is necessary for model training. Therefore, in order to optimize the service quality, the NWDAF may undertake the following actions: - The NWDAF may return a confidence parameter as stated in clause 6.1.3 expressing the confidence in the prediction produced. Prediction may be returned with zero confidence as described below. This confidence is likely to grow in the case of subscriptions. - The value of the confidence depends on the level or urgency expressed by the parameter "preferred level of accuracy" as listed in clause 6.1.3, the parameter "time when analytics information is needed" as listed in clause 6.1.3 and the availability of data. If no sufficient data is collected to provide an estimation for the preferred level of accuracy before the time deadline, the service shall return a zero confidence. Otherwise, the NWDAF may wait until enough data is collected before providing a response or a first notification. - In order to be prepared for future requests on analytics from NFs/OAM, the NWDAF, upon operator configuration, may collect data on its own initiative, e.g. on samples of UEs and retain the data collected in the data storage. NOTE 1: The NWDAF can send an error response to the analytics consumer to indicate that statistics are unavailable if the NWDAF was not prepared for future requests and did not collect data on its own initiative. The volume and maximum duration of data storage is also subject to operator configuration. The NWDAF may decide to reduce the amount of data collected to reduce signalling load, by either prioritizing requests received from analytics consumers, or reducing the extent (e.g. duration, scope) of data collection, or modifying the sampling ratios. When using sampling ratio, the NWDAF may, depending on the analytics required and based on local configuration, provide additional partitioning criteria to the NFs to allow for a better UEs representation and to request that the NFs first partition the UEs before applying sampling ratio (see Event Reporting Information as specified in TS 23.502 [3]). The NWDAF may provide one or multiple partitioning criteria in its request for data collection from NFs. In order to optimize the performance and accuracy of data collection and reporting and reduce the impact on data producers, the NWDAF may request subscriptions to the NFs with the variable reporting periodicity parameter. Depending on the condition provided as part of this parameter, different reporting periodicity are used. The condition is the load of the NF. The NWDAF may skip data collection phase when the NWDAF already has enough information to provide requested analytics. The data which NWDAF may collect is listed for each analytics in input data clause and is decided by the NWDAF. NOTE 2: NWDAF can skip data collection phase for some specific input data per the requested analytics e.g. when some of the data is already available at NWDAF for the requested analytics, or when NWDAF considers that some of the data is not needed at all to provide the requested analytics as per the analytics consumer request (e.g. based on preferred level of accuracy or based on the time when analytics are needed). Event exposure subscriptions for data collection from the AMF and the SMF may need to survive after the removal of UE context in the AMF including event exposure subscriptions, or upon the creation of new UE context in AMF or SMF. In order for event exposure subscriptions in AMF and SMF to be (re)created in these cases, the NWDAF may subscribe to the events in AMF and/or SMF via UDM for a UE identified by a SUPI or a group of UEs identified by an Internal-Group-Id, as specified in clause 4.15.4.4 of TS 23.502 [3]. In hierarchical interactions among NWDAFs, without standalone DCCF, or co-located DCCF, the efficiency of data collection can be achieved by inter-NWDAF instance cooperation among NWDAF instances on different levels of the hierarchy. An efficient data collection means that the same data required for the same Analytics ID or different Analytics IDs should not be collected multiple times by the different NWDAFs of the hierarchy.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.2 Data Collection from NFs
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.2.1 General	The Data Collection from NFs and SCP is used by NWDAF to subscribe/unsubscribe at any 5GC NF to be notified for data on a set of events. The Data Collection from NFs and SCP is based on the services of AMF, SMF, UDM, PCF, NRF, NSACF, UPF, LMF and AF (possibly via NEF): - Event Exposure Service offered by each NF as defined in clause 4.15 and clause 5.2 of TS 23.502 [3]. - other NF services (e.g. Nnrf_NFDiscovery and Nnrf_NFManagement in NRF as defined in clause 4.17 of TS 23.502 [3]) This data collection service is used directly in order to retrieve behaviour data for individual UEs or groups of UEs (e.g. UE reachability) and also to retrieve global UE information (e.g. Number of UEs present in a geographical area). Table 6.2.2.1-1: NF Services consumed by NWDAF for data collection Service producer Service Reference in TS 23.502 [3] or other indicated specification AMF Namf_EventExposure (NOTE 2) 5.2.2.3 5.2.3.5 SMF Nsmf_EventExposure (NOTE 2) 5.2.8.3 5.2.3.5 PCF Npcf_EventExposure (for a group of UEs identified by an Internal-Group-Id or any UE) Npcf_PolicyAuthorization_Subscribe (for a specific UE) 5.2.5.7 UDM Nudm_EventExposure 5.2.3.5 NEF Nnef_EventExposure 5.2.6.2 AF Naf_EventExposure 5.2.19.2 NRF Nnrf_NFDiscovery 5.2.7.3 Nnrf_NFManagement 5.2.7.2 NSACF Nnsacf_SliceEventExposure 5.2.21.4 UPF Nsmf_EventExposure or Nupf_EventExposure 5.2.8.3 5.2.26.2 SCP Nscp_EventExposure Service 5.2.28.2 LMF Nlmf_DataExposure Clause 8.3.4 of TS 23.273 [39] NOTE 1: There is no data collected from the PCF by the NWDAF defined in this Release of the specification. NOTE 2: The Nudm_EventExposure can be used when NWDAF uses the procedures specified in clause 4.15.4.4 of TS 23.502 [3] to subscribe to AMF or SMF via UDM. To retrieve data related to a specific UE, there are two cases: - If no Area of interest is indicated by the consumer, the NWDAF shall first determine which NF instances are serving this UE as stated in table 6.2.2.1-2 unless the NWDAF has already obtained this information due to recent operations related to this UE. - If an Area of interest is indicated, the NWDAF can: - First determine the AMF serving the UE and subscribe UE location from the AMF. Once the UE is in or moves into the Area of interest, the NWDAF determines which NF instances are serving this UE as stated in table 6.2.2.1-2 unless the NWDAF has already obtained this information due to recent operations related to this UE; or - Determine the NF instances of a given type of network function serving the Area of interest by querying NRF unless the NWDAF has already obtained this information due to recent operations related to this UE. Table 6.2.2.1-2: NF Services consumed by NWDAF to determine which NF instances are serving a UE Type of NF instance (serving the UE) to determine NF to be contacted by NWDAF Service Reference in TS 23.502 [3] UDM NRF Nnrf_NFDiscovery 5.2.7.3 AMF UDM Nudm_UECM 5.2.3.2 SMF UDM Nudm_UECM 5.2.3.2 BSF NRF Nnrf_NFDiscovery 5.2.7.3 PCF BSF Nbsf_Management 5.2.13.2 NEF NRF Nnrf_NFDiscovery 5.2.7.3 NWDAF NRF UDM Nnrf_NFDiscovery Nudm_UECM 5.2.7.3 5.2.3.2 NSACF NRF Nnrf_NFDiscovery 5.2.7.3 GMLC NRF Nnrf_NFDiscovery 5.2.7.3 The UDM instance should be determined using NRF as described in clause 4.17.4 of TS 23.502 [3] and factors to determine as described in clause 6.3.8 of TS 23.501 [2]. The AMF, SMF instances should be determined using a request to UDM providing the SUPI. To determine the SMF serving a PDU session, the NWDAF should in addition provide the DNN and S-NSSAI of this PDU Session; otherwise the NWDAF will obtain a list of possibly multiple SMFs (e.g. one per PDU session). The BSF instance is discovered and selected according to TS 23.503 [4], clause 6.1.1.2.2. The PCF instance serving UE PDU Session(s) should be determined using a request to BSF with the allocated UE address, DNN and S-NSSAI. To collect data (e.g. current number of UEs registered in a network slice or current number of PDU Sessions established in a network slice) from NSACF, the NSACF instance is discovered and selected as specified in clause 6.3.22 of TS 23.501 [2]. When NWDAF receives a request addressed to an Internal Group ID from a consumer, NWDAF may need to initiate data collection from several 5GC NFs, such as AMF, SMF, UDM, PCF, AF (e.g. via NEF), etc. If an Area of interest is indicated by the consumer, NWDAF may first discover the instances of the required 5GC NFs deployed in the network, e.g. by querying NRF, otherwise: - For discovering the UDM, NWDAF can query the NRF with the Internal Group ID as the target of the query. - For discovering AMF, SMF, PCF, NEF and AF, NWDAF may need to discover all instances in the network by using the Nnrf_NFDiscovery service. NOTE 3: It is assumed that all members of an Internal Group ID belong to the same UDM Group ID. NWDAF can select a UDM instance supporting the UDM Group ID of the Internal Group ID. Then, if data needs to be collected from AMF, SMF, UDM and PCF, NWDAF may initiate the data collection with the Internal Group ID as the target, e.g. subscribing to the event exposure in all the instances of a given type of network function. This subscription to all the instances of required source of event exposure handles, e.g. mobility of UEs across AMFs, or initiation of new PDU sessions with different allocated SMFs. For collecting data from AMF and SMF, NWDAF may collect the data directly from AMF and/or SMF, or indirectly via UDM, according to clause 4.15.4.4 of TS 23.502 [3]. The indirect method may be required if the event exposure subscription from NWDAF, for a UE identified by a SUPI or a group of UEs identified by an Internal-Group-Id, needs to survive the removal of UE context in the AMF including event exposure subscriptions, or upon the creation of new UE context in AMF or SMF serving the UE or group of UEs. In this case the UDM is responsible for (re)creating event exposure subscriptions in AMF and SMF, as specified in clause 4.15.4.4 of TS 23.502 [3]. The NWDAF determines to collect data from a trusted AF supporting specific Event ID(s) and serving specific application(s) based on internal configuration. The NEF instance that is serving a specific network slices and/or applications of a UE should be determined using NRF using optional request parameters as defined in clause 6.3.14 of TS 23.501 [2] If NWDAF needs to collect data from an AF deployed outside the operator's domain, the NWDAF shall contact NEF with a SUPI or Internal Group ID as the target of the data collection. NEF is responsible for translation of SUPI to GPSI, or internal to external group identifiers, by querying UDM, prior to contacting the AF. NOTE 4: It is assumed that an AF is provisioned with the list of UE IDs (GPSIs or SUPIs) belonging to an External or Internal Group ID. NWDAF may collect data directly from UPF for a specific UE identified by a SUPI, a group of UEs identified by an Internal-Group-Id or for any UEs as defined in clause 5.8.2 of TS 23.501 [2] and in clause 4.15.4.5 of TS 23.502 [3]. NWDAF may subscribe indirectly via SMF or directly to UPF for UPF data collection. The direct subscription to UPF event exposure service is only for data collection for any UE e.g., to collect user data usage information for NWDAF NF Load analytic, and if the subscription is not including any of the parameters described in Table 4.15.4.5.1-1 of TS 23.502 [3]. NOTE 5: To avoid causing high UPF load due to extensive reporting related to all traffic flows, the NWDAF can preferably subscribe for reporting for some UEs only. To retrieve required data for any UE, the NWDAF may subscribe to events from the AMF and/or SMF instances or UPF instances it has determined, setting the target of event reporting to "any UE" and the event filter(s) according to the Analytics Filter Information. Alternatively, if the required data is communication related and for any UE within an Area of interest, the NWDAF can obtain from the AMF instances it has determined a list of UEs located within the Area of Interest. Based on the obtained UE list, for each UE in the list, the NWDAF retrieves the SMF serving the UE and the NWDAF subscribes to data from the relevant SMF per each specific UE. The indirect event exposure subscription to AMF or SMF via UDM is not available for "any UE" or "any UE within an Area of interest". If the required data is collected from UE via AF as described in clause 6.2.8 and the Target of Analytics Reporting received from consumer is "any UE", the NWDAF may either set the target of event reporting to "any UE" in the data collection request to the AF, or may determine a list of UEs from AMF and/or SMF based on the Analytics Filter Information and send the data collection request to the AF for the determined list of UEs. NOTE 6: If NWDAF requires collecting data from either AMF or SMF for "any UE" or "any UE within an Area of Interest", NWDAF can use the direct Event Exposure subscription to AMF or SMF, since subscriptions to "any UE" or "any UE within an Area of Interest" are persistent by nature in AMF or SMF, due to not being linked to a UE context. To retrieve data related to "any UE" based on Analytics Filter Information, the NWDAF shall first determine which NF instances are matching the Analytics Filter Information (see clause 6.7.5.1) as stated in table 6.2.2.1-3 unless the NWDAF has already obtained this information due to recent operations related to this Analytics Filter Information. Table 6.2.2.1-3: NF Services consumed by NWDAF to determine which NF instances are matching analytics filters Type of NF instance (matching analytics filters) to determine NF to be contacted by NWDAF Service Reference in TS 23.502 [3] AMF, SMF, UPF NRF Nnrf_NFDiscovery 5.2.7.3 To retrieve data related to Analytics IDs for "any UE" with Analytics Filter Information defining an area of interest in terms of TA or Cells and/or with specific S-NSSAIs, NWDAF requires the network slice association information to properly determine the AMFs to collect data from as well as the proper queries to OAM for data collection. NOTE 7: Examples of Analytics ID requiring NWDAF to use network slice association information for data retrieval are: network performance clause 6.6.1; user data congestion clause 6.8.1; QoS Sustainability clause 6.9; Dispersion Analytics clause 6.10.1; observed service experience for a Network Slice clause 6.4.1; and slice load analytics clause 6.3. The network slice association information comprises the TAs associated with each AMF and for each TAI its associated access type, cells and list of supported S-NSSAIs (including indication of S-NSSAIs restricted by AMF). Additionally, the mapping of cells per TAI and supported S-NSSAIs (including indication of S-NSSAIs restricted by AMF) for TAI for each AMF can change and NWDAF shall obtain this accurate information in order to properly retrieve data for analytics generation. In order to derive the network slice association information, NWDAF may be configured with the mapping of cells per TAI and the S-NSSAIs per TAI. NWDAF may subscribe to the "S-NSSAIs per TAI mapping" event exposed by AMF. The NWDAF may use the configured information (when the analytics subscription or request is at cell granularity) and the area of interest in the analytics subscription or request to retrieve from AMF the list of supported S-NSSAIs (including indication of S-NSSAIs restricted by AMF) per TAI and access types per TAI for each AMF in the required area of interest. NWDAF consumes the "S-NSSAIs per TAI mapping" event exposed by AMF using, as target of event reporting, the list of TAIs based on the area of interest received in the Analytics Filter Information or identified by the mapping of the Cells per TAI matching to the Cell granularity included in the Analytics Filter Information. The AMF "S-NSSAIs per TAI mapping" event output contains, for each of the TAIs requested by NWDAF, its associated access type and the list of supported S-NSSAIs (including indication of S-NSSAIs restricted by AMF). To retrieve data from SMFs for Analytics IDs subscription or requests for "any UE" including Analytics Filter Information with specific Applications, DNNs, DNAIs and area of interest per TA granularity, NWDAF shall first discover the SMF serving the area of interest via NRF. NOTE 8: Examples of Analytics ID requiring NWDAF to collect data related to PDU sessions associated with an AoI with TA granularity are: network performance clause 6.6.1; user data congestion clause 6.8.1, QoS Sustainability clause 6.9. NWDAF may directly consume events from the discovered and selected SMF using the event target set to "any PDU session" and event filters with the same parameters of the Analytics Filter Information, i.e. list of Application IDs and/or DNNs and/or DNAI and the area of interest related to the requested Analytics ID. 1. When SMF supports the exchange of UE Location parameter when SMF interacts with AMF via Nsmf_PDUSession_Create/Update/CreateSMContext/UpdateSMContext due to session establishment, modification, or release, service request, or handover procedures (as defined in clause 5.2.8.2 of TS 23.502 [3]), SMF can directly map the PDU sessions to an AoI with TA granularity. If there are any changes in PDU sessions in the area of interest, for the Application ID and/or DNN and/or DNAI subscribed by NWDAF, SMF notifies the detected changes to NWDAF via Nsmf_EventExposure_Notify service operation, enabling NWDAF to keep an updated map of SMF and PDU sessions associated with the Analytics Filter Information in an area of interest. 2. When SMF does not support the exchange of UE Location parameter when SMF interacts with AMF but supports the mapping of PDU sessions per TA (as defined in clause 5.6.11 of TS 23.501 [2]), SMF may subscribe to UE mobility event notifications of AMF as described in clause 5.3.4.4 of TS 23.501 [2] using event ID "UE moving in or out of Area of Interest" and Event Filters as described in Table 5.2.2.3.1-1 from TS 23.502 [3] to retrieve the list of SUPIs (and GPSIs if available) in the area of interest. Based on the retrieved list of SUPIs in the area of interest, SMF identifies the PDU sessions in the area of interest. 3. When SMF does not support the exchange of UE Location parameter when SMF interacts with AMF nor supports the mapping of PDU sessions per TA (as defined in clause 5.6.11 of TS 23.501 [2]), SMF rejects the request from NWDAF. Upon the reject, NWDAF identifies the need to create the mapping of PDU sessions per TA. NWDAF subscribes to UE mobility event notifications of AMF as described in clause 5.3.4.4 of TS 23.501 [2] using event ID "UE moving in or out of Area of Interest" and Event Filters as described in Table 5.2.2.3.1-1 from TS 23.502 [3] to retrieve the list of SUPIs (and GPSIs if available) in the area of interest. Based on the retrieved list of SUPIs in the area of interest, NWDAF subscribes to the SMFs serving the UEs in the area of interest and derives the mapping of PDU sessions per TA. To train an ML model or to do ML model performance monitoring for LMF-based AI/ML positioning, the NWDAF can collect input data from LMF as defined in clause 6.22.4 of TS 23.273 [39]. An NWDAF may require to discover and select other NWDAFs for UE related analytics. In this case, the NWDAF may discover from UDM if an NWDAF is already collecting data related to the UE, as specified in clauses 5.2 and 6.1C.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.2.2 Procedure for Data Collection from NFs	The procedure in Figure 6.2.2.2-1 is used by NWDAF to subscribe/unsubscribe at NFs in order to be notified for data collection on a related event (s), using Event Exposure Services as listed in Table 6.2.2.1-1. Depending on local regulation requirements, user consent for UE related data collection and usage of collected data may be required. User consent is defined for a specific purpose such as, e.g. analytics or model training. NWDAF checks user consent taking the purpose for data collection and usage of these data into account. Figure 6.2.2.2-1: Event Exposure Subscribe/unsubscribe for NFs 1. The NWDAF checks if data is to be collected for a user, i.e. SUPI or GPSI, then, depending on local policy and regulations, the NWDAF checks the user consent by retrieving the user consent information from UDM using Nudm_SDM_Get including data type "User consent". If user consent is not granted, NWDAF does not subscribe to event exposure for events related to this user and the data collection for this SUPI or GPSI stops here. 2. If the user consent is granted, the NWDAF subscribes to UDM to notifications of changes on subscription data type "User consent" for this user using Nudm_SDM_Subscribe. 3. The NWDAF subscribes to or cancels subscription for a (set of) Event ID(s) by invoking the Nnf_EventExposure_Subscribe/Nnf_EventExposure_Unsubscribe service operation. NOTE 1: The Event ID(s) are defined in TS 23.502 [3]. 4. If NWDAF subscribes to a (set of) Event ID(s), the NFs notify the NWDAF (e.g. with the event report) by invoking Nnf_EventExposure_Notify service operation according to Event Reporting Information in the subscription. When the Reporting type is provided at step 1, the NWDAF determines that the events are disappeared, if the same events are included in the notification compared to the previous notification. Otherwise, NWDAF determines the events are newly appeared or changed. Also, the NWDAF restores the events that are not included in the notification, but included in the previous notification. If the Granularity of dynamics is applied to the subscription, the NWDAF shall infer the events in the NF from the events in the previous notification with the applied Granularity of dynamics. NOTE 2: The Event Reporting Information are defined in TS 23.502 [3]. NOTE 3: The NWDAF can use the immediate reporting flag as defined in Table 4.15.1-1 of TS 23.502 [3] to meet the request-response model for data collection from NFs. NOTE 4: This procedure is also used when the NWDAF subscribes for data from a trusted AF. 5. The UDM may notify the NWDAF on changes of user consent at any time after step 2. 6-7. If user consent is no longer granted for a user for which data has been collected, the NWDAF shall unsubscribe to any Event ID to collect data for that SUPI or GPSI. The NWDAF may unsubscribe to be notified of user consent updates from UDM for each SUPI for which data consent has been revoked.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.2.3 Procedure for Data Collection from AF via NEF	The procedure in Figure 6.2.2.3-1 is used by NWDAF to collect information from AFs via the NEF. NOTE 1: In this release, AF registers its available data to NWDAF via OAM configuration at NEF. The AF collectable data information includes: AF identification, AF service identification (e.g. endpoint information of Naf_EventExposure), available data to be collected per application (e.g. identified by Event ID(s)). Figure 6.2.2.3-1: Data Collection from AF via NEF 1a. After the registration of AF available data at the NEF, NEF generates an event exposure with new EventID to be associated with available data to be collected from AF. NEF invokes Nnrf_NFManagement_NFUpdate_request service operation to update its registration information (i.e. NEF Profile) including the generated Event IDs and associated AF identification, Application ID(s) (i.e. internal application ID or Application ID known in the core network). 1b. NRF stores the received NEF registration information including available data to be collected from AF. 1c. NRF sends Nnrf_NFManagement_NFUpdate_response message to NEF. 1d. When NWDAF needs to discovery the available data from AFs and the appropriated NEF to collect this data, NWDAF invokes Nnrf_NFDiscovery_Request_request service operation using as parameter the NEF NF Type and optionally a list of Event ID(s), AF identification and application ID. 1e. NRF matches the requested query for available data in AFs with the registered NEF Profiles and sends this information via Nnrf_NFDiscovery_Request_response message to NWDAF. NOTE 2: After the registration and discovery procedure described in step 1, NWDAF identifies the available data per AF per application and the proper NEF to collect such data. 2. The NWDAF subscribes to or cancels subscription to data in AF via NEF by invoking the Nnef_EventExposure_Subscribe or Nnef_EventExposure_Unsubscribe service operation. If the event subscription is authorized by the NEF, the NEF records the association of the event trigger and the NWDAF identity. NOTE 3: User consent for retrieving user data in AF via NEF is not specified in this Release. 3. Based on the request from the NWDAF, the NEF subscribes to or cancels subscription to data in AF by invoking the Naf_EventExposure_Subscribe/ Naf_EventExposure_Unsubscribe service operation. 4. If the NEF subscribes to data in AF, the AF notifies the NEF with the data by invoking Naf_EventExposure_Notify service operation according to Event Reporting Information in the subscription. 5. If the NEF receives the notification from the AF, the NEF notifies the NWDAF with the data by invoking Nnef_EventExposure_Notify service operation. When the Reporting type is provided at step 2, the NWDAF determines that the events are disappeared, if the same events are included in the notification compared to the previous notification. Otherwise, NWDAF determines the events are newly appeared or changed. Also, the NWDAF restores the events that are not included in the notification, but included in the previous notification. If the Granularity of dynamics is applied to the subscription, the NWDAF shall infer the events in the AF from the events in the previous notification and the applied Granularity of dynamics.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.2.4 Procedure for Data Collection from NRF	The NWDAF may use NRF services and Network Function service framework procedures as defined in clause 5.2.7 and clause 4.17 of TS 23.502 [3]: - NF/NF service discovery procedures (in clause 4.17.4 of TS 23.502 [3]) and Nnrf_NFDiscovery service (in clause 5.2.7.3 of TS 23.502 [3]) in order to dynamically discover the NF instances and services of the 5GC. Such discovery may be performed on a periodic basis, or under specific circumstances. - NF/NF service status subscribe/notify procedures (in clause 4.17.7 of TS 23.502 [3]) and Nnrf_NFManagement service (in clause 5.2.7.2 of TS 23.502 [3]) in order to be notified about the change of status of an NF. The service operations for obtaining status information are NFStatusSubscribe and NFStatusNotify, from the Nnrf_NFManagement service. The information provided by the NRF to the NWDAF with the Nnrf_NFDiscovery_Request and the Nnrf_NFManagement_NFStatusNotify service operations are the NF Profiles which includes the supported NF services of the NFs as defined in clause 5.2.7 of TS 23.502 [3]. Such information can be used to set-up and maintain a consistent network map for data collection and also, depending on use cases, to perform analytics (e.g. NF load analytics as defined in clause 6.5). If the NWDAF needs to keep a consistent network map for data collection from AMFs and SMFs associated with a list of TAs, the subscription to NFStatusSubscribe for such type of NFs may include the list of TAs target.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.2.5 Usage of Exposure framework by the NWDAF for Data Collection	The NWDAF shall subscribe (and unsubscribe) to the Event exposure service from NF(s) reusing the framework defined in clause 4.15 of TS 23.502 [3]. This framework supports the possibility for the NWDAF to indicate/request: - Events-ID: one or multiple Event ID(s) defined in clause 4.15.1 of TS 23.502 [3]. - Target of Event Reporting defined in clause 4.15.1 of TS 23.502 [3]: the objects targeted by the Events. Within a subscription, all Event ID(s) are associated with the same target of event reporting. In the case of NWDAF, the objects can be UE(s), UE group(s), any UE. - Event Filter Information defined in clause 4.15.1 of TS 23.502 [3]. This provides Event Parameter Types and Event Parameter Value(s) to be matched against. - A Notification Target Address and a Notification Correlation ID as defined in clause 4.15.1 of TS 23.502 [3], allowing the NWDAF to correlate notifications received from the NF with this subscription. - Event Reporting Information described in TS 23.502 [3] Table 4.15.1-1 and the muted stored events exposure as described in clause 6.2.7. - Expiry time as defined in clause 4.15.1 of TS 23.502 [3]. The notifications from NFs/AFs contain on top of the Event being reported (and of dedicated information being reported for this event): - the Notification Correlation Information provided by the NWDAF in its request; - (when applicable to the event) the Target Id e.g. UE ID (SUPI and if available GPSI); and - a time stamp.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.3 Data Collection from OAM
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.3.1 General	The NWDAF may collect relevant management data from the services in the OAM as configured by the PLMN operator. ‐ NG RAN or 5GC performance measurements as defined in TS 28.552 [8]. ‐ 5G End to end KPIs as defined in TS 28.554 [10]. ‐ 5G application layer measurements from the UEs in the specified area for specified end user service type as defined in TS 28.404 [54], TS 28.405 [55] and TS 28.406 [56]. NWDAF shall use the following services to have access to the information provided by OAM: - Generic performance assurance and fault supervision management services as defined in TS 28.532 [6]. ‐ PM (Performance Management) services as defined in TS 28.550 [7]. ‐ FS (Fault Supervision) services defined in TS 28.545 [9]. - MDA (Management Data Analytics) services as defined in TS 28.104 [45]. NWDAF can be configured to invoke the existing OAM services to retrieve the management data that are relevant for analytics generation, which may include NF resources usage information (e.g. usage of virtual resources assigned to NF) and NF resource configuration information (e.g. life cycle changes of NF resource configurations). OAM perform the required configuration in order to provide the information requested by NWDAF subscription and perform the tasks, e.g. data collection, data processing, associated with the subscribed request from NWDAF. Another usage of OAM services is when the target of data collection is a specific UE, via MDT based retrieval of information: - Measurement collection for MDT as defined in TS 37.320 [20]. In addition, NWDAF can be provisioned with Network Slice information (i.e. as defined by the NetworkSliceInfo specified in TS 28.541 [22]) when a slice is created or modified via OAM configuration mechanism as defined in TS 28.541 [22] and TS 28.532 [6].
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.3.2 Procedure for data collection from OAM	The interactions between NWDAF and OAM for data collection are illustrated in Figure 6.2.3.2-1. The data collected depends on the use cases. This figure is an abstraction of the OAM performance data file report management service that is defined TS 28.532 [6]. The actual OAM services and reporting mechanisms that NWDAF may use are specified in TS 28.532 [6], TS 28.550 [7] or TS 28.545 [9]. The flow below assumes the NWDAF is configured on how to subscribe to the relevant OAM services. OAM shall setup the required mechanisms to guarantee the continuous data collection requested by NWDAF. Figure 6.2.3.2-1: Data collection from OAM performance data file report management service 1. (Clause 11.6.1.3.2 of TS 28.532 [6]), Subscribe (Input): NWDAF subscribes to the notification(s) related to the services provided by the management service producer. 2. (Clause 11.6.1.3.3 of TS 28.532 [6]), Subscribe (Output): management service producer responses to NWDAF if the subscription is success or not. 3. Data processing: management service producer prepares the data. 4. (Clause 11.6.1.1 of TS 28.532 [6]), Notification (notifyFileReady): management service producer notifies the data file is ready. As the final step, NWDAF fetches data by using file transfer protocols as defined in clause 11.6.2 of TS 28.532 [6]. NOTE 1: The call flow in Figure 6.2.3.2-1 only shows a subscribe/notify model for the simplicity, however both request-response and subscription-notification models are supported. NOTE 2: NWDAF is configured with the Network Slice information (i.e. NetworkSliceInfo including a DN (Distinguished Name) of the NetworkSlice managed object relating to the network slice instance associated to the S-NSSAI and NSI ID if available as defined in TS 28.541 [22]). Based on the Network Slice information, the NWDAF uses the DN (Distinguished Name) to identify the NetworkSlice managed object relating to the S-NSSAI and NSI ID and consumes the management services to collect the management data of the corresponding NetworkSlice managed object (including the NRF serving the network slice, the NFs associated to the network slice, the NG RAN or 5GC performance measurements defined in TS 28.552 [8], or the 5G end to end KPIs defined in TS 28.554 [10]) provided by OAM.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.4 Correlation between network data and service data	The Correlation information in input data which helps NWDAF correlate data from different NFs, OAM and UE application(s) is defined in Table 6.2.4-1, which is subject to all the network data analytics. NOTE: For simplicity, the correlation information is not listed in the input data per network data analytics. Table 6.2.4-1: Correlation Information Correlation Information Description Timestamp, IP address 5-tuple To correlate the data from AF and from UPF. Timestamp, AN Tunnel Info (Clause 9.3.2.2 of TS 38.413 [16]) To correlate the UPF data and OAM data which are reported by the RAN (e.g. Reference Signal Received Power or Reference Signal Received Quality as defined in Table 6.4.2-3). Timestamp, UE IP address To correlate the data from UPF and SMF. Timestamp, SUPI To correlate data from SMF and AMF. Timestamp, SUPI, DNN, S-NSSAI or UE IP address To correlate data from SMF and PCF. Timestamp, RAN UE NGAP ID (Clause 9.3.3.2 of TS 38.413 [16]) and Global RAN Node ID To correlate the AMF data and OAM data reported by the RAN (e.g. Reference Signal Received Power or Reference Signal Received Quality as defined in Table 6.4.2-3). Timestamp, Application ID, IP filter information To correlate data from SMF and AF. Timestamp, UE ID or UE IP address, Application ID, DNN, S-NSSAI To correlate data from 5GC NF (e.g. SMF, UPF) and UE Application (via the AF).
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.5 Time coordination across multiple NWDAF instances
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.5.1 General	In certain situations, an NWDAF Service Consumer expects to receive analytics by a given time. In particular, when an NWDAF Service Consumer is collecting analytics from multiple NWDAFs it can be necessary to coordinate the timing of the analytics subscriptions/requests from the same NWDAF service consumer. The NWDAF Service Consumer may use "time when analytics information is needed parameter" (see clause 6.1.3) as a dynamic timer to indicate the minimum time it is going to wait (i.e. "expected waiting time") to receive the analytics collected from multiple NWDAFs.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.5.2 Procedure for time coordination across multiple NWDAFs	Figure 6.2.5.2-1: Procedure for time coordination across multiple NWDAFs 1a-1b. On analytics request/subscription, the NWDAF Service Consumer indicates the "expected waiting time" as "time when analytics is needed" parameter to those NWDAFs from which it expects to receive the analytics latest by the "time when analytics information is needed", using either the Nnwdaf_AnalyticsInfo_Request or Nnwdaf_AnalyticsSubscription_Subscribe service operation. In this example, NWDAF1 and NWDAF2 are the NWDAFs with tightly related analytics. 2a-2b. Each NWDAF generates the requested analytics based on data from related data sources. In this example, NWDAF1 processes data from NF1 and NWDAF2 processes data from NF2. 3a-3b. [Optional] If the "time when analytics information is needed" is reached, but the analytics is not ready, the NWDAF may indicate a "revised waiting time" in an error response or error notification, using either the Nnwdaf_AnalyticsInfo_Request response or Nnwdaf_AnalyticsSubscription_Notify service operation, depending on the service used in step 1. 4a-4b. [Optional] On receiving an indicated "revised waiting time" as part of an error response or error notification, the NWDAF Service consumer may use the "revised waiting time" to update the "time when analytics information is needed" parameter for future analytics requests/subscriptions to the same group of NWDAFs. 5a-5b. If the value of the "time when analytics information is needed" was updated in step 4, the NWDAF Service Consumer, in future requests or within current subscription, indicates the new expected waiting time as "time when analytics information is needed" to all NWDAFs with tightly related analytics, using either the Nnwdaf_AnalyticsInfo_Request or Nnwdaf_AnalyticsSubscription_Subscribe service operation. NOTE 1: Steps 3a-3b and steps 4a-4b may happen in different orders depending on the timing of analytics collection (from other NFs, e.g. NF1 or NF2) or processing. NOTE 2: Parameter "time when analytics is needed" as in steps 1a-1b, 3a-3b, 4a-4b and 5a-5b can be per individual Analytics ID.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.6 Enhanced Procedures for Data Collection
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.6.0 General	Data collection may be performed via DCCF, MFAF and ADRF, when such NFs are deployed, or via NWDAF (hosting DCCF and/or ADRF). The NF services used for DCCF, MFAF, ADRF and the NWDAF (hosting DCCF and/or ADRF) are listed in Table 6.2.6.0-1. Table 6.2.6.0-1: NF Services for the enhanced data collection procedures Service producer Service Reference NWDAF Nnwdaf_DataManagement 7.4 DCCF Ndccf_DataManagement Ndccf_ContextManagement 8.2 8.3 MFAF Nmfaf_3daDataManagement Nmfaf_3caDataManagement 9.2 9.3 ADRF Nadrf_DataManagement 10.2 DCCF, MFAF and NWDAF hosting DCCF shall use the same services listed in clause 6.2.1 from OAM and NFs (including AFs directly or via NEF) to collect data. Additionally, the new services for data exposure from DCCF, MFAF, ADRF and NWDAF (hosting DCCF and/or ADRF) as specified in clause 8, 9, 10 and 7.4 may also be used for data collection. The NWDAF, DCCF, ADRF shall obtain the proper information to perform data collection for a UE identified by a SUPI, a group of UEs identified by an Internal-Group-Id or any UE following the principles of clause 6.2.1. When NWDAF or DCCF need to discover the sources of data collection, they follow the principles defined in clause 6.2.2.1 in the case of data collection from NFs; in clause 5A.2 in the case of DCCF deployed in the network and requiring data from NWDAFs or ARDFs that independently collect data; clause 5.2 in the case of NWDAF (hosting DCCFs) and collecting data from other NWDAFs.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.6.1 Bulked Data Collection
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.6.1.0 General	NWDAFs may provide bulked data to consumers as an alternative to providing individual events (i.e. subscription to multiple event IDs to obtain the data required for an analytics generation). The bulked data is the set of data samples from the collected event notifications to be used for an Analytics ID for a consumer of NWDAF. A data sample may be a notification, in which case the bulked data may comprise a group of received notifications, or a data sample may be information extracted from a notification and processed, in which case the bulked data comprises the processed information. The bulked data can be used for the purpose of analytics inference or ML Model training. The bulked data is generated based on the set of data samples from event notifications collected from NFs/OAM and the properties for the selection of such data, Consumers of bulked data or operators may define different rules (e.g. aggregation formats or processes) for generation of bulked data for training or inference. NFs capable to expose bulked data have the following capabilities: - Exposing runtime collected data (e.g. data from NFs/AFs/OAM retrieved via notification mechanisms), or historical collected data (e.g. data from NFs/AFs/OAM that were at some point collected, then stored), or both; - Applying selection processes of data samples or processing mechanisms for the generation of the bulked data according to bulked data formatting and processing instructions provided by consumers of the bulked data or defined by an operator. The bulked data formatting and processing instructions may include the formatting and processing instructions as specified in clause 5A.4 and further instructions as described in clause 6.2.6.1.1. Such instructions define the allowed and/or restricted properties and/or processes to be applied to the set of data from the collected event notifications to be used for the bulked data, the properties and processes being: - The properties associated with the Bulked data formatting and processing are filters over the data to be associated with the bulked data. The properties for bulked data are defined as per Data Specification parameters in clause 6.2.6.1.1. - The processes associated with the bulked data formatting and processing are mechanisms applied to the data to be associated with the bulked data and are defined according to the Formatting and Processing parameter defined in clause 5A.4 and the further instructions defined in clause 6.2.6.1.1. These processes may comprise: definition if data to be used for composing the bulked data is directly extracted from collected events, or the data is extracted from event notification of the same event type and pre-processed, or both; applying anonymization of data fields in the bulked data to avoid exposing undesired information, aggregation levels (i.e. per cell, per UEs, or temporal, e.g. per hours or days). NOTE: Pre-process data from collected event notifications of the same event type refers to the usage of data manipulation processes in order to aggregate, concatenate, process data from multiple collected event notifications from the same event type that results in a single processed value. - Having the mapping of the Service Operation that have to be used for collecting data of the bulked data associated with an Analytics ID.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.6.1.1 Services for Bulked Data Collection	NWDAF may expose the Nnwdaf_DataManagement_Subscribe service operation with a request for bulked data including the following input parameters: - Data Specification: - Event ID(s) or Analytics ID(s); - In the case of Event IDs, the Data Specification fields includes the fields Target of Event Reporting and Event Filter Information as defined in clause 4.15.1 of TS 23.502 [3] and Bulked Data Type parameter, which can be set to ''raw data samples'' (i.e. data is directly extracted from collected events) or ''pre-processed data samples'' (i.e. data from collected events is processed and the processed data is included in the bulked data) or a combination of both; - If the Analytics ID(s), the Data Specification fields contain: - Target of Reporting including a tuple with Analytics ID; Bulked Data Type, which can be set to ''raw data samples'' (i.e. data is directly extracted from collected events) or ''pre-processed data samples'' (i.e. data from collected events is processed and the processed data is included in the bulked data) or a combination of both; - Filter Information may include fields related to the Analytics ID such as: Target of Analytics Information (e.g. any UE, list of UEs, groups of UEs); Analytics Filter Information (e.g. area of interest, DNN, Application, S-NSSAI). The Analytics ID also determines the Service Operation from NFs, OAM to be used and type of data (i.e. Event IDs, OAM measurements) to be collected and associated with the bulked data. - Service Operation in the case of Event ID, defines the service operation to be used by NWDAF, DCCF, MFAF, or ADRF to request data (e.g. Namf_EventExposure_Subscribe or OAM Subscribe) - Bulked Data Formatting and Processing: the parameters defined in clause 4.15.1 of TS 23.502 [3] for Event Reporting Information and Formatting and Processing instructions as defined in clause 5A.4. - A Notification Target Address (+ Notification Correlation ID), where the Notification Correlation ID is the unique identification for the bulked data being generated for the requesting consumer. - ADRF ID or NWDAF ID (or ADRF Set ID or NWDAF Set ID) storing historical data (optional). If known to the consumer, this may be specified to direct a DCCF or an NWDAF to the repository containing historical data. - (Optional) ADRF information indicating whether the collected data for the generation of the bulked data are to be stored in an ADRF and optionally an ADRF ID. - (Optional, in case the requested data is Event IDs) Data Source identification to collect the data, e.g. NF Instance (or NF Set) ID from which the data needs to be collected. The output parameter of the Nnwdaf_DataManagement_Subscribe service operation comprise the subscription correlation ID, which identifies the requested bulked data. The input parameters of Nnwdaf_DataManagement_Notify service operation shall contain the Notification Correlation ID and the generated bulked data when the fetch flag = false. When the fetch flag = true the notifications will contain the Notification Correlation ID, the Fetch Correlation ID and a target address where the generated bulked data may be retrieved. In the case of unsuccessful bulked data generation, the notification will contain an indication of an unsuccessful bulked data generation, optionally with expired bulked data deadline. The input parameters for the service operation Nnwdaf_DataManagement_Fetch include: the Notification correlation ID (+list of Fetch Correlation ID), which identifies the requested bulked data. The output parameters for the service operation Nnwdaf_DataManagement_Fetch include: - the generated bulked data. The generated bulked data exposed by the above listed service operations comprises: - the dataset (i.e. the resulting set of data samples and/or set of pre-processed data samples from the collected event notifications) generated based on the parameters of bulked data request and Bulked Data Formatting and Processing; - timestamp when the data sample is associated with a bulked data.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.6.2 Procedure for Data Collection from NWDAF	The procedure in Figure 6.2.6.2-1 is used by NWDAF service consumer to invoke the data management services at NWDAFs in order to retrieve runtime and historical data. Figure 6.2.6.2-1: Data Collection from NWDAF via Data Management Service 1. NWDAF service consumer (e.g. NWDAF, DCCF) identifies that further data from an NWDAF is required in order to perform some operation related to Analytics ID. The triggers for further data collection are related to: a) the local policies of NWDAF or DCCF (e.g. preparation for future requests for Analytics ID as specified in clause 6.2.2.1); b) a request for analytics generation requiring data not available or not directly reachable via the NWDAF service consumer (e.g. out of the serving area); c) a request for model training; d) a request for data collection that NWDAF service consumer cannot provide by itself. NOTE 1: If the NWDAF service consumer is a DCCF, the discovery of the proper NWDAF is defined in clause 6.2.6.3.6. If the NWDAF service consumer is a NWDAF, the NWDAF service consumer can discover the appropriate NWDAF(s) as defined in clause 5.2. 2a. NWDAF service consumer invokes Nnwdaf_DataManagement_Subscribe service from NWDAF to request a required data. The request comprises the Data Specification as well as Data Formatting and Processing instructions as defined in clause 5A.4, Notification Target Address (+ Notification Correlation ID). When the required data is Event IDs, the NWDAF service consumer may include the Data Source, e.g. NF Instance (or NF Set) ID from which the data needs to be collected. The NWDAF service consumer may include ADRF information indicating whether the data are to be stored in an ADRF and optionally an ADRF ID. The NWDAF service consumer may include ADRF ID or NWDAF ID (or ADRF Set ID or NWDAF Set ID) storing historical data (optional), directing NWDAF to the repository containing historical data. The NWDAF checks if required data is related to a user, i.e. SUPI or GPSI, then, depending on local policy and regulations, as described in clause 6.2.9, the NWDAF checks or has checked the user consent by retrieving the user consent information from UDM using Nudm_SDM_Get including data type "User consent". If user consent is not granted, NWDAF sends a response to the NWDAF service consumer in step 2b, indicating that user consent for data collection was not granted and the data collection for this SUPI or GPSI stops here. If the user consent is granted, the NWDAF can provide the required data to the NWDAF service consumer by performing the following steps 2b-7 and the NWDAF subscribes to UDM to notifications of changes on subscription data type "User consent" for this user using Nudm_SDM_Subscribe. When receiving the notification that user consent has been revoked, the NWDAF shall provide a Termination Request in Nnwdaf_DataManagement_Notify to request the NWDAF service consumer to cancel the subscription to the required data. 2b. Based on the received request, NWDAF creates a new data for the requesting consumer. NWDAF sends Nnwdaf_DataManagement_Subscribe service response with a confirmation of successful request and the subscription correlation ID identifying the requested data. NOTE 2: Subscription Correlation ID allows the NWDAF service consumer to request to NWDAF any changes in the generation of a requested data. 3. NWDAF determines whether the request data is available at such NWDAF. NWDAF maintains a local association of requested Event IDs or Analytics IDs to the list of triggered event subscription identifications from data sources to generate the requested data. Based on this local association, the NWDAF checks if the data to be collected is available at itself. If the data is available, NWDAF uses such data to generate the requested data. When data sources are NFs, the NWDAF discovers the proper NFs as defined in clause 6.2.2.1. When the data sources are other NWDAFs, the NWDAF discovers the other NWDAFs as defined in clause 5.2. When the data source is DCCF, the NWDAF discovers the proper DCCF as defined in clause 6.3.19 of TS 23.501 [2]. 4a. (Optional) If NWDAF receives a request for data that is not available or not reachable by such NWDAF (e.g. out of serving area), NWDAF determines the sources for the data that is not available, if the information has not been included in the subscription to the requested data. 4b. (Optional) NWDAF may trigger further data collection using any of the available mechanisms in clause 6.2.2 (e.g. if the data subscribed in step 2a partially matches data that are already being collected by the NWDAF from a data source and a modification of the subscription to the data source would satisfy both the existing data collection as well as the newly requested data) and clause 6.2.6 (e.g. recursively using data collection services from other needed NWDAFs, DCCFs, ADRFs, NFs). NWDAF updates its local association of the mapping of the requested data (Event ID or Analytics ID) to the identification of the request/subscription for data collection from the further data sources. 5. Based on the properties of the received request, NWDAF generates the requested data including the available or collected data (e.g. from other NWDAFs, DCCFs or ADRFs, NFs). 6a. If the fetch flag is set to true in step 2a, NWDAF waits until the requested data is ready and sends a Nnwdaf_DataManagement_Notify service message with fetch instructions. The requested data is ready when the NWDAF has generated the data and completed processing and formatting as described in clause 5A.4. 6b. If the fetch flag is set to false in step 2a, NWDAF uses the Nnwdaf_DataManagement_Notify service to send the Notification Correlation ID and requested data to the NWDAF service consumer. 7(a.b). Alternatively, if the Nnwdaf_DataManagement_Notify service message with fetch instructions is received in step 6a, the NWDAF service consumer shall fetch the required data from NWDAF via Nnwdaf_DataManagement_Fetch service operation within the fetch deadline specified in the fetch instructions. The NWDAF service consumer invokes the Nnwdaf_DataManagement_Fetch service operation with the input parameters including the Fetch Correlation ID, that identifies the data to be fetched and receives a response with the requested data. 8. The NWDAF service consumer uses the requested data for performing further processing. If the NWDAF service consumer is an NWDAF the requested data can be used for analytics generation or model training or for further exposing such data to other NWDAFs. If the NWDAF service consumer is a DCCF, the requested data can be provided to a DCCF data consumer. 9. When the NWDAF service consumer determines that no more data is required or if receiving a Termination Request from the NWDAF, e.g. due to user consent revocation for the data collection related to a user, it unsubscribes to the requested data from NWDAF. If NWDAF had triggered further data collection in Step 3a and 3b, NWDAF also unsubscribe to all data sources. NOTE 3: It is also possible that instead of providing the dataset of the generated data in steps 6a, 6b, 7b, the NWDAF provides a reference to where the dataset can be retrieved by the NWDAF service consumer.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.6.3 Data Collection using DCCF
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.6.3.1 General	This clause specifies procedures for data collection using the DCCF described in clause 5A for cases other than obtaining analytics from an NWDAF (which is specified in clause 6.1.2). Two options are supported: data delivered via the DCCF, according to clauses 6.2.6.3.2 and 6.2.6.3.3 and data delivered via a messaging framework according to clauses 6.2.6.3.4 and 6.2.6.3.5. Which option to be used is determined by DCCF configuration. Due to e.g. UE mobility, the source DCCF or MFAF may no longer be able to serve the UE. In such cases, the data consumer may select a new DCCF as described in clause 6.2.6.3.7, or source DCCF may execute the DCCF relocation procedure as described in clause 6.2.6.3.8. The data consumer may indicate in the subscription request whether the DCCF should execute the relocation procedure or send a notification to the data consumer as specified in the reselection procedure. If the data consumer does not send an indicator in the subscription request allowing DCCF relocation initiated by the DCCF the consumer may execute DCCF reselection based on internal logic. When data is collected for a group of UEs identified by an Internal-Group-Id or any UE, the procedure of DCCF relocation initiated by the DCCF is not applicable. In such case, the consumer may select one or more Target DCCFs and initiate a new subscription for the UEs.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.6.3.2 Data Collection via DCCF	The procedure depicted in Figure 6.2.6.3.2-1 is used by a data consumer (e.g. NWDAF) to obtain data and be notified of events via the DCCF using Ndccf_DataManagement_Subscribe service operation. Whether the data consumer directly contacts the Data Source or goes via the DCCF is based on configuration of the data consumer. Figure 6.2.6.3.2-1: Data Collection via DCCF 1. The data consumer subscribes to data via the DCCF by invoking the Ndccf_DataManagement_Subscribe (Service_Operation, Data Specification, Formatting Instructions, Processing Instructions, NF (or NF-Set) ID, ADRF Information) service operation as specified in clause 8.2.2. The data consumer may specify one or more notification endpoints. If data to be collected is subject to user consent: if the data consumer checked user consent, the data consumer shall provide user consent check information (i.e. an indication that it has checked user consent), otherwise, the data consumer shall provide a purpose for the data collection. Service_Operation is the service operation to be used by the DCCF to request data (e.g. Namf_EventExposure_Subscribe or OAM Subscribe). Data Specification provides Service Operation-specific parameters (e.g. event IDs, UE-ID(s), target of event reporting) used to retrieve the data. Formatting and Processing Instructions are as defined in clause 5A.4. The data consumer may include the Data Source, e.g. NF Instance (or NF Set) ID from which the data needs to be collected. The data consumer may include ADRF information indicating whether the data are to be stored in an ADRF and optionally an ADRF ID. 2. The DCCF checks if data is to be collected for a user, i.e. SUPI or GPSI, then, depending on local policy and regulations, the DCCF checks the user consent by retrieving the user consent information from UDM using Nudm_SDM_Get including data type "User consent" and taking into account purpose for data collection as provided in step 1. If user consent is not granted, DCCF does not subscribe to event exposure for events related to this user, the data collection for this SUPI or GPSI stops here and DCCF sends a response to the data consumer indicating that user consent for data collection was not granted. If the user consent is granted, the DCCF subscribes to UDM to notifications of changes on subscription data type "User consent" for this user using Nudm_SDM_Subscribe. If the data consumer is NWDAF and it provides user consent check information (i.e. an indication that it has checked user consent) in Ndccf_DataManagement_Subscribe in step 1, which has been obtained by the NWDAF from UDM before, then the DCCF can do data collection for a user based on the user consent information from the NWDAF and skip retrieving it from UDM. 3. The DCCF determine the NF type(s) and/or OAM to retrieve the data based on the Service Operation requested in step 1. If the NF instance or NF Set ID is not provided by the data consumer. the DCCF determines the NF instances that can provide data as described in clause 5A.2 and clause 6.2.2.2. If the consumer requested storage of data in an ADRF but the ADRF ID is not provided by the data consumer, or the collected data is to be stored in an ADRF according to configuration on the DCCF, the DCCF selects an ADRF to store the collected data. 4. The DCCF determines whether the data requested in step 1 are already being collected, as described in clause 5A.2. If the data requested are already being collected from the Data Source by a data consumer, the DCCF adds the data consumer to the list of data consumers that are subscribed for these data, then the DCCF determines that no subscriptions to the Data Source need to be created or modified. 5. If the data subscribed in step 1 partially matches data that are already being collected by the DCCF from a Data Source and a modification of this subscription to the Data Source would satisfy both the existing data subscriptions as well as the newly requested data, the DCCF invokes Nnf_EventExposure_Subscribe (Subscription Correlation ID) with parameters indicating how to modify the previous subscription (as specified in clause 5A.2). The DCCF adds the data consumer to the list of data consumers that are subscribed for these data. If some of the newly requested data can only be provided by new Data Source, the DCCF creates new subscription(s) to the new Data Source for the newly requested data. If the data requested at step 1 are not already available or not being collected yet, the DCCF subscribes to data from the NF using the Nnf_EventExposure_Subscribe service operation as specified in clause 5A.2 and clause 6.2.2.2, with DCCF indicated as Notification Target Address. The DCCF adds the data consumer to the list of data consumers that are subscribed for these data. 6. When new output data are available, the Data Source uses Nnf_EventExposure_Notify to send the data to the DCCF. 7. The DCCF uses Ndccf_DataManagement_Notify to send the data to all notification endpoints indicated in step 1. Data sent to notification endpoints may be processed and formatted by the DCCF so they conform to delivery requirements for each data consumer or notification endpoint as specified in clause 5A.4. The DCCF may store the information in ADRF if requested by the consumer or if required by DCCF configuration, using procedure as specified in clause 6.2B.3. NOTE: According to Formatting Instructions provided by the data consumer, multiple notifications from a Data Source can be combined in a single Ndccf_DataManagement_Notify so many notifications from the Data Source result in fewer notifications (or one notification) to the data consumer. Alternatively, a notification can instruct the data notification endpoint to fetch the data from the DCCF before an expiry time. 8a. If DCCF needs to retrieve data from OAM, procedure for data collection from OAM as per steps 1-4 from clause 6.2.3.2 is used. 8b. The DCCF uses Ndccf_DataManagement_Notify to send the data to all notification endpoints indicated in step 1. Data sent to notification endpoints may be processed and formatted by the DCCF, so they conform to delivery requirements for each data consumer or notification endpoint as specified in clause 5A.4. The DCCF may store the information in ADRF if requested by the consumer or if required by DCCF configuration, using procedure as specified in clause 6.2B.3. 9. If a Ndccf_DataManagement_Notify contains a fetch instruction, the notification endpoint sends a Ndccf_DataManagement_Fetch request to fetch the data from the DCCF. 10. The DCCF delivers the data to the notification endpoint 11. The UDM may notify the DCCF on changes of user consent at any time after step 2. If user consent is no longer granted for a user for which data has been collected and there are no other consumers for the data, the DCCF shall unsubscribe to any Event ID to collect data for that SUPI or GPSI. The DCCF shall further update or terminate affected subscriptions of the Data Consumer. The DCCF may unsubscribe to be notified of user consent updates from UDM for each SUPI for which user consent has been revoked. 12. When the data consumer no longer wants data to be collected it invokes Ndccf_DataManagement_Unsubscribe (Subscription Correlation ID), using the Subscription Correlation Id received in response to its subscription in step 1. The DCCF removes the data consumer from the list of data consumers that are subscribed for these data. 13. If there are no other data consumers subscribed to the data, the DCCF unsubscribes with the Data Source.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.6.3.3 Historical Data Collection via DCCF	The procedure depicted in figure 6.2.6.3.3-1 is used by data consumers (e.g. NWDAF) to obtain historical data, i.e. data related to past time period. The data consumer requests data using Ndccf_DataManagement_Subscribe service operation. Whether the data consumer uses this procedure or directly contacts the ADRF or NWDAF is based on configuration. Figure 6.2.6.3.3-1: Historical Data Collection via DCCF 1. The data consumer requests data via DCCF by invoking the Ndccf_DataManagement_Subscribe (Service_Operation, Data Specification, Time Window, Formatting Instructions, Processing Instructions, ADRF ID or NWDAF ID (or ADRF Set ID or NWDAF Set ID) service operation as specified in clause 8.2.2. The data consumer may specify one or more notification endpoints to receive the data. If data to be collected is subject to user consent: if the data consumer checked user consent, the data consumer shall provide user consent check information (i.e. an indication that it has checked user consent), otherwise, the data consumer shall provide a purpose for the data collection. "Service_Operation" is the service operation used to acquire the data from a data source. "Data Specification" provides Service_Operation-specific parameters (e.g. event IDs, UE-ID(s)) used to retrieve the data. "Time Window" specifies a past time period and comprises a start and stop time. "Formatting and Processing Instructions" are as defined in clause 5A4. The data consumer may optionally include the ADRF or NWDAF instance (or ADRF Set or NWDAF Set) ID where the stored data resides. 2. The DCCF checks if data is to be collected for a user, i.e. SUPI or GPSI, then, depending on local policy and regulations, the DCCF checks the user consent by retrieving the user consent information from UDM using Nudm_SDM_Get including data type "User consent" and taking into account purpose for data collection as provided in step 1. If user consent is not granted, DCCF does not subscribe to event exposure for events related to this user, the data collection for this SUPI or GPSI stops here and DCCF sends a response to the data consumer indicating that user consent for data collection was not granted. If the user consent is granted, the DCCF subscribes to UDM to notifications of changes on subscription data type "User consent" for this user using Nudm_SDM_Subscribe. If the data consumer is NWDAF and it provides user consent check information (i.e. an indication that it has checked user consent) in Ndccf_DataManagement_Subscribe in step 1, which has been obtained by the NWDAF from UDM before, then the DCCF can do data collection for a user based on the user consent information from the NWDAF and skip retrieving it from UDM. 3. If an ADRF or NWDAF instance or ADRF Set ID or NWDAF Set ID is not provided by the data consumer, the DCCF determines if any ADRF or NWDAF instances might provide the data as described in clause 5B and 5A.2. NOTE 1: An ADRF or NWDAF might have previously registered data it is collecting with the DCCF. 4. (conditional) If the DCCF determines that an ADRF instance might provide the data, or an ADRF instance or Set was supplied by the data consumer, the DCCF sends a request to the ADRF, using Nadrf_DataManagement_RetrievalSubscribe (Data Specification, Notification Target Address=DCCF) service operation as specified in clause 10.2. The ADRF responds to the DCCF with an Nadrf_DataManagement_RetrievalSubscribe response indicating if the ADRF can supply the data. If the data can be provided, the procedure continues with steps 6a and 6c. 5. (conditional) If the DCCF determines that an NWDAF instance might provide the data or an NWDAF instance or Set was supplied by the data consumer, the DCCF sends a request to the NWDAF using Nnwdaf_DataManagement_Subscribe (Data Specification, Notification Target Address=DCCF) as specified in clause 7.4.2. 6a-d. The ADRF uses Nadrf_DataManagement_RetrievalNotify or the NWDAF uses Nnwdaf_DataManagement_Notify to send the requested data (e.g. one or more stored notifications archived from a data source) to the DCCF. The data may be sent in one or more notification messages. 7a-b. The DCCF uses Ndccf_DataManagement_Notify to send data to all notification endpoints indicated in step 1. Notifications are sent to the Notification Target Address(es) using the data consumer Notification Correlation ID(s) received in step 1. Data sent to notification endpoints may be processed and formatted by the DCCF, so they conform to delivery requirements specified by the data consumer. NOTE 2: According to Formatting Instructions provided by the data consumer, multiple notifications from an ADRF or NWDAF can be combined in a single Ndccf_DataManagement_Notify so many notifications from the ADRF or NWDAF results in fewer notifications (or one notification) to the data consumer. Alternatively, a Ndccf_DataManagement_Notify can instruct the data notification endpoint to fetch the data from the DCCF before an expiry time. 8. If a notification contains a fetch instruction, the notification endpoint sends a Ndccf_DataManagement_Fetch request to fetch the data from the DCCF. 9. The DCCF delivers the data to the notification endpoint. 10. The UDM may notify the DCCF on changes of user consent at any time after step 2. If user consent is no longer granted for a user for which data has been collected and if there are no other consumers for the data, the DCCF shall unsubscribe to any data collection for that SUPI or GPSI. The DCCF shall further update or terminate affected subscriptions of the Data Consumer. The DCCF may unsubscribe to be notified of user consent updates from UDM for each SUPI for which user consent has been revoked. 11. When the data consumer no longer wants data to be collected or has received all the data it needs, it invokes Ndccf_DataManagement_Unsubscribe (Subscription Correlation ID) as specified in clause 8.2.3, using the Subscription Correlation Id received in response to its subscription in step 1. 12. If the data are being provided by an ADRF and there are no other data consumers subscribed to the data, the DCCF unsubscribes with the ADRF using Nadrf_DataManagement_RetrievalUnsubscribe as specified in clause 10.2.7. 13. If the data are being provided by an NWDAF and there are no other data consumers subscribed to the data, the DCCF unsubscribes with the NWDAF using Nnwdaf_DataManagement_Unsubscribe as specified in clause 7.4.3.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.6.3.4 Data Collection via Messaging Framework	This procedure depicted in Figure 6.2.6.3.4-1 is used by a data consumer (e.g. NWDAF) to obtain data and be notified of events using the DCCF and a Messaging Framework. The 3GPP DCCF Adaptor (3da) Data Management service and 3GPP Consumer Adaptor (3ca) Data Management service of the Messaging Framework Adaptor Function (MFAF) are used to interact with the 3GPP Network and the Messaging Framework. Whether the data consumer directly contacts the Data Source or goes via the DCCF is based on configuration. Figure 6.2.6.3.4-1: Data Collection via Messaging Framework 1. The data consumer subscribes to data via the DCCF by invoking the Ndccf_DataManagement_Subscribe (Service_Operation, Data Specification, Formatting Instructions, Processing Instructions, NF (or NF-Set) ID, ADRF Information, Data Consumer Notification Target Address (+ Notification Correlation ID)) service operation as specified in clause 8.2.2. The data consumer may specify one or more notification endpoints and the NF or NF set to collect data from. If data to be collected is subject to user consent: if the data consumer checked user consent, the data consumer shall provide user consent check information (i.e. an indication that it has checked user consent), otherwise, the data consumer shall provide a purpose for the data collection. Service_Operation is the service operation to be used by the DCCF to request data (e.g. Namf_EventExposure_Subscribe or OAM Subscribe). Data Specification provides Service_Operation-specific required parameters (e.g. event IDs, UE-ID(s), target of event reporting) and optional input parameters used to retrieve the data. Formatting and Processing Instructions are as defined in clause 5A.4. The data consumer may optionally include the Data Source NF Instance (or NF Set) ID. The data consumer may include ADRF information indicating whether the data are to be stored in an ADRF and, optionally, an ADRF ID. NOTE 1: Data consumer requesting data to be stored in ADRF allows the collected data to be available to other data consumers in the future. 2. The DCCF checks if data is to be collected for a user, i.e. SUPI or GPSI, then, depending on local policy and regulations, the NWDAF checks the user consent by retrieving the user consent information from UDM using Nudm_SDM_Get including data type "User consent" and taking into account purpose for data collection as provided in step 1. If user consent is not granted, NWDAF does not subscribe to event exposure for events related to this user, the data collection for this SUPI or GPSI stops here and DCCF sends a response to the data consumer indicating that user consent for data collection was not granted. If the user consent is granted, the DCCF subscribes to UDM to notifications of changes on subscription data type "User consent" for this user using Nudm_SDM_Subscribe. If the data consumer is NWDAF and it provides user consent check information (i.e. an indication that it has checked user consent) in Ndccf_DataManagement_Subscribe in step 1, which has been obtained by the NWDAF from UDM before, then the DCCF can do data collection for a user based on the user consent information from the NWDAF and skip retrieving it from UDM. 3. If the NF instance or NF Set ID is not provided by the data consumer, the DCCF determines the NF instances that can provide data as described in clause 5A.2 and clause 6.2.2.2. If the consumer requested storage of data in an ADRF, but the ADRF ID is not provided by the data consumer, or the collected data is to be stored in an ADRF according to configuration on the DCCF, the DCCF selects an ADRF to store the collected data. 4. The DCCF determines whether the data requested in step 1 are already being collected, as described in clause 5A.2. If the data requested are already being collected from the Data Source by a data consumer, the DCCF adds the data consumer to the list of data consumers that are subscribed for these data, then the DCCF determines that no subscriptions to the Data Source need to be created or modified. 5. The DCCF sends an Nmfaf_3daDataManagement_Configure (Data Consumer Information, MFAF Notification Information, Formatting Conditions, Processing Instructions) to configure the MFAF to map notifications received from the Data Source to outgoing notifications sent to endpoints and to instruct the MFAF how to format and process the outgoing notifications. The DCCF may also instruct the MFAF to store data into ADRF by providing an ADRF ID, if requested by the data consumer in step 1, together with the NF Id of the data source. Data Consumer Information contains for each notification endpoint, the data consumer Notification Target Address (+ Data Consumer Notification Correlation ID to be used by the MFAF when sending notifications in step 9. MFAF Notification Information is included if a Data Source is already sending the data to the MFAF. MFAF Notification Information identifies Event Notifications received from the Data Sources and comprises the MFAF Notification Target Address (+ MFAF Notification Correlation ID). If the MFAF does not receive MFAF Notification information from the DCCF, the MFAF selects a MFAF Notification Target Address (+ MFAF Notification Correlation ID) and sends the MFAF Notification Information, containing MFAF Notification Target Address (+ MFAF Notification Correlation ID), to the DCCF in the Nmfaf_3daDataManagement_Configure Response. 6. If the data subscribed in step 1 partially matches data that are already being collected by the DCCF from a Data Source and a modification of this subscription to the Data Source would satisfy both the existing data subscriptions as well as the newly requested data, the DCCF invokes Nnf_EventExposure_Subscribe (Subscription Correlation ID) with parameters indicating how to modify the previous subscription (as specified in clause 5A.2). The DCCF adds the data consumer to the list of data consumers that are subscribed for these data. If some of the newly requested data can only be provided by new Data Source, the DCCF creates new subscription(s) to the new Data Source for the newly requested data. If the data requested at step 1 are not already available or not being collected yet, the DCCF subscribes to data from the NF using the Nnf_EventExposure_Subscribe (Data Specification, MFAF Notification Target Address (+ MFAF Notification Correlation ID)) service operation as specified in clause 5A.2 and clause 6.2.2.2, using the MFAF Notification Target Address (+ MFAF Notification Correlation ID) received in step 5. The DCCF adds the data consumer to the list of data consumers that are subscribed for these data. 7. When new output data are available, the Data Source uses Nnf_EventExposure_Notify to send the data to the MFAF. The Notification includes the MFAF Notification Correlation ID. 8. The MFAF uses Nmfaf_3caDataManagement_Notify to send the data to all notification endpoints indicated in step 6. Notifications are sent to the Notification Target Address(es) using the Data Consumer Notification Correlation ID(s) received in step 6. Data sent to notification endpoints may be processed and formatted by the MFAF, so they conform to delivery requirements specified by the data consumer. The MFAF may store the information in ADRF if requested by consumer or if required by DCCF configuration NOTE 2: According to Formatting Instructions provided by the data consumer, multiple notifications from a Data Source can be combined in a single Nmfaf_3caDataManagement_Notify, so many notifications from the Data Source results in fewer notifications (or one notification) to the data consumer. Alternatively, a notification can instruct the data notification endpoint to fetch the data from the MFAF before an expiry time. 9. If a Nmfaf_3caDataManagement_Notify contains a fetch instruction, the notification endpoint sends a Nmfaf_3caDataManagement_Fetch request to fetch the data from the MFAF. 10. The MFAF delivers the data to the notification endpoint. 11. The UDM may notify the DCCF on changes of user consent at any time after step 2. If user consent is no longer granted for a user for which data has been collected and if there are no other consumers for the data, the DCCF shall unsubscribe to any Event ID to collect data for that SUPI or GPSI. The DCCF shall further update or terminate affected subscriptions of the Data Consumer. The DCCF may unsubscribe to be notified of user consent updates from UDM for each SUPI for which user consent has been revoked. 12. When the data consumer no longer wants data to be collected, it invokes Ndccf_DataManagement_Unsubscribe (Subscription Correlation ID), using the Subscription Correlation Id received in response to its subscription in step 1. The DCCF removes the data consumer from the list of data consumers that are subscribed for these data. 13. If there are no other data consumers subscribed to the data, the DCCF unsubscribes with the Data Source. 14. The DCCF de-configures the MFAF so it no longer maps notifications received from the Data Source to the notification endpoints configured in step 5.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.6.3.5 Historical Data Collection via Messaging Framework	The procedure depicted in figure 6.2.6.3.5-1 is used by data consumers (e.g. NWDAF) to obtain historical data, i.e. data related to past time period. The data consumer obtains data using Ndccf_DataManagement_Subscribe service operation as specified in clause 8.2.2, where the subscription results in one or more notifications depending on how the data is retrieved from the ADRF or NWDAF and how the data is formatted. Whether the data consumer uses this procedure or directly contacts the ADRF or NWDAF is based on configuration. Figure 6.2.6.3.5-1: Historical Data Collection via Messaging Framework 1. The data consumer requests data via DCCF by invoking the Ndccf_DataManagement_Subscribe (Service Operation, Data Specification, Time Window, Formatting Instructions, Processing Instructions, ADRF ID or NWDAF ID (or ADRF Set ID or NWDAF Set ID) service operation as specified in clause 8.2.2. The data consumer may specify one or more notification endpoints to receive the data. If data to be collected is subject to user consent: if the data consumer checked user consent, the data consumer shall provide user consent check information (i.e. an indication that it has checked user consent), otherwise, the data consumer shall provide a purpose for the data collection. Service_Operation is the service operation used to acquire the data from a data source, Data Specification provides Service_Operation-specific required parameters (e.g. event IDs, UE-ID(s) and optional input parameters used to retrieve the data. Time Window specifies a past time period and comprises a start and stop time and Formatting and Processing Instructions are as defined in clause 5A4. The data consumer may optionally include the ADRF or NWDAF instance (or ADRF Set or NWDAF Set) ID where the stored data resides. 2. The DCCF checks if data is to be collected for a user, i.e. SUPI or GPSI, then, depending on local policy and regulations, the DCCF checks the user consent by retrieving the user consent information from UDM using Nudm_SDM_Get including data type "User consent" and taking into account purpose for data collection as provided in step 1. If user consent is not granted, DCCF does not subscribe to event exposure for events related to this user, the data collection for this SUPI or GPSI stops here and DCCF sends a response to the data consumer indicating that user consent for data collection was not granted. If the user consent is granted, the DCCF subscribes to UDM to notifications of changes on subscription data type "User consent" for this user using Nudm_SDM_Subscribe. If the data consumer is NWDAF and it provides user consent check information (i.e. an indication that it has checked user consent) in Ndccf_DataManagement_Subscribe in step 1, which has been obtained by the NWDAF from UDM before, then the DCCF can do data collection for a user based on the user consent information from the NWDAF and skip retrieving it from UDM. 3. If an ADRF or NWDAF instance or ADRF Set ID or NWDAF Set ID is not provided by the data consumer, the DCCF determines if any ADRF or NWDAF instances might provide the data as described in clause 5B and 5A.2. NOTE 1: An ADRF or NWDAF might have previously registered data it is collecting with the DCCF. 4. The DCCF sends an Nmfaf_3daDataManagement_Configure (Data Consumer Information, Formatting Conditions, Processing Instructions) to configure the MFAF to map notifications received from the ADRF or NWDAF to outgoing notifications sent to endpoints and to instruct the MFAF how to format and process the outgoing notifications. "Data Consumer Information" contains for each notification endpoint, the data consumer Notification Target Address (+ Data Consumer Notification Correlation ID) to be used by the MFAF when sending notifications. The MFAF selects an MFAF Notification Target Address (+ MFAF Notification Correlation ID) and sends the MFAF Notification Information, containing MFAF Notification Target Address (+ MFAF Notification Correlation ID), to the DCCF in the Nmfaf_3daDataManagement_Configure Response. 5. (conditional) If the DCCF determines that an ADRF instance might provide the data, or an ADRF instance or Set was supplied by the data consumer, the DCCF sends a request to the ADRF, using Nadrf_DataManagement_RetrievalSubscribe (Data Specification, MFAF Notification Information) containing the MFAF Notification Target Address (+ MFAF Notification Correlation ID) received in step 4 as specified in clause 10.2. 6. The ADRF responds to the DCCF with an Nadrf_DataManagement_RetrievalSubscribe response indicating if the ADRF can supply the data. If the data can be provided, the procedure continues with step 9. 7. (conditional) If the DCCF determines that an NWDAF instance might provide the data, or an NWDAF instance or NWDAF Set was supplied by the data consumer, the DCCF sends a request to the NWDAF, using Nnwdaf_DataManagement_Subscribe (Data Specification, MFAF Notification Information) as specified in clause 7.4.2. MFAF Notification Information contains the MFAF Notification Target Address (+ MFAF Notification Correlation ID) received in step 4. 8. The NWDAF responds to the DCCF with an Nnwdaf_DataManagement_Subscribe response indicating if the NWDAF can supply the data. 9. The ADRF uses Nadrf_DataManagement_RetrievalNotify or the NWDAF uses Nnwdaf_DataManagement_Notify to send the requested data (e.g. one or more stored notifications archived from a data source) to the MFAF. The data may be sent in one or more notification messages. 10. The MFAF uses Nmfaf_3caDataManagement_Notify to send data to all notification endpoints indicated in step 4. Notifications are sent to the Notification Target Address(es) using the Data Consumer Notification Correlation ID(s) received in step 4. Data sent to notification endpoints may be processed and formatted by the MFAF, so they conform to delivery requirements specified by the data consumer. NOTE 2: According to Formatting Instructions provided by the data consumer, multiple notifications from an ADRF or NWDAF can be combined in a single Nmfaf_3caDataManagement_Notify so many notifications from the ADRF or NWDAF results in fewer notifications (or one notification) to the data consumer. Alternatively, a Nmfaf_3caDataManagement_Notify can instruct the data notification endpoint to fetch the data from the MFAF before an expiry time. 11. If a notification contains a fetch instruction, the notification endpoint sends a Nmfaf_3caDataManagement_Fetch request as specified in clause 9.3.3 to fetch the data from the MFAF. 12. The MFAF delivers the data to the notification endpoint. 13. The UDM may notify the DCCF on changes of user consent at any time after step 2. If user consent is no longer granted for a user for which data has been collected and if there are no other consumers for the data, the DCCF shall unsubscribe to any data collection for that SUPI or GPSI. The DCCF shall further update or terminate affected subscriptions of the Data Consumer. The DCCF may unsubscribe to be notified of user consent updates from UDM for each SUPI for which user consent has been revoked. 14. When the data consumer no longer wants data to be collected or has received all the data it needs, it invokes Ndccf_DataManagement_Unsubscribe (Subscription Correlation ID), using the Subscription Correlation Id received in response to its subscription in step 1. 15. If the data are being provided by an ADRF and there are no other data consumers subscribed to the data, the DCCF unsubscribes with the ADRF using Nadrf_DataManagement_RetrievalUnsubscribe as specified in clause 10.2.7. 16. If the data are being provided by an NWDAF and there are no other data consumers subscribed to the data, the DCCF unsubscribes with the NWDAF using Nnwdaf_DataManagement_Unsubscribe as specified in clause 7.4.3. 17. The DCCF de-configures the MFAF so it no longer maps notifications received from the ADRF or NWDAF to the notification endpoints configured in step 4.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.6.3.6 Data collection profile registration	In some cases data consumers (e.g. NWDAF or ADRF) collect data from data source NF directly, e.g. when NWDAF is co-located with 5GC NF. To enable data consumers can get the data which has been collected by NWDAF or ADRF directly (i.e. not via DCCF), the NWDAF or ADRF may register/update the data collection profile to the DCCF during/after the procedure of data collection. DCCF can then determine some requested data is available in NWDAF or ADRF and can coordinate data collection based on the data collection profile. The procedure depicted in Figure 6.2.6.3.6-1 is used by data source (e.g. NWDAF or ADRF) to register data profile to DCCF. Figure 6.2.6.3.6-1: Procedure for the NWDAF or ADRF register data profile to DCCF 1. An ADRF or NWDAF instance is collecting or has collected data directly, e.g. from collocated NF. 2. The ADRF or NWDAF requests to register/update data collection profile (Service Operation, Analytics/Data Specification, ADRF ID or NWDAF ID) to DCCF by invoking the Ndccf_ContextManagement_Register or Ndccf_ContextManagement_Update. The registration/ update request can be triggered by the acceptation of subscription for data collection responded by the data source (e.g. collocated NF), it can be before the start of data collection or after the completion of data collection. DCCF determines the data collection status of NWDAF or ADRF based on the Analytics/Data Specification, i.e. DCCF determines whether the required data is being collected or has been collected. "Service Operation" identifies the service used to collect the data or analytics from a Data Source (e.g.: Namf_EventExposure_Subscribe or Nnwdaf_AnalyticsSubscription_Subscribe). "Analytics/Data Specification" is the "Service Operation" specific parameters that identify the collected data (i.e.: Analytics ID(s) / Event ID (s), Target of Analytics Reporting or Target of Event Reporting, Analytics Filter or Event Filter, etc.). ADRF ID or NWDAF ID specify the ADRF or NWDAF which registers data collection profile. 3. The DCCF responds to the ADRF or NWDAF with a Ndccf_ContextManagement_Register Response or Ndccf_ContextManagement_Update Response. 4. To obtain historical data and if the data consumer is configured to collect data via the DCCF using Ndccf_DataManagement_Subscribe service operation, the data consumer uses the procedures described in clause 6.2.6.3.2 or clause 6.2.6.3.3. 5. The ADRF or NWDAF requests to delete a registration of data collection or analytics collection to the DCCF by invoking the Ndccf_ContextManagement_Deregister, triggered for instance by a request of the service consumer or by a storage life-time expiry of related data.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.6.3.7 DCCF (re-)selection initiated by consumer	The procedure depicted in Figure 6.2.6.3.7-1 is used by a data consumer (e.g. NWDAF or central DCCF) to obtain data related to UE(s), to be notified by the DCCF when the DCCF can no longer serve the UE(s) and to then reselect the DCCF. Figure 6.2.6.3.7-1: Procedure for DCCF relocation initiated by consumer 0. The data consumer subscribes to source DCCF. 1. Source DCCF may notify the data consumer that it cannot serve the subscription anymore, e.g. when location of UE(s) falls outside the serving area of the DCCF. A cause code is added with the notification (e.g. UE(s) moved outside DCCF serving area). The DCCF may send pending data to the data consumer. 2. The data consumer for the DCCF determines to select a new instance of DCCF. The data consumer discovers and selects the target DCCF as described in clause 6.3.19 of TS 23.501 [2]. The data consumer may perform the DCCF selection due to internal triggers, notification of a UE mobility event or by receiving the notification from the source DCCF in step 1. 3. The data consumer sends a subscription request to the target DCCF using Ndccf_DataManagement_Subscribe request. 4. The data consumer may unsubscribe from the source DCCF. 5. Target DCCF may subscribe to relevant data source(s), if not yet subscribed.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.6.3.8 DCCF and MFAF relocation initiated by DCCF	The procedure depicted in Figure 6.2.6.3.8-1 is used to support DCCF and MFAF reselection when the source DCCF or MFAF can no longer serve the UE. Figure 6.2.6.3.8-1: Procedure for DCCF relocation initiated by DCCF 0. The data consumer subscribes to source DCCF. The data consumer may indicate in the subscription request with an indicator that the DCCF may execute the relocation procedure. NOTE 1: If the source DCCF or target DCCF does not support relocation, the consumer may execute DCCF relocation based on internal logic. 1. Source DCCF subscribes UE mobility events from AMF. The UE ID is provided by the data consumer in step 0. 2. If UE moves out of the service area of the source DCCF, source DCCF determines UE DCCF subscription context to be transferred to target DCCF, e.g., triggered by a UE mobility event notification from AMF. If UE moves out of the service area of the source MFAF, source DCCF determines UE MFAF data subscription to be transferred to target MFAF. 3. Source DCCF may query the NRF to discover and select the target DCCF and/or MFAF, e.g. based on the UE location information received from AMF. Conditional on MFAF transfer: 4. Source DCCF uses the Nmfaf_3daDataManagement service to request transfer of UE MFAF data subscription context to the target MFAF. 5. Target MFAF retrieves MFAF subscription context from Source MFAF. 6. Target MFAF accepts the data subscription context transfer. The source MFAF stops data collection and can remove the related context. 7. Target MFAF responds to the DCCF indicating the transfer is complete. The response may contain a Transaction Reference ID from the Target MFAF. The DCCF regards the data collection context at the Source NWDAF as terminated. 8. If a DCCF relocation does not occur, Source DCCF updates the data subscription to the data source by changing the Notification Target Address and Notification Correlation ID to the MFAF Notification Address and the MFAF Notification Correlation ID that received in step 7. Conditional on DCCF transfer: 9. Source DCCF using Ndccf_DataManagement_Transfer Request service operation to transfer of UE data subscription context to the target DCCF. 10. Target DCCF accepts the data subscription(s) context transfer. 11. If an MFAF is being used, the Target DCCF uses the Nmfaf_3daDataManagement_Configure service to configure the target MFAF. 12. Target DCCF confirms UE data subscription context transfer to the source DCCF. The confirmation includes the Subscription Correlation ID used by the Target DCCF. 13. [Optional] Target DCCF subscribes to the relevant data source(s), if it is not yet subscribed to the data source(s) for the data required for the data subscription context and repeat step 1 to subscribes for UE mobility events from AMF. 14. Source DCCF informs the data consumer about the successful UE DCCF or MFAF data subscription context transfer using a Ndccf_DataManagement_Notify message. The notification may contain a Subscription Correlation ID provided by the target DCCF. 15. [Optional] Source DCCF unsubscribes with the data source(s) that are no longer needed for the remaining UE data subscriptions. For DCCF or MFAF transfer: NOTE 3: At this point, UE data subscription transfer is deemed completed. 16-18. Target DCCF or MFAF collects data from the data source(s) and notifies the data consumer using a Ndccf_DataManagement_Notify message.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.7 Data Collection with Event Muting Mechanism
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.7.1 General	Additional mechanisms to limit signalling between Event Producer NF (e.g. AMF, SMF) and Event Consumer NF (NWDAF, DCCF) are provided, with the Event Provider NFs enhanced with the optional capability of muting the notification of the events while storing for a limited time and limited size the events until the Event Consumer NF retrieves such mute stored events.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.7.2 Procedure for Data Collection with Event Muting Mechanism	The mute storage of events mechanism in the DCCF, the NWDAF, or NFs (if configured to support event muting mechanism) reuses the Event Reporting Information field of Event Exposure Framework to include the following flags: - Deactivate notification flag: The event consumer NF includes in the subscription to an event ID the deactivation flag to indicate to the event provider NF to collect, store the requested events but halt the notification to the consumer. The number of stored events may be limited based on Event Producer NF configuration; when this number is reached or another exception occurs (e.g., full buffer), the Event Producer NF performs the actions described later in this clause. - Retrieval notification flag: The event consumer NF includes in an event subscription modification request the subscription identification and the retrieval notification flag to indicate to the event producer NF to send the past collected events not already sent to this consumer NF. After sending the past collected events the event producer continues to store events without sending notifications to the event consumer. When an Event Consumer NF requests notification muting from an Event Producer NF using the Deactivation notification flag, the Event Consumer can in addition specify requested Event Producer NF actions to be taken when an exception occurs at the Event Producer NF. The Event Consumer NF may specify the following parameters in the Event Reporting Information field of Event Exposure Framework: - Event Producer NF action on buffered notifications: 'Send All', 'Discard All' or 'Drop Old'. - Event Producer NF action on subscription: 'close', 'continue with muting', 'continue without muting'. The Event Producer NF evaluates the requested action from the Event Consumer NF according to local policy (if configured) and in the response to the Event Consumer NF provides an accept indication if the request can be satisfied. If the request is accepted, the response from the Event Producer NF may indicate the following: - The maximum number of notifications that the Event Producer NF expects to be able to store. - An estimate of the duration for which notifications can be buffered. Using the event muting mechanism NWDAF, DCCF can subscribe to events from NFs such as AMF and SMF (if configured to support event muting mechanism), to avoid constant notifications and retrieve the mute stored events when it requires. The procedure in Figure 6.2.7.2-1 is used by Event Consumer NF to control the frequency of data collection from Event Producer NFs (except DCCF and NWDAF) via Event Exposure. For data collection via DCCF and NWDAF, the consumer may mute the notifications by using the formatting instructions as specified in clause 5A.4. Figure 6.2.7.2-1: Procedure for muting event notification 0a. The Event Consumer NF, such as NWDAF or DCCF, is configured with local policies that are used to determine when the muted storage of events is triggered and actions to be requested from the Event Producer NF if an exception occurs at the NF Producer (e.g., full buffer). The Event Producer NF is configured with local polices specifying default actions to take should an exception occur and policies for handling Event Consumers requests containing requested actions to be taken by the Event Producer NF if an exception occurs. 0b. The Event Consumer NF, such as NWDAF or DCCF, may receive a request with the Formatting and Processing parameters indicating Event Clubbing. The DCCF or NWDAF may utilize event muting when collecting data from NFs (if configured to support event muting mechanism). 1. The Event Consumer NF, DCCF or NWDAF subscribes for a (set of) Event ID(s) by invoking the Nnf_EventExposure_Subscribe service operation including in event reporting information the deactivate notification flag and optionally requested Event Producer NF actions to be taken when an exception occurs at the Event Producer NF. If the Event Producer NF supports the deactivate notification flag and the requested actions to be taken by the Event Producer NF when an exception occurs, the Event Producer NF sends a response back including the Subscription Correlation ID and an indication of successful deactivation of notifications. The Event Consumer NF may request the Event Producer NF to store data related to Event ID(s), or aggregated data related to UE(s). The response from the Event Producer NF may indicates the maximum number of notifications that the Event Producer NF expects to be able to store and / or an estimate of the duration for which notifications can be buffered. If the Event Producer NF does not support the deactivate notification flag or does not accept requested actions to be taken by the Event Producer NF when an exception occurs, the Event Producer NF sends a response back including an indication of failure and cause code. In this case, the Event Consumer NF re-sends the subscription request without including in the event reporting information the deactivate notification flag or with different requested actions to be taken by the Event Producer NF when an exception occurs. NOTE: If the Event Producer NF receives a subscription without the deactivate notification flag, the steps 2 - 6 are not executed and the Event Producer NF performs the event notification as defined in clause 4.15 of TS 23.502 [3]. 2. Based on the request from Event Consumer NF, DCCF or NWDAF, the Event Producer NF triggers a window of event collection for the Event Consumer NF, DCCF or NWDAF subscription with the indication of "deactivate notification flag". The Event Producer NF keeps the association between the Event ID, Subscription Correlation ID (which identifies the consumer of the event), subscriber information (e.g. notification target information) and the status of the transaction between the Event Consumer NF, DCCF or NWDAF and the Event Producer as "collecting events / non-notification". 3. Based on local policies or based on the Notification Time Window indicated in the Formatting and Processing parameters of the received request in step 0b, the Event Consumer NF, DCCF or NWDAF decides when to request the muted stored events from the Event Producer NF. 4. Event Consumer NF invokes the Nnf_EventExposure_Subscribe service operation from the Event Producer NF including, the Event ID, the Subscription Correlation ID and the retrieval notification flag. These parameters denote the identification of the transaction required by the Event Consumer NF, i.e. retrieve muted stored events for a subscribed Event ID and trigger a new time window of muted stored event generation without notification. 5. Event Producer NF based on the parameters received in the request from Event Consumer NF verifies whether there is a subscription to the requested Event ID with a deactivate notification flag. In positive case, Event Producer NF identifies and sends the past collected events muted during the period between the received retrieval notification flag and the last deactivate flag received from the Event consumer NF for the Event ID, the Subscription Correlation ID. If an exception (e.g. full buffer) occurs, the Event Producer NF performs the actions determined in steps 0a and 1. 6. The Event Producer NF checks whether overall event reporting information (e.g. the maximum time window for the subscription of such Event ID) has expired. If yes, it does not trigger another round of event muted storage and deactivates the subscription. If not expired, the Event Producer NF trigger another time window for muted stored of produced events, sets back the deactivated notification flag for the Event ID and Subscription Correlation ID. If the Event Consumer NF wants to change an existing subscription to an Event Producer NF using muted stored events into a regular notification of events, it shall invoke Nnf_EventExposure_Subscribe service operation from Event Producer NF without deactivate notification flag.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.8 Data Collection from the UE Application
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.8.1 General	The NWDAF may interact with an AF to collect data from UE Application(s) as an input for analytics generation and ML Model training. The AF can be in the MNO domain or an AF external to MNO domain. The data collection request from NWDAF may trigger the AF to collect data from the UE Application. The AF in this clause is referred as the Data Collection AF which is described in TS 26.531 [32]. The UE Application establishes a connection to the AF in the MNO domain or external to MNO domain over user plane via a PDU session. The AF communicates with the UE Application and collects data from UE Application. For both an AF in trusted domain and an AF in untrusted domain (which supports to collect data from a UE Application), the SLA between the operator and the Application Service Provider (i.e. ASP) determines per Application ID in use by the ASP: - The AF for the UE Application to connect to (e.g. based on an FQDN). - The information that the UE Application shares with the AF, subject to user consent. - Possible Data Anonymization, Aggregation or Normalization algorithms (if used). - The authentication information that enable the AF to verify the authenticity of the UE's Application that provides data. NOTE 1: The mutual authentication information that is used by the UE Application and the AF and how user consent is obtained is out o scope of the present document. However, the authentication information could contain GPSI. The AF (which supports the data collection) is configured based on the SLA above. NOTE 2: Data Anonymization, Aggregation or Normalization algorithms within the SLA are defined per individual UE. A UE Application (which supports to providing data to an AF) is configured by the ASP with the Application ID to use in the communication with the AF and then the UE Application is configured per Application ID with the following information: - The address of the AF to contact. - The parameters that the UE Application is authorized to provide to the AF. - The authentication information to enable the UE Application to verify the authenticity of the AF that requests data. NOTE 3: The authentication and authorization information that is used by the UE Application and the AF for collection and how user consent is obtained is out of SA2's scope. However, the authentication information could contain GPSI. NOTE 4: The configuration procedure for the above information from the ASP to the UE's Application is out of scope of the present document. NOTE 5: The Application ID configured in the UE Application can either be an OSAppId as defined in TS 23.503 [4] or an OS independent Application Identifier (e.g. for applications running on a web browser). The Target for Event Reporting in the Naf_EventExposure request may be set to: - an external UE ID (i.e. GPSI) or an external Group ID, in case the AF is located in the untrusted domain; - a SUPI or an internal Group ID, in case the AF is located within the trusted domain. The GPSI may be an External Identifier for individual UE as defined in TS 23.501 [2] that includes the domain name. This domain name and the Application ID configured in the UE Application are different from each other.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.8.2 Procedure for data collection from the UE Application
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.8.2.1 Connection establishment between UE Application and AF	The UE Application receives the data collection configuration from ASP. The configuration information is as described in clause 6.2.8.1. The UE Application establishes a user plane connection to the AF. Data collection procedure from the UE Application is performed via the user plane connection. NOTE 1: Whether multiple user plane connections are established, or a single user plane connection is established for different applications between each UE Application and AF is based on implementation that is out of 3GPP scope. NOTE 2: The Connection establishment procedure from the UE Application to the AF as above is out of scope of the present specification. For the 3GPP defined services, the Connection establishment procedure is described in TS 26.531 [32]. For the non-3GPP defined services, the Connection establishment procedure is out of 3GPP's scope. NOTE 3: In order to preserve resources (e.g. battery, quota) for the end user, a user plane connection to the AF can be established only when the UE has an active PDU Session for the UE Application and it is actively using the network (i.e. the user plane connection to the AF does not need to be established when the UE Application is inactive, or used in an off-line mode). Both direct data collection procedure (from the UE Application to the AF, either in trusted domain or untrusted domain) and indirect data collection procedure (from the UE Application to the Application server and from the Application server to the AF) are described in TS 26.531 [32]. The AF stores the IP address received from the UE (in the PDU session used) in order to request data collection from the UE Application. The UE IP address is used by the AF to identify the user plane connection. The UE Application provides the Application ID configured in the UE Application to the AF as described in TS 26.531 [32].
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.8.2.2 AF registration and discovery	The AF registers its available NF profile to the NRF. The AF in trusted domain registers to the NRF by using the Nnrf_NFManagement service that is defined in clause 5.2.7.2 of TS 23.502 [3]. The AF in untrusted domain registers the available NF profile to the NRF via the NEF as described in clause 6.2.2.3. The AF discovery and selection is described in clause 6.3.25 of TS 23.501 [2].
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.8.2.3 Data Collection Procedure from UE	Figure 6.2.8.2.3-1: Data Collection Procedure from UE 1. An NF subscribes to analytics from the NWDAF as described in clause 6.1.1.1, that includes Analytics ID, Analytics Filter Information including e.g. AoI, Internal Application ID(s) and Target of Analytics Reporting. NWDAF may also initiate the data collection prior to this subscription. Depending on local regulations and operator policy, the NWDAF checks the user consent for data collection in UDM. If the user consent is not given, the below steps are not performed. NOTE: Subscription to analytics can be triggered directly towards NWDAF or can be done via DCCF using procedure in clause 6.1.4.2. 2. NWDAF discovers and selects the AF that provides data collection (based on the AF profiles registered in NRF) as described in clause 6.3.25 of TS 23.501 [2]. Step 3a is used for the AF in trusted domain while step 3b is used for the AF in untrusted domain. 3a. NWDAF subscribes to the AF in trusted domain for UE data collection (i.e. input data from UE for analytics), by using Naf_EventExposure_Subscribe as defined in clause 5.2.19.2.2 of TS 23.502 [3]. The NWDAF request contains an Application ID known in the core network and the UE Application provides the Application ID configured in the UE Application. The AF binds the NWDAF request for an Application Id and the UE data collection for an Application Id configured in the UE. 3b. NWDAF subscribes to the AF in untrusted domain for UE data collection (i.e. input data from UE for analytics), by using step 2 and step 3 of the procedure that is described in Figure 6.2.2.3-1. NOTE: For steps 3a and 3b, data collection can also be triggered using DCCF, as specified in clause 6.2.6.3. 4. The AF collects the UE data using either direct or indirect data collection procedure in clause 6.2.8.2.1. The establishment of the connection can be performed at any time prior to this. The AF links the data collection request from step 3 to the user plane connection as described in clause 6.2.8.2.4. NOTE 1: The Direct data collection and indirect data collection procedure is described in TS 26.531 [32]. Step 5a is used for the AF in trusted domain while step 5b is used for the AF in untrusted domain. 5a. The AF in trusted domain receives the input data from the UE and processes the data (e.g. anonymizes, aggregates and normalizes) according to the SLA that is configured in the AF described in clause 6.2.8.1 and Event ID(s) and Event Filter(s) set during step 3a. The trusted AF then notifies the NWDAF on the processed data according to the NWDAF subscription in step 3a. 5b. The AF in untrusted domain receives the input data from the UE and processes the data (e.g. anonymizes, aggregates and normalizes) according to the SLA that is configured in the AF described in clause 6.2.8.1 and Event ID(s) and Event Filter(s) set during step 3b. The untrusted AF notifies the NWDAF on the processed data by using step 5b (i.e. Step 4 and step 5 of the procedure that described in Figure 6.2.2.3-1). NOTE 2: If NWDAF requests the same data from multiple UEs, i.e. a determined list of UEs or "any UE" as the Target of Analytics Reporting, the AF can process (e.g. anonymize, aggregate and normalize) the data from multiple UEs according to the Event ID(s) and Event Filter(s) received from NWDAF during step 3a or 3b before notifying the NWDAF on the processed data in step 5a (if the AF is in trusted domain) or step 5b (if the AF is in untrusted domain). 6. The NWDAF produces analytics using the UE data received from the AF. 7. The NWDAF provides analytics to the consumer NF. If the Target of Analytics Reporting that was received from the consumer in step 1 includes an Internal Group ID, NWDAF includes such Internal Group ID in step 3a or step 3b to AF. In the case of step 3b, NEF translates the Internal Group ID to an External Group ID. If the Target of Analytics Reporting that was received from consumer in step 1 is "any UE", NWDAF may either set the target of event reporting to "any UE" in step 3a or 3b to AF, or may determine a list of SUPIs from AMF and/or SMF based on the Analytics Filter Information and sends the step 3a or 3b to AF for the determined list of UEs. NOTE 3: It is assumed that the AF is provisioned with the list of UE IDs (GPSIs or SUPIs) belonging to an External or Internal Group ID.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.8.2.4 Correlation between UE data collection and the NWDAF data request	6.2.8.2.4.1 General The UE IP address is used to identify the user plane connection established between the UE application and the AF for data collection, while the AF receives the Naf_EventExposure_Subscribe to request for the specific UE data collection by using SUPI (for AF in trusted domain) or external UE ID (i.e. GPSI) (for AF in untrusted domain). AF is required to correlate the UE IP address to the SUPI or to GPSI. If the AF supports requests addressed to External Group ID (for AF in untrusted domain) or Internal Group ID (for AF in internal trust domain), the AF must correlate the list of external UE ID (i.e. GPSI) or SUPI, respectively, with the group(s) the UE belongs to, so that the AF can further correlate the UE ID (external or internal) to the UE IP address. AF may indicate in NF profile and register to NRF in clause 6.2.8.2.2 if it supports to do the mapping itself or ask NWDAF to do it. If the AF is in a trusted domain, it may also indicate the supported list of S-NSSAI, DNN combinations to NRF in NF profile. NOTE 1: If any method in Annex A is used, the AF always indicates in the NF profile that it supports mapping by itself. Accordingly, if AF supports the mapping, for AF in trusted domain, it is required to correlate the UE IP address and SUPI as described in clause 6.2.8.2.4.2 or by other means, e.g. as described in Annex A. This, after receiving the data collection request from NWDAF and there is no mapping information storage in the AF. For AF in untrusted domain, the procedure to correlate the UE IP address and GPSI is described in clause 6.2.8.2.4.4 or by some other means, e.g. as described in Annex A. If there is NAT between the UE and the AF, one of the methods in Annex A or method in clause 6.2.8.2.4.5 can be used. If there is no NAT between the UE and the AF, NWDAF may collect the mapping information as described in clause 6.2.8.2.4.4 before sending request to AF in step 3a or step 3b in Figure 6.2.8.2.3-1. If the user plane session between the UE and the AF is released, the AF / NWDAF removes the stored correlation information between UE IP address and UE SUPI / GPSI. For all procedures defined in this clause 6.2.8.2.4.1, a specific combination of S-NSSAI/DNN shall be corresponding to a single PDU session for a UE to access the AF (either in trusted domain or untrusted domain). NOTE 2: Based on implementation, for the UE to access the Data Collection AF, only a single PDU Session is allowed to be established to the Data Collection AF, by configuring a specific S-NSSAI/DNN for the Data Collection AF only. 6.2.8.2.4.2 AF in trusted domain correlates UE data collection and NWDAF request This is only valid if there is no NAT between the UE and the AF. If the AF receives the Naf_EventExposure_Subscribe/Request including Target for Event Reporting set to SUPI and not including the UE's IP address and the AF does not locally store the UE's IP address, the AF finds the PDU session(s) serving the SUPI, DNN, S-NSSAI from UDM and the allocated IPv4 address or IPv6 prefix or both from SMF as described in Figure 6.2.8.2.4.2-1. Figure 6.2.8.2.4.2-1: AF in trusted domain correlates UE data collection and NWDAF request 0. At the establishment of the user plane connection between the UE Application and the AF, the AF stores the UE IP address (for both direct and indirect reporting) as described clause 6.2.8.2.1. 1. The AF receives a request to retrieve input data as described in clause 6.2.8.2.3 including a SUPI. The AF finds the SMF serving the PDU session(s) for this SUPI using Nudm_UECM_Get_Request including SUPI, type of requested information set to SMF Registration Info and the S-NSSAI and DNN, as defined in clause 5.3.2.5.7 in TS 29.503 [26]. 2. The UDM provides the SMF id and the corresponding PDU Session id, S-NSSAI, DNN using Nudm_UECM_Get_Response to the AF. Using the AF supported S-NSSAI, DNN and the received information from UDM, AF determines the PDU session used for the user plane connection between UE and AF. 3. The AF sends Nsmf_EventExposure_Subscribe to the SMF identified in step 2, including the Target for Event Reporting set to the PDU Session id(s) provided in step 2 and the Event ID set to IP address/prefix allocation/change. 4. The SMF provides the allocated IPv4 address or IPv6 prefix to the AF. 5. The AF correlates the UE data that includes the UE IP address and the NWDAF request for a SUPI using the retrieved IPv4 address or IP v6 prefix. If the user plane session between the UE and the AF is released, the AF shall remove the stored correlation information between the UE IP address / prefix and SUPI. 6.2.8.2.4.3 AF in untrusted domain correlates UE data collection and NWDAF request This is only valid if there is no NAT between the UE and the AF. If the AF receives the Naf_EventExposure_Subscribe from NWDAF, via NEF, including Target for Event Reporting set to GPSI and not including the UE's IP address and the AF does not locally store the UE's IP address, the AF request the NEF to provide the allocated IPv4 address or IPv6 prefix or both as described in Figure 6.2.8.2.4.3-1. NOTE 1: The NWDAF can also provide the UE IP address to the AF as described in clause 6.2.8.2.4.1. Figure 6.2.8.2.4.3-1: AF in untrusted domain correlates UE data collection and NWDAF request 0. Same step as step 0 in figure 6.2.8.2.4.3-1. 1. The AF receives a request to retrieve input data as described in clause 6.2.8.2.3 including a GPSI. The AF requests NEF to provide the IPv4address or IPv6 prefix or both serving the PDU session for this GPSI towards the AF using Nnef_UEAddress_Get_Request. 2. The NEF is configured with the DNN, S-NSSAI to access this AF. The NEF finds the SMF serving the PDU session(s) for this GPSI, DNN, S-NSSAI using Nudm_UECM_Get_Request including type of requested information set to SMF Registration Info and the S-NSSAI and DNN, as defined in clause 5.3.2.5.7 of TS 29.503 [26]. NOTE 2: If there are more than one (DNN, S-NSSAI) combination to access this AF, the NEF will find the SMF(s) serving the PDU session(s) to any of these (DNN, S-NSSAI) combinations. 3. The UDM provides the SMF id(s) and the tuple (PDU Session id (S-NSSAI, DNN) using Nudm_UECM_Get_Response to the NEF. Using the configuration in NEF, as described in step 2, the NEF determines the PDU session used for the user plane connection between UE and AF. 4. The NEF sends Nsmf_EventExposure_Subscribe to the SMF(s) identified in step 3, including the Target for Event Reporting set to the PDU Session id(s) provided in step 3 and the Event ID set to IP address/prefix allocation/change. 5. The SMF provides the allocated IPv4 address or IPv6 prefix or both to the NEF. 6. The NEF provides the allocated IPv4 address or IPv6 prefix or both provided by SMF in step 5 to the AF. 7. The AF correlates the UE data that includes the UE IP address and the NWDAF request for the GPSI using the retrieved IPv4 address or IP v6 prefix. If the user plane session between the UE and the AF is released, the AF shall remove the stored correlation information between the UE IP address / prefix and GPSI. 6.2.8.2.4.4 NWDAF correlates UE data collection and NWDAF request for trusted AF and untrusted AF This is only valid if there is no NAT between the UE and the AF. NWDAF receives the analytics subscription from consumer and discover an AF as described in clause 6.2.8.2.3. NWDAF finds the PDU session(s) serving the SUPI, DNN, S-NSSAI from UDM and the allocated IPv4 address or IPv6 prefix from SMF as described in Figure 6.2.8.2.4.4-1. Figure 6.2.8.2.4.4-1: NWDAF correlates UE data collection and NWDAF request 1. The NWDAF finds the SMF(s) serving the PDU session(s) for this SUPI or GPSI using Nudm_UECM_Get_Request including SUPI or GPSI, type of requested information set to SMF Registration Info and the list of S-NSSAI and DNN combinations, as defined in clause 5.3.2.5.7 in TS 29.503 [26]. The NWDAF acquires the DNN, S-NSSAI used to access the AF using Nnrf_NFDiscovery_Request service operation or is configured with the DNN, S-NSSAI used to access the AF. 2. The UDM provides the SMF id(s) and the corresponding PDU Session id(s), per S-NSSAI, DNN combination using Nudm_UECM_Get_Response to the NWDAF. Based on the S-NSSAI, DNN used to access the AF in step 1, NWDAF determines the PDU session used for the user plane connection between UE and AF. 3. The NWDAF sends Nsmf_EventExposure_Subscribe to the SMF identified in step 2, including the Target for Event Reporting set to the PDU Session id(s) provided in step 2 and the Event ID set to IP address/prefix allocation/change. 4. The SMF provides the allocated IPv4 address or IPv6 prefix to the NWDAF. 5. Step 3a for AF in trusted domain or step 3b for AF in untrusted domain in Figure 6.2.8.2.3-1 is performed with the exception that NWDAF sets the allocated IPv4 address or IPv6 prefix that were received in step 4 as target of event reporting. If NWDAF subscribed for the PDU session used for the user plane connection between the UE and the AF is released notification in step 3, the SMF informs the NWDAF that the UE IP address / prefix is released via Nsmf_EventExposure_Notify. Based on this information, the NWDAF shall remove the stored correlation information between the UE IP address / prefix and SUPI. 6.2.8.2.4.5 AF correlates UE data collection and NWDAF request when there is NAT between UE and AF This is only valid if there is NAT between the UE and the AF, and the AF has retrieved the UE (private) IP address assigned by 5GC for the PDU session, either from the SMF, or it was provided from the NWDAF in order to request data collection from the UE Application. Figure 6.2.8.2.4.5-1: AF correlates UE data collection and NWDAF request when there is NAT between UE and AF 0. The AF gets or retrieves the UE private IP address as described in clauses 6.2.8.2.4.2, 6.2.8.2.4.3, and 6.2.8.2.4.4. For untrusted AF, the following steps 1-7 apply. For trusted AF, the following steps 3-6 apply and NEF is replaced by the AF. 1. The AF requests NEF to provide the UE public IP address towards the AF using Nnef_UEAddress_Subscribe (UE (private) IP address, IP address of remote end) indicating immediate reporting. 2. The NEF authorizes the AF request. If the authorisation is not granted, the NEF replies to the AF indicating authorisation failure and the following steps will be skipped; otherwise, the NEF proceeds with the following steps. 3. The NEF sends the Nnrf_NFDiscovery_Request to obtain the address of the UPF implementing NAT information exposure functionality for the UE (private) IP address. The request includes the UE private IP address. The NEF may also include the DNN and S-NSSAI associated with the AF ID, as well as the IP domain. 4. The NRF responds with a Nnrf_NFDiscovery_Response including the UPF address of the UPF implementing NAT information exposure functionality for the UE (private) IP address. 5. The NEF uses the Nupf_EventExposure_Subscribe service operation to request UE public address and port information from the UPF. The request includes the UE (private) IP address and an IP address of the remote end and indicates immediate reporting. 6. The UPF sends the UE public IP address and source TCP/UDP port immediately by invoking the Nupf_EventExposure_Notify to the NEF. 7. The NEF provides the UE public IP address and source TCP/UDP port provided by UPF in step 6 to the AF using Nnef_UEAddress_Notify. 8. The AF correlates the UE data by using the retrieved UE public IP address and port information. NOTE: When new information (e.g. UE public IP address and source TCP/UDP port) corresponding to the UE (private) IP address and IP address of the remote end provided in step 5, is available, steps 6-8 are performed again. 6.2.8.2.4a Void
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.9 User consent for analytics	Depending on local policy or regulations, to protect the privacy of user data, the data collection, ML Model training and analytics generation for a SUPI or GPSI, Internal or External_Group_Id or "any UE" may be subject to user consent bound to a purpose, such as analytics or ML Model training. The user consent is subscription information stored in the UDM, which includes: a) whether the user authorizes the collection and usage of its data for a particular purpose; b) the purpose for data collection, e.g. analytics or model training. The NWDAF retrieves the user consent to data collection and usage from UDM for a user, i.e. SUPI prior to collecting user data from an NF as described in clause 6.2.2 and from a DCCF as described in clause 6.2.6 and prior to retrieving data from NWDAF (either directly according to clause 6.2.6.2 or via DCCF according to clause 6.2.6.3) or ADRF (either directly or via DCCF according to clause 6.2.6.3). In roaming scenario, the H-RE-NWDAF is the enforcement point to check user consent. The H-RE-NWDAF retrieves the roaming-related user consent for a user from the UDM. NOTE 1: The content of the roaming-related user consent is specified in clause X.7 and Annex V of TS 33.501 [49]. If a request for analytics is for "any UE", meaning that the consumer requests analytics for all UEs registered in an area, such as a S-NSSAI or DNN or AoI, then the NWDAF resolves "any UE" into a list of SUPIs using the Namf_EventExposure service with EventId "Number of UEs served by the AMF and located in an area of interest" and retrieves user consent for each SUPI. If a request for analytics is for an Internal or External Group Id, NWDAF resolves it into a list of SUPIs and retrieves user consent for each SUPI. If user consent for a user is granted, then the NWDAF subscribes to user consent updates in UDM using Nudm_SDM_Subscribe service operation. Otherwise, the NWDAF excludes the corresponding SUPI from the request to collect data and generate analytics or ML Model on the other users for which user consent is granted if the request is for a group of UEs identified by an Internal-Group-Id or "any UE". When data is collected from the UE Application, the ASP is responsible to obtain user consent to share data with the MNO. If the UDM notifies that the user consent changed, then the NWDAF checks if the user consent is not granted for the purpose of analytics or model training. If user consent was revoked for a UE, the NWDAF stops data collection for that UE. For analytics subscriptions to UE related analytics with the Target of Analytics Reporting set to that UE, the NWDAF stops generation of new analytics and stops providing affected analytics to consumers. For ML Model subscriptions with Target of ML Model Reporting set to that UE, the NWDAF containing MTLF stops (re-)training of ML Model(s) using data from the UE and stops providing the ML Model(s) to consumers (NWDAF containing AnLF) for analytics. If the Target of Analytics Reporting or Target of ML Model Reporting is either an Internal or External Group Id or a list of SUPIs or "any UE", the NWDAF skips those SUPIs that do not grant user consent for the purpose of analytics or model training. The NWDAF may unsubscribe to be notified of user consent updates from UDM for users for which data consent has been revoked. NOTE 2: The NWDAF can provide analytics or ML Model to consumers that request analytics or ML Model for an Internal or External Group Id, or for "any UE", skipping those users for which consent is not granted or is revoked. The Analytics ID that needs to check user consent before collecting input data are those that collect input data per user, i.e. per SUPI, GPSI, Internal or External Group Id, or those with the Target of Analytics Reporting or Target of ML Model Reporting set to a SUPI, GPSI or External or Internal Group Id and are described in clause 6.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.10 Data collection by H-RE-NWDAF from V-RE-NWDAF for outbound roaming users	This procedure may be used by the RE-NWDAF in the HPLMN as service consumer to subscribe/unsubscribe to notifications about input data from the VPLMN for outbound roaming users (from the HPLMN perspective). The H-RE-NWDAF and V-RE-NWDAF in the procedure are NWDAFs with roaming exchange capability. Figure 6.2.10-1: data collection by H-RE-NWDAF from V-RE-NWDAF for outbound roaming users 1. For subscription to collected data related to the UE(s), the H-RE-NWDAF checks the user consent of related users depending on local policy or regulations. NOTE 1: See clause X.7 and Annex V TS 33.501 [49] for details of the user consent check procedures. See clause X.8 of TS 33.501 [49] for protection of data exchange in roaming case. 2. The H-RE-NWDAF of HPLMN discovers the V-RE-NWDAF of VPLMN that supports the Nnwdaf_RoamingData service using the NRF as specified in Clause 5.2. NOTE 2: The access to the Nnf_EventExposure services is expected to be restricted to NF service consumers within the same PLMN to prevent bypassing checks based on user consent and operator policy. 3. The H-RE-NWDAF subscribes/unsubscribes to notifications about input data by invoking the Nnwdaf_RoamingData_Subscribe / Nnwdaf_RoamingData_Unsubscribe service operation. It optionally may indicate the IDs of AMFs and for local breakout also SMFs in the VPLMN handling related UEs, as obtained from the UDM. 4. The V-RE-NWDAF checks if the HPLMN is authorised to subscribe to the input data based on VPLMN operator polices (that may depend on the HPLMN and may indicate permissible or restricted input data and related parameters). If the HPLMN is not authorized to subscribe to the input data, the subscription request must be rejected with a proper cause in the response to the H-RE-NWDAF and the following steps are skipped. 5. The V-RE-NWDAF may trigger new data collection from NF(s) (as indicated via the AMF ID(s) or SMF ID(s)) if needed and monitors the requested input data using procedures as described in clauses 6.2.1 to 6.2.8. 6. The V-RE-NWDAF may restrict the exposed input data based on VPLMN operator polices (that may depend on the HPLMN) and may store them for auditing. 7. The V-RE-NWDAF responds to or notifies the H-RE-NWDAF with the subscribed available input data.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.11 Data collection by V-RE-NWDAF from H-RE-NWDAF for inbound roaming users	This procedure may be used by an RE-NWDAF in the VPLMN to subscribe/unsubscribe to notifications about input data from the HPLMN for inbound roaming users (from the VPLMN perspective). H-RE-NWDAF and V-RE-NWDAF in the procedure are NWDAFs with roaming exchange capability. Figure 6.2.11-1: Data Collection by V-RE-NWDAF from H-RE-NWDAF for inbound roaming users 1. The V-RE-NWDAF of VPLMN discovers an H-RE-NWDAF of the HPLMN that supports the Nnwdaf_RoamingData service using the NRF as specified in clause 5.2. NOTE 1: The access to the Nnf_EventExposure services is expected to be restricted to NF service consumers within the same PLMN to prevent bypassing checks based on user consent and operator policy 2. The V-RE-NWDAF subscribes/unsubscribes to input data information by invoking Nnwdaf_RoamingData_Subscribe / Nnwdaf_RoamingData_Unsubscribe service operation. 3. The H-RE-NWDAF checks if the VPLMN is authorised to subscribe to the indicated input data based on the HPLMN operator polices (that may depend on the VPLMN and may indicate permissible or restricted input data and related parameters) and user consent of related users. If the VPLMN is not authorized to subscribe to the input data, the subscription request must be rejected with a proper cause in the response to the V-RE-NWDAF and the following steps are skipped. NOTE 2: See clause X.7 and Annex V of TS 33.501 [49] for details of the user consent check procedures. See clause X.8 of TS 33.501 [49] for protection of data exchange in roaming case. 4. The H-RE-NWDAF may trigger new data collection if needed and monitors the requested input data, using procedures as described in clauses 6.2.1 to 6.2.8. 5. The H-RE-NWDAF may restrict the exposed input data based on HPLMN operator polices (that may depend on the VPLMN) and may store them for auditing. 6. The H-RE-NWDAF responds to or notifies the V-RE-NWDAF with the subscribed available input data.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.12 Data Collection using LCS
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.12.1 General	The NWDAF may collect location information for a target UE or a group of target UEs using LCS. The collected location related information can include: ‐ Location estimate of the UE in geographical coordinates and/or local coordinates expressed as a shape as defined in TS 23.032 [34] or local coordinate reference system; - Time stamp of location estimate; - Velocity of the UE as defined in TS 23.032 [34]; - Information about the positioning method used to obtain the location estimate of the UE; - Indication of area event, when UE enters, is within or leaves the Geographical area; - Indication of motion event when UE moves by more than some predefined straight line distance from a previous location. NOTE: The location information that can be retrieved is defined within the location service response in clause 5.5 of TS 23.273 [39]. NWDAF shall use Ngmlc service as defined in TS 23.273 [39] to collect the location information using LCS. Only Mobile Terminated Location Request (MT-LR) is supported, including both Immediate Location Request and Deferred Location Request. NWDAF may determine to query LCS system instead of AMF to obtain UE's location information based on the following attributes as received from NWDAF consumer: - Analytics ID (e.g. UE Mobility, QoS Sustainability, Relative Proximity, Movement Behaviour); - Preferred granularity of location information.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.12.2 Procedure for data collection using LCS	The interactions between NWDAF and LCS for data collection are illustrated in Figure 6.2.12.2-1. The data collected depends on the use cases. This figure is an abstraction of how NWDAF collects location information using LCS. The actual procedures that NWDAF may use are as follows: - For a target UE, both 5GC-MT-LR procedure for the commercial location service as specified in clause 6.1.2 and deferred 5GC-MT-LR procedure as specified in clause 6.3 of TS 23.273 [39] can be utilized; - For a group of target UEs, bulk operation of LCS Service Request Targeting to Multiple UEs as specified in clause 6.8 of TS 23.273 [39] can be utilized. Figure 6.2.12.2-1: Data collection using LCS 1. NWDAF requests the location information from GMLC about a target UE (that may be identified by a SUPI) or a group of target UEs (identified by a group ID). 2. GMLC interacts within LCS, i.e. with AMF/LMF as described in TS 23.273 [39], to obtain the UE's location information. If privacy verification is required, GLMC will interact with UE via AMF before sending the location information to NWDAF. 3.1 If it is Immediate Location Request, GMLC sends the location service response including the location information for the target UE (or a group of target UEs) within a short time period as specified in clause 4.1a.4 of TS 23.273 [39] to the NWDAF. 3.2 If it is Deferred Location Request, GMLC sends the location service response including the indication of event occurrence and location information if requested for the target UE (or group of target UEs) at some future time (or times) as specified in clause 4.1a.5 of TS 23.273 [39] to the NWDAF.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.13 Rating untrusted AF data sources
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.13.1 General	When using an untrusted AF as data source, NWDAF may consider the data source rating results. The rating of untrusted AF is based on the quality of data collected. Such rating may be triggered when the accuracy check based on the calculation between the predicted and ground-truth data indicates low performance, while the untrusted data source rating may be performed based on NWDAF internal logic. In the selection of the appropriate data sources, the NWDAF may use the rate of untrusted AF data sources as a criterion to calculate the expected confidence degree.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.13.2 Procedure for rating untrusted AF data sources	The process of rating untrusted AF data sources is depicted in Figure 6.2.13.2. For realizing potential issues, the NWDAF containing AnLF subscribes to the NWDAF containing MTLF, which performs an accuracy calculation based on the predicted and ground-truth data or alternatively the NWDAF containing AnLF can calculate the accuracy locally by comparing the predicted and ground-truth data. Figure 6.2.13.2: NWDAF containing AnLF-based untrusted AF data source rating 1. NWDAF containing AnLF subscribes to NWDAF containing MTLF for obtaining an ML Model using the Nnwdaf_ModelProvision_Subscribe service operation. The NWDAF containing AnLF may include a threshold (as described in clause 6.2E.2) to indicate when the NWDAF containing MTLF needs to execute the accuracy monitoring operations. Option 1: Accuracy report from NWDAF containing MTLF 2a. NWDAF containing MTLF evaluates the ML Model Accuracy according to clause 6.2E.2. 2b. An accuracy report is sent to the NWDAF containing AnLF, e.g. when the reporting threshold is met by invoking Nnwdaf_MLModelProvision_Notify service operation. Option 2: NWDAF containing AnLF computes accuracy 2c. NWDAF containing AnLF calculates the accuracy by comparing the predictions with ground truth data. 2d. NWDAF containing AnLF is aware that the ML Model used has a low accuracy either by receiving the accuracy report in step 2b or monitoring the accuracy by itself in step 2c. NWDAF containing AnLF determines that it needs to check further the data sources and compute data source rating. The decision conditions upon which it needs to initiate data source rating for a data source is based on NWDAF containing AnLF implementation. 3a-3b. NWDAF containing AnLF initiates rating of a data source by requesting and receiving supplementary data, i.e. via Nnwdaf_DataManagement_Fetch / Ndccf_DataManagement_Notify, from different data sources (if available) to verify the data source quality or correctness. Such data can be for example performance data from the OAM which are supplementary to the data from untrusted AFs, or data from UPF supplementary to the data from untrusted AFs. 4. NWDAF containing AnLF updates the rating for the sources where untrusted data is deviated from the supplementary trusted data per Event ID. NOTE 1: An NWDAF containing AnLF determines the rating of an untrusted AF data source based on internal operations. 5a. NWDAF containing AnLF stores the untrusted AF data source rating locally. 5b. NWDAF containing AnLF may send the untrusted AF data source rating to UDSF, if available. NWDAF containing AnLF uses the Nudsf_UnstructuredDataManagement_Create service operation. NOTE 2: To avoid an untrusted AF to be permanently excluded as a data source, the NWDAF containing AnLF can re-rate the untrusted AF based on its internal logic. For example, it can rate the untrusted AF after some timer expired. 6. A NWDAF consumer subscribes to a certain Analytics ID, using Nnwdaf_AnalyticsSubscription_Subscribe service operation. Either step 7a or step 7b is executed, before collecting the data needed for the subscribed Analytic ID. 7a. The NWDAF containing AnLF retrieves the untrusted AF rating of the data sources locally. 7b. The NWDAF containing AnLF retrieves the untrusted AF rating of the data sources from the UDSF using the to use Nudsf_UnstructuredDataManagement_Query service operation. 8. If the rating of one or more untrusted AF is below a threshold (i.e. that is pre-configured), then the NWDAF containing AnLF can: (i) select an alternative untrusted AF (if available) with higher rating; or (ii) request supplementary data from other trusted data sources. 9. The NWDAF containing AnLF subscribes to a new data source to receive alternative or supplementary data if a new data source is selected in step 8. 10. The NWDAF containing AnLF may use the rate of untrusted AF data sources as a criterion to calculate the confidence level of the respective analytics output. 11. The NWDAF containing AnLF provides the analytics output to the analytics consumer, using the Nnwdaf_AnalyticsSubscription_Notify service operation. In the case of ML Model (re)training, if the NWDAF containing MTLF is the same NWDAF containing AnLF in step 5b, it may also use the rate of untrusted AF data sources by performing steps 7b and 8-9 and then, (re)trains the ML Model.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.14 Analytics Collection from MDAF
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.14.1 General	The MDA functional overview and service framework as defined in Figure 5.1-1 of TS 28.104 [45] is used by NWDAF to trigger the MDA MnS to request analytics from the MDA Management Function (MDAF). Before NWDAF requests analytics from the MDA Management Function, the NWDAF discovers the MDA Management Function via the MnS discovery service producer as defined in clause 5 of TS 28.537 [46].
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.2.14.2 Procedure for analytics collection from MDAF	Figure 6.2.14.2-1: Procedure for analytics collection from MDAF Precondition: Initially MDAF(s) or MDA MnS producers register their capabilities, i.e. MnS information or MnS profile as described in clause 5 of TS 28.537 [46] to a MnS discovery service producer. The MnS discovery service producer may contain all or partial information related to the capabilities of MDA MnS producer. 1. An analytics consumer issues a request or subscription of network analytics towards the selected NWDAF as described in clause 6.1. 2. NWDAF discovers the MDA Management Function from the MnS discovery service producer by sending a MnS producer discovery service operation. NOTE 1: The service operation can possibly include parameters for MDA MnS discovery, e.g. requested MDA Type, Area of Interest, Network Slice information, etc. The detailed parameters are defined in TS 28.537 [46] and TS 28.622 [41]. 3. The MnS discovery service producer provides the relevant MnS information of the MDA Management Function to the NWDAF. NOTE 2: MnS information refer to the data describing a MnS producer and their capabilities, which is used by the consumer to discover the producers of specific Management Services and to derive the addresses of the Management Services as defined in TS 28.537 [46] and TS 28.622 [41]. 4. If the MnS information of more than one MDAF is received, NWDAF selects the most suitable MDAF and gets the address of the selected MDAF from the MnS information. Then NWDAF sends analytics request to the MDA Management Function by triggering a MDARequest service operation as defined in clause 9.3.2 of TS 28.104 [45] requesting management data analytics such as Slice Coverage Analysis, Mobility Performance Analysis, Fault Prediction analysis. The analyticScope may contain Area of interest, Network Slice information, NF type etc. NOTE 3: Definitions and details of the parameters for MDARequest service operation can be found in TS 28.104 [45]. NOTE 4: The selection of the most suitable MDAF can be based on MnS information of the MDA MnS, e.g. MDA type as one MDA capability, and the request/subscription received from analytics consumer in step 1. The detailed information and procedure between the NWDAF and the MDAF for MDA MnS selection is defined in TS 28.537 [46], TS 28.622 [41] and TS 28.104 [45]. 5. The MDA Management Function provides the analytics to the NWDAF by triggering an MDAReporting service operation. 6. NWDAF provides the analytics response back to the analytics consumer after processing the analytics provided by the MDA Management Function together with other data receives from NF sources according to the Analytics ID defined in clause 6. 6.2A Procedure for ML Model Provisioning 6.2A.0 General This clause presents the procedure for the ML Model provisioning. An NWDAF containing AnLF may be locally configured with (a set of) IDs of NWDAFs containing MTLF and the Analytics ID(s) supported by each NWDAF containing MTLF to retrieve trained ML Models or may use the NWDAF discovery procedure specified in clause 5.2 for discovering NWDAFs containing MTLF. An NWDAF containing MTLF may determine that further training for an existing ML Model is needed when it receives the ML Model subscription or the ML Model request. A NWDAF containing MTLF may retrieve trained ML Models from other NWDAF containing MTLF as described in clause 5.3. The NWDAF containing MTLF determines to train a ML Model either based on the request from NWDAF containing AnLF, or based on local configuration. The NWDAF containing MTLF further determines the FL procedure is required but it can not act as an FL server, therefore, the NWDAF containing MTLF should discover an FL server NWDAF as described in clause 5.2 and request the ML Model provisioning from the FL server NWDAF. How to protect the trained ML Model file from being used without authorization e.g. being forwarded by an NWDAF containing MTLF that has retrieved a model from another NWDAF containing MTLF is defined in Annex X, clause X.10 of TS 33.501  [49]. 6.2A.1 ML Model Subscribe/Unsubscribe The procedure in Figure 6.2A.1-1 is used by an NWDAF service consumer, i.e.: - an NWDAF containing AnLF to subscribe/unsubscribe at another NWDAF, i.e. an NWDAF containing MTLF; - an NWDAF containing MTLF to subscribe/unsubscribe at another NWDAF containing MTLF, i.e. an FL server NWDAF; or - an LMF to subscribe/unsubscribe at an NWDAF containing MTLF, to be notified when ML Model Information related to the ML Model subscription becomes available, using Nnwdaf_MLModelProvision services as defined in clause 7.5. When the service consumer is an NWDAF containing AnLF, the ML Model Information is used by the NWDAF containing AnLF to derive analytics. When the service consumer is an LMF, how the ML Model Information is used by the LMF is described in TS 23.273 [39]. The service is also used by an NWDAF Service Consumer to modify existing ML Model Subscription(s). An NWDAF can be at the same time a consumer of this service provided by other NWDAF(s) and a provider of this service to other NWDAF(s). Figure 6.2A.1-1: ML Model for analytics subscribe/unsubscribe 1. The NWDAF service consumer (e.g. NWDAF containing AnLF, NWDAF containing MTLF) subscribes to, modifies, or cancels subscription for a (set of) trained ML Model(s) associated with a/an (set of) Analytics ID(s) by invoking the Nnwdaf_MLModelProvision_Subscribe / Nnwdaf_MLModelProvision_Unsubscribe service operation. If the NWDAF service consumer is LMF, it includes an indication that a model for LMF-based AI/ML Positioning is requested. NOTE 1: How to implement the indication of requesting a model for LMF-based AI/ML Positioning in a backward compatible and extensible manner is up to stage 3. The parameters that can be provided by the NWDAF service consumer are listed in clause 6.2A.2. The service consumer optionally indicates its support for multiple ML Models if available. When a ML Model subscription is received, the NWDAF containing MTLF may: - determine whether existing trained ML Model(s) can be used for the subscription; or - determine whether triggering further training for the existing trained ML Models is needed for the subscription. If the NWDAF containing MTLF determines that further training is needed, this NWDAF may initiate data collection from 5GC NFs (e.g. AMF/DCCF/ADRF), UE Application (via AF) or OAM as described in clause 6.2, to generate the ML Model. For ML model training for LMF-based AI/ML positioning, the NWDAF containing MTLF collects input data from the LMF as described in TS 23.273 [39]. If the service invocation is for a subscription modification or subscription cancelation, the NWDAF service consumer includes an identifier (Subscription Correlation ID) to be modified in the invocation of Nnwdaf_MLModelProvision_Subscribe. 2. If the NWDAF service consumer subscribes to a (set of) trained ML Model(s) associated to a (set of) Analytics ID(s), the NWDAF containing MTLF notifies the NWDAF service consumer by invoking Nnwdaf_MLModelProvision_Notify service operation with: - For each Analytics ID requested by the service consumer, a set of pair(s) of unique ML Model identifier and the associated Information. NOTE 2: The structure and format of the ML Model identifier and its uniqueness are up to stage 3. NOTE 3: Parameters defined for Multiple ML Models are for Analytics accuracy enhancement. The content of trained ML Model Information that can be provided by the NWDAF containing MTLF is specified in clause 6.2A.2. The NWDAF containing MTLF also invokes the Nnwdaf_MLModelProvision_Notify service operation to notify an available re-trained ML Model when the NWDAF containing MTLF determines that the previously provided trained ML Model required re-training at step 1. When step 1 is for a subscription modification (i.e. including Subscription Correlation ID), the NWDAF containing MTLF may provide either a new trained ML Model different to the previously provided one, or a re-trained ML Model by invoking Nnwdaf_MLModelProvision_Notify service operation. When the content includes ML Model provide indicator, the NWDAF service consumer (i.e. an NWDAF containing AnLF, an NWDAF containing MTLF or LMF) may trigger ML Model retrieval procedure in clause 6.2B.7 to retrieve ML Model(s) from ADRF. 6.2A.2 Contents of ML Model Provisioning The consumers of the ML Model provisioning services (e.g. NWDAF containing AnLF, NWDAF containing MTLF) as described in clause 7.5 and clause 7.6 may provide the input parameters as listed below: - A list of Analytics IDs: identifies the analytics for which the ML Model is used. - [OPTIONAL] Vendor ID: identifies the vendor of the consumer (e.g. of NWDAF containing AnLF, NWDAF containing MTLF). For each Analytics ID, the following parameters may be provided: - [OPTIONAL] Use case context: indicates the context of use of the analytics to select the most relevant ML Model. NOTE 1: The NWDAF containing MTLF can use the parameter "Use case context" to select the most relevant ML Model, when several ML Models are available for the requested Analytics ID(s). The values of this parameter are not standardized. - [OPTIONAL] ML Model Interoperability Information. This is vendor-specific information that conveys, e.g., requested model file format, model execution environment, etc. The encoding, format, and value of ML Model Interoperable Information is not specified since it is vendor specific information, and is agreed between vendors, if necessary for sharing purposes. - [OPTIONAL] ML Model Filter Information: indicates the applicable conditions of the trained ML model and enables the consumer to select which ML Model for the analytics is requested, e.g. S-NSSAI, Area of Interest. Parameter types in the ML Model Filter Information are the same as parameter types in the Analytics Filter Information which are defined in procedures. - [OPTIONAL] Target of ML Model Reporting: indicates the object(s) for which ML Model is requested, e.g. specific UEs i.e. a list of SUPIs, a group of UEs i.e. a list of Internal-Group-Ids or any UE (i.e. all UEs). - [OPTIONAL] Requested representative ratio: a minimum percentage of UE(s) in the group whose data is a non-empty set and can be used in the model training when the Target of ML Model Reporting is a group of UEs i.e. a list of Internal-Group-Ids. - [OPTIONAL] ML Model Target Period: indicates time interval [start, end] for which ML Model for the Analytics is requested. The time interval is expressed with actual start time and actual end time (e.g. via UTC time). - [OPTIONAL] Inference Input Data information: contains information about various settings that are expected to be used by the consumer (e.g. NWDAF containing AnLF) during inferences such as: - the "Input Data" that are expected be used, each of them optionally accompanied by metrics that show the granularity with which this data will be used (i.e., a sampling ratio, the maximum number of input values, and/or a maximum time interval between the samples of this input data). NOTE 2: This can be a subset of the possible Input Data specified for a certain analytics type. - the data sources that are expected to be used, indicated as a list of NF instance (or NF set) identifiers. - [OPTIONAL] Number of ML model(s), indicating the maximum number of ML models that the NWDAF containing MTLF could provide to the consumer (e.g. NWDAF containing AnLF) for the Analytics ID. - ML Model Reporting Information with the following parameters: - (Only for Nnwdaf_MLModelProvision_Subscribe) ML Model Reporting Information Parameters as per Event Reporting Information Parameter defined in Table 4.15.1-1, TS 23.502 [3]. - A Notification Target Address (+ Notification Correlation ID) as defined in clause 4.15.1 of TS 23.502 [3], allowing to correlate notifications received from the NWDAF containing MTLF with this subscription. - [OPTIONAL] Indication of supporting multiple ML Models. - [OPTIONAL] Accuracy level(s) of Interest. NOTE 3: Multiple ML Models Filter Information are composed by Accuracy level(s) of Interest and Number of ML Model(s). - [OPTIONAL] Time when model is needed: indicates the latest time when the consumer expects to receive the ML Model(s). - [OPTIONAL] ML Model Monitoring Information: - desired ML Model metric. NOTE 4: In this Release, only "ML Model Accuracy" is defined as ML Model metric. - [OPTIONAL] (only for Nnwdaf_MLModelProvision_Subscribe service operation) ML Model monitoring reporting mode: such as Accuracy reporting interval or pre-determined status. Depending on the reporting mode, the NWDAF containing MTLF reports the ML Model accuracy to NWDAF containing AnLF either periodically or when the ML Model accuracy is crossing an ML Model Accuracy threshold, i.e. the accuracy either becomes higher or lower than the ML Model Accuracy threshold. - [OPTIONAL] ML Model Accuracy Threshold: indicating the accuracy threshold of the ML Model requested by the consumer (as a kind of pre-determined status). It also can be used as an indication that the MTLF is triggered to execute the accuracy monitoring operations for the ML Model provisioned to AnLF. - [OPTIONAL] DataSetTag and ADRF ID if available: indicates the inference data (including input data, prediction and the ground truth data at the time which the prediction refers to) stored in ADRF which can be used by MTLF to retrain or reprovision of the ML Model. - [OPTIONAL] ML Model identifier: indicates the Model that the data corresponding to the DataSetTag is related to (in the case of subscription modification). The NWDAF containing MTLF provides to the consumer of the ML Model provisioning service operations as described in clause 7.5 and 7.6, the output information as listed below: - (Only for Nnwdaf_MLModelProvision_Notify) The Notification Correlation Information. - For each Analytics ID requested by the service consumer, a set of pair (s) of unique ML Model identifier and the following information. - ML Model Information, which includes: - the ML Model file address; or - ADRF (Set) ID. When ADRF (Set) ID is provisioned and the MTLF authorizes the NF Service Consumer to retrieve all ML Models corresponding to a Storage Transaction ID, a Storage Transaction ID may also be provisioned. - [OPTIONAL] ML Model provider information, includes the NF Instance Identifer to identify the NWDAF containing MTLF which is the ML Model provider and is responsible for training/updating this ML Model. NOTE 5: For example, if one MTLF provides the Model which is generated by FL server NWDAF, the NF Instance Identifier identifies the FL server NWDAF but not the MTLF. - [OPTIONAL] ML Model Filter Information and/or Target of ML Model Reporting may be also provided. They are provided when the ML Model provisioning request include same Analytics IDs but with different Targets of ML Model Reporting and/or ML Model Filter Information. - [OPTIONAL] ML Model provide indicator: indicates that the ML Model corresponding to the ML Model identifier is updated (e.g. re-trained ML Model). - [OPTIONAL] ML Model degradation indicator: indicates whether the provided ML Model is degraded. - [OPTIONAL] Validity period: indicates time period when the provided ML Model Information applies. - [OPTIONAL] Spatial validity: indicates Area where the provided ML Model Information applies. NOTE 6: Spatial validity and Validity period are determined by MTLF internal logic and it is a subset of AoI if provided in ML Model Filter Information and of ML Model Target Period, respectively. - [OPTIONAL] ML Model representative ratio: indicating the percentage of UE(s) in the group whose data is used in the ML Model training when the Target of ML Model Reporting is a group of UE(s). - [OPTIONAL] Training Input Data Information: contains information about various settings that have been used by MTLF during ML model training, such as: - the "Input Data" that have been used, each of them optionally accompanied by metrics that show the data characteristics and granularity with which this data has been used (i.e. a sampling ratio, the maximum number of input values and/or a maximum time interval between the samples of this input data, data range including maximum and minimum values, mean and standard deviation and data distribution when applicable) and the time, i.e. timestamp and duration, when this data was obtained. - the data sources related to the "Input Data" that were used for ML Model training, which have been identified by a list of NF instance (or NF set) identifiers. NOTE 7: This can be a subset of the possible Input Data specified for a certain analytics type. NOTE 8: Data source information enables ML Model selection when different models are available for an Analytics ID, or it enables a consumer to avoid selecting a ML Model that used data from a specific data source at a particular time or used data characterized by specific data characteristics. - [OPTIONAL] ML Model Accuracy Information: indicates the accuracy of the ML Model if related ML Model Monitoring Information was provided, which includes: - the metric value of the ML Model. - [OPTIONAL] used ML Model metric. To calculate the UE location, the LMF as the consumer can also use the ML Model provisioning services as described in clause 7.5 and clause 7.6 to retrieve ML model from NWDAF containing MTLF.The ML Model retrieving procedure by the LMF from NWDAF containing MTLF and the applicability of the parameters of the contents of ML Model provisioning are as defined in TS 27.273 [39]. 6.2A.3 ML Model request The procedure in Figure 6.2A.3-1 is used by an NWDAF service consumer, i.e.: - an NWDAF containing AnLF to request and get ML Model Information from another NWDAF, i.e. an NWDAF containing MTLF; - an NWDAF containing MTLF to request and get ML Model Information from another NWDAF containing MTLF working as FL server; or - an LMF to request and get ML Model Information from NWDAF containing MTLF, using Nnwdaf_MLModelInfo services as defined in clause 7.6. When the service consumer is an NWDAF containing AnLF, the ML Model is used by the NWDAF containing AnLF to derive analytics. When the service consumer is an LMF, the ML Model Information is used by the LMF as described in TS 23.273 [39]. An NWDAF can be at the same time a consumer of this service provided by other NWDAF(s) and a provider of this service to other NWDAF(s). Figure 6.2A.3-1: ML Model Request 1. The NWDAF service consumer (e.g. NWDAF containing AnLF, NWDAF containing MTLF) requests a (set of) ML Model(s) associated with a/an (set of) Analytics ID(s) by invoking Nnwdaf_MLModelInfo_Request service operation. If the NWDAF service consumer is LMF, it includes an indication that a model for LMF-based AI/ML Positioning is requested. The parameters that can be provided by the NWDAF Service Consumer are listed in clause 6.2A.2. The service consumer optionally indicates its support for multiple ML Models if available. Whe an ML Model Information request is received, the NWDAF containing MTLF may: - determine whether existing trained ML Model(s) can be used for the request; or - determine whether triggering further training for the existing trained ML Models is needed for the request. If the NWDAF containing MTLF determines that further training is needed, this NWDAF may initiate data collection from 5GC NFs (e.g. AMF/DCCF/ADRF), UE Application (via AF) or OAM as described in clause 6.2, to generate the ML Model. For ML model training for LMF-based AI/ML positioning, the NWDAF containing MTLF collects input data from the LMF as described in TS 23.273 [39]. 2. The NWDAF containing MTLF responds to the NWDAF service consumer by invoking Nnwdaf_MLModelInfo_Request response service operation including: - a set of pair(s) of unique ML Model identifier, the ML Model Information for each Analytics ID that the NWDAF service consumer requests. The content of ML Model Information that can be provided by the NWDAF containing MTLF is specified in clause 6.2A.2. 6.2B Analytics Data and ML Model Repository procedures 6.2B.1 General Collected data and analytics may be stored in ADRF, using procedure as specified in clause 6.2B.2 and clause 6.2B.3. Collected data and analytics may be deleted from ADRF, using procedure as specified in clause 6.2B.4. ML Model may be stored in ADRF, using procedure as specified in clause 6.2B.5. ML Model may be deleted from ADRF, using procedure as specified in clause 6.2B.6. ML Model(s) may be retrieved from ADRF, using procedure as specified in clause 6.2B.7. 6.2B.2 Historical Data and Analytics storage The procedure depicted in figure 6.2B.2-1 is used by consumers (e.g. NWDAF, DCCF or MFAF) to store historical data and/or analytics, i.e. data and/or analytics related to past time period that has been obtained by the consumer. After the consumer obtains data and/or analytics, consumer may store historical data and/or analytics in an ADRF. Whether the consumer directly contacts the ADRF or goes via the DCCF or via the Messaging Framework is based on configuration. The consumer may include in the storage request the DataSetTag attribute which the data records are to be associated to when stored by ADRF. The DataSetTag attribute is defined in Table 6.2B-1. Data records can be associated to multiple DataSetTag attributes. Table 6.2B-1: DataSetTag attribute Information Description DataSetTag Identifies the data set. DataSetDescription Provides human-readable information about the characteristics of the data set. The consumer may include in the storage request the Data Synthesis and Compression (DSC) information. The detail of DSC information is up to implementation, which is out of 3GPP scope. NOTE: DSC information can include the following information: - indication that the data have been generated using a data synthesis tool; - indication that the data have been generated using a data compression tool; - the information about the data synthesis and/or compression technique. Figure 6.2B.2-1: Historical Data and Analytics storage 0a-c. NWDAF, DCCF or ADRF are configured with default operator storage policies as described in clause 5B.1. 1. The consumer sends data and/or analytics to the ADRF by invoking the Nadrf_DataManagement_StorageRequest (collected data with timestamp, analytics with timestamp, Service Operation, Analytics Specification or Data Specification, Storage Handling information, optionally DataSetTag, optionally DSC information) service operation. The NWDAF or DCCF may provide notification endpoint information to the ADRF for use by the ADRF to send notifications (implicit subscription) alerting the DCCF or NWDAF that data are about to be deleted (see step 6). 2a-c. Based on Storage Handling information (if available) and Storage Policy, the ADRF, DCCF or NWDAF determines the Storage Approach (lifetime for storing data and whether consumer is notified prior to data deletion). 3. The ADRF stores the data and/or analytics sent by the consumer. The ADRF may, based on implementation, determine whether the same data and/or analytics is already stored or being stored based on the information sent in step 1 by the consumer NF and, if the data and/or analytics is already stored or being stored in the ADRF, the ADRF decides to not store again the data and/or analytics sent by the consumer. If the DataSetTag attribute is included for data and/or analytics already stored or being stored, then ADRF associates the data records with such DataSetTag. 4. The ADRF sends Nadrf_DataManagement_StorageRequest Response message to the consumer indicating that data and/or analytics is stored, whether the ADRF determined at step 3 that data or analytics is already stored and the Storage Approach. Conditional on ADRF Managing the Storage Approach 5. The ADRF determines that the lifetime of the stored data or analytics has expired (according to the Storage Approach). 6. If indicated by the Storage Approach, the ADRF sends a notification alerting the DCCF or NWDAF that data are about to be deleted. NOTE: This is an implicit subscription. 7. The DCCF or NWDAF indicate in the response to the ADRF whether they will retrieve the Data or Analytics. 8 The DCCF or NWDAF may retrieve the Data or Analytics from the ADRF as described in clause 5B.1. Conditional on the DCCF or NWDAF Managing the Storage Approach 9 The NWDAF or DCCF determine that the lifetime of the stored data or analytics has expired (according to the Storage Approach). 10. The NWDAF or DCCF optionally retrieve the data or analytics from the ADRF as described in clause 5B.1. 11. The NWDAF or DCCF request that the data or analytics be deleted from the ADRF. 12. The ADRF deletes the data or analytics if: - the Storage Approach in the ADRF indicates alerting the consumer is not required prior to data or analytics deletion; - in step 7 the DCCF or NWDAF indicated data or analytics will not be retrieved prior to deletion; - data or analytics retrieval in step 8 has completed; or - A request to delete data or analytics was received in step 11. If the ADRF received a response from the NWDAF or DCCF in step 7 indicating data or analytics will be retrieved but retrieval is not initiated before an adequate fixed time has elapsed, the ADRF may autonomously delete the data or analytics. 6.2B.3 Historical Data and Analytics Storage via Notifications The procedure depicted in figure 6.2B.3-1 is used by consumers (NWDAF, DCCF) to store received notifications in the ADRF. The consumer requests the ADRF to initiate a subscription for data and/or analytics. Data and/or analytics provided in notifications as a result of the subsequent subscription by the ADRF are stored in the ADRF. The consumer may include in the subscription request the DataSetTag attribute defined in Table 6.2B-1. Figure 6.2B.3-1: Historical Data and Analytics Storage via Notifications 0a-c NWDAF, DCCF or ADRF are configured with default operator storage policies as described in clause 5B.1. 1a-d. Based on provisioning or based on reception of a DataManagement subscription request (e.g. see clause 6.2.6.3.2), the DCCF or the NWDAF determines that notifications are to be stored in an ADRF. The subscription request may contain Storage Handling information with a requested ADRF storage lifetime and a request to be notified before data are deleted from the ADRF. 2a-b. The DCCF or the NWDAF determines the ADRF where data and/or analytics needs to be stored and requests that the ADRF subscribes to receive notifications. The determination may be made based on configuration or information supplied by the data consumer as described in clauses 6.1.4 and 6.2.6.3. The request to the ADRF specifies the data and/or analytics to which the ADRF will subscribe by invoking the Nadrf_DataManagement_StorageSubscriptionRequest service operation. The request may include the DataSetTag attribute which the data records are to be associated to when stored by ADRF. If the Storage Policy is not configured on the NWDAF or DCCF, the NWDAF or DCCF sends the Storage Handling Request to the ADRF. The NWDAF or DCCF may provide notification endpoint information to the ADRF for use by the ADRF to send notifications (implicit subscription) alerting the DCCF or NWDAF that data are about to be deleted (see step 12). 3. [Optional] The ADRF may, based on implementation, determine whether the same data and/or analytics is already stored or being stored, based on the information sent in step 2 by the consumer. If the DataSetTag attribute is included for data and/or analytics already stored or being stored, then ADRF associates the data records with such DataSetTag. 3a-c. [Optional] Based on Storage Handling information and Storage Policy, the ADRF, DCCF or NWDAF determines the Storage Approach (lifetime for storing data and whether consumer is notified prior to data deletion). 4a-b. [Optional] If the data and/or analytics is already stored and/or being stored in the ADRF, the ADRF sends Nadrf_DataManagement_StorageSubscriptionRequest Response message to the consumer indicating that data and/or analytics is stored. The ADRF includes the Storage Approach if determined in step 3c. 5a-b. The DCCF or NWDAF sends a subscription response to the Data or Analytics Consumer. The response may contain the Storage Approach determined by the ADRF, NWDAF or DCCF. 6a-b. ADRF subscribes to the DCCF or the NWDAF to receive notifications, providing its notification endpoint address and a notification correlation ID. In step 6b, the ADRF uses Nnwdaf_DataManagement_Subscribe (as depicted in Figure 6.2B.3-1) to obtain data from the NWDAF or Nnwdaf_AnalyticsSubscription_Subscribe to obtain analytics from the NWDAF (not depicted). 7. The DCCF, the MFAF or the NWDAF sends Analytics or Data notifications containing the notification correlation ID provided by the ADRF to ADRF notification endpoint address. The Analytics or Data notifications shall contain timestamp. The ADRF stores the notifications. Conditional on DCCF or NWDAF determining to stop storing data or analytics 8a-b. The DCCF or the NWDAF determines to stop storing notifications in the ADRF. 9a-b. The DCCF or the NWDAF requests that the ADRF unsubscribes to receive notifications. 10a-b. The ADRF sends a request to the DCCF or the NWDAF to unsubscribe to data notifications. The NWDAF may interact with the Data Source and the DCCF may interact with the Data Source and/or MFAF. Delivery of notifications from the DCCF/MFAF or NWDAF to the ADRF are subsequently halted. If Nnwdaf_AnalyticsSubscription_Subscribe was used in step 6b, then the Nnwdaf_AnalyticsSubscription_Unsubscribe is used in step 10b. Conditional on ADRF Managing the Storage Approach 11. The ADRF determines that the lifetime of the stored data or analytics has expired (according to the Storage Approach). 12. If indicated by the Storage Approach, the ADRF sends a notification alerting the DCCF or NWDAF that data are about to be deleted. NOTE: This is an implicit subscription. Conditional if the Storage Approach is based on Default Operator Policies provisioned on the NWDAF or DCCF (see step 3a-c and clause 5B.1) 13. The DCCF or NWDAF indicate in the response to the ADRF whether they will retrieve the Data or Analytics. 14. The DCCF or NWDAF may retrieve the Data or Analytics from the ADRF as described in clause 5B.1. Conditional if the Storage Approach is based on a Storage Handling Request received from the Data or Analytics Consumer (see step 3a-c and clause 5B.1) 15a-b. The NWDAF or DCCF sends a notification to the Data or Analytics Consumer alerting it that the data or analytics are about to be deleted. 16a-b. The Data or Analytics Consumer indicates in the response to the NWDAF or DCCF whether it will retrieve the Data or Analytics. 17. The DCCF or NWDAF indicate in the response to the ADRF whether the consumer will retrieve the Data or Analytics. 18. The Data or Analytics Consumer retrieves the stored data or analytics as described in clause 5B.1. Conditional on DCCF or NWDAF Managing the Storage Approach 19a-b. The NWDAF or DCCF determines that the lifetime of the stored data has expired (according to the Storage Approach). Conditional if the Storage Approach is based on Default Operator Policies provisioned on the NWDAF or DCCF (see step 3a-c and clause 5B.1) 20. The NWDAF or DCCF optionally retrieve the data or analytics from the ADRF as described in clause 5B.1. 21. The NWDAF or DCCF request that the data or analytics be deleted from the ADRF. Conditional if the Storage Approach is based on a Storage Handling Request received from the Data or Analytics Consumer (see step 3a-c and clause 5B.1) 22a-b. The NWDAF or DCCF sends a notification to the Data or Analytics Consumer alerting it that the data or analytics are about to be deleted. 23a-b. The Data or Analytics Consumer indicates in the response to the NWDAF or DCCF whether it will retrieve the Data or Analytics. 24. The Data or Analytics Consumer retrieves the stored data or analytics as described in clause 5B.1. 25. The NWDAF or DCCF request that the data or analytics be deleted from the ADRF. 26. The ADRF deletes the data or analytics if: - the Storage Approach in the ADRF indicates alerting the consumer is not required prior to data or analytics deletion; - in Steps 13 or 17 the DCCF or NWDAF indicated data or analytics will not be retrieved prior to deletion; - data or analytics retrieval in steps 14 or 18 has completed or - A request to delete data or analytics was received in steps 21 or 25. If the ADRF received a response from the NWDAF or DCCF in steps 14 or 17 indicating data or analytics will be retrieved but retrieval is not initiated before an adequate fixed time has elapsed, the ADRF may autonomously delete the data or analytics. 6.2B.4 Data removal from an ADRF The procedure depicted in figure 6.2B.4-1 is used by consumers (DCCF, NWDAF) to remove data previously stored in an ADRF. Figure 6.2B.4-1: Data Removal from an ADRF 1. A consumer requests that specified data be deleted from the ADRF using Nadrf_DataManagement_Delete request service operations. The request may include the DataSetTag attribute which the stored data records are associated to. 2. The ADRF deletes all copies of the stored data. 3. The ADRF indicates the result (i.e. data deleted, data not found, data found but not deleted) using Nadrf_DataManagement_Delete response service operations. NOTE: As described in clauses 6.2B.2 and 6.2B.3, data or analytics may be removed from an ADRF when a storage lifetime expires. The NWDAF, DCCF or ADRF can provide an alert to the consumer and the consumer may retrieve data or analytics prior to deletion by the ADRF. 6.2B.5 ML Model Storage in ADRF The procedure depicted in figure 6.2B.5-1 is used by NWDAF containing MTLF to store or update ML Model(s) in an ADRF. Figure 6.2B.5-1: ML Model Storage in ADRF 0. NWDAF containing MTLF determines to store or update ML Model(s) in ADRF based on MTLF policy. 1. NWDAF containing MTLF requests to store or update a (set of) ML Model(s) to the ADRF by invoking Nadrf_MLModelManagement_StorageRequest service, optionally including allowed NF instance list for the ML Model identifier as described in TS 33.501 [49]. 2. The ADRF locally maintains the association between the ML Model identifier and the NF instance ID of NWDAF containing MTLF and allowed NF instance list (if there is any in step 1). [Optional] If instead of the ML Model(s), the ML Model address(es) is/are included in request, ADRF downloads the ML Model(s) based on the ML Model address(es) and locally stores the ML Model(s). 3. The ADRF sends Nadrf_MLModelManagement_StorageRequest Response message to the consumer including the ML Model storage or ML Model Update result indication. 6.2B.6 ML Model removal from ADRF The procedure depicted in figure 6.2B.6-1 is used by NWDAF containing MTLF to delete ML Model from an ADRF. Figure 6.2B.6-1: ML Model removal from ADRF 1. NWDAF containing MTLF requests to delete the ML Model(s) previously stored in the ADRF using Nadrf_MLModelManagement_Delete request service operation. 2. The ADRF deletes both the stored ML Model(s) and related ML Model information. 3. The ADRF indicates the result (i.e. ML Model deleted, ML Model not found, ML Model found but not deleted) using Nadrf_MLModelManagement_Delete response service operation. 6.2B.7 ML Model retrieval from ADRF The procedure depicted in Figure 6.2B.7-1 is used by consumers (NWDAF containing MTLF, NWDAF containing AnLF or LMF) to retrieve ML Models from an ADRF. Figure 6.2B.7-1: Procedure for ML Model(s) retrieval from ADRF 1. The NWDAF service consumer (NWDAF containing AnLF or NWDAF containing MTLF) subscribes/requests for a (set of) trained ML Model(s) associated with a/an (set of) Analytics ID(s) by invoking the Nnwdaf_MLModelProvision_Subscribe / Nnwdaf_MLModelInfo_Request service. LMF may subscribe ML Model from NWDAF containing MTLF as described in clause 6.2A.1. LMF may request ML Model from NWDAF containing MTLF as described in clause 6.2A.3. 2. The NWDAF containing MTLF determines whether the set of ML Model(s) associated with a/an (set of) Analytics ID(s) should be retrieved from the ADRF. When NWDAF containing MTLF authorizes the NF consumer to retrieve the ML Model(s) stored in the ADRF directly, steps 3 and 4 is skipped. If NWDAF containing MTLF determines that the set of ML Model(s) corresponding Analytics ID(s) requested in step 1 needs to be retrieved from ADRF and the NF consumer is agnostic to where the ML Model(s) is stored, then Steps 3 and 4 is performed. NOTE: How NWDAF containing MTLF and ADRF authorizes the NF Consumer is specified in clause X.10 of TS 33.501 [49]. 3. The ADRF service consumer (NWDAF containing MTLF) requests for the ML Model stored in ADRF by invoking the Nadrf_MLModelManagement_RetrievalRequest Request (Storage Transaction Identifier or one or more unique ML Model identifier(s)) service operation. 4. The ADRF verifies the service consumer (NWDAF containing MTLF) as described in Annex X.10 of TS 33.501 [49]. If verification is successful, the ADRF sends Nadrf_MLModelManagement_RetrievalRequest Response (ML Model file address of Model file(s) stored in ADRF) service operation. 5. The NWDAF containing MTLF notifies/ response to the NWDAF service consumer with the tuple Analytics ID, one or more tuples of unique ML Model identifier and ML Model Information. The ML Model information may contain the ML Model file address or ADRF (Set) ID. The ADRF(Set) ID is included only when the NWDAF containing MTLF authorizes the NF consumer to retrieve the ML Model(s) stored in the ADRF in step 2. When ADRF (Set) ID is provided and the NWDAF containing MTLF authorizes the NF Service Consumer to retrieve all ML Models corresponding to a Storage Transaction ID, the Storage Transaction ID may be provided. In other cases, the NWDAF containing MTLF only provides ML Model identifier(s). 6. If in step 5, the NWDAF service consumer which is also the service consumer of ADRF in this steps 6 and 7 (NWDAF containing AnLF, NWDAF containing MTLF or LMF) received ADRF (Set) ID (where the ML Model(s) requested in step 1 is stored) and/or the ML Model provide indicator, then the service consumer may invoke the Nadrf_MLModelManagement_RetrievalRequest (Storage Transaction Identifier or one or more unique ML Model identifier(s)) service operation to get the ML Model stored in ADRF. 7. The ADRF verifies the service consumer as described in Annex X.10 of TS 33.501 [49]. If verification is successful, the ADRF sends Nadrf_MLModelManagement_RetrievalRequest Response (ML Model identifier(s) and address(es) of Model file(s) stored in ADRF) to the NWDAF (ADRF) service consumer. 6.2C Horizontal Federated Learning among Multiple NWDAFs 6.2C.1 General This clause specifies how NWDAF containing MTLF can leverage Horizontal Federated Learning technique to train an ML Model. 6.2C.2 Procedures 6.2C.2.1 Registration and Discovery procedure for Federated Learning Figure 6.2C.2.1-1: Registration and Discovery procedure for Federated Learning Steps 1 to 3 are the NWDAF registration procedure. 1-3. NWDAF containing MTLF as FL Server NWDAF or FL Client NWDAF registers to NRF with its NF profile, which includes NWDAF NF Type (see clause 5.2.7.2.2 of TS 23.502 [3]), Analytics ID(s), Address information of NWDAF, Service Area, FL capability type information (i.e. FL server and/or FL client) and Time interval supporting FL as described in clause 5.2. Steps 4 to 6 are the NWDAF Discovery procedure. 4-6. NWDAF containing MTLF determines ML Model requires FL based on operator policy (e.g. pre-configured list of ML Models), Analytic ID, Service Area/DNAI or data can not be obtained directly from data producer NF (e.g. due to privacy reasons). If the NWDAF containing MTLF can not perform as FL Server NWDAF, the MTLF first discovers and selects FL Server NWDAF from NRF by invoking the Nnrf_NFDiscovery_Request service operation. The following criteria might be used: Analytic ID of the ML Model required, Model filter information as defined in clause 6.2A.2, FL capability Type (i.e. FL server), Time Period of Interest, Service Area. Once the FL Server NWDAF (the requested or the selected one) is determined, the FL Server NWDAF discovers and selects other NWDAF(s) containing MTLF as FL Client NWDAF(s) from NRF by invoking the Nnrf_NFDiscovery_Request service operation. The following criteria might be used: Analytic ID of the ML Model required, FL capability Type (i.e. FL client), Service Area, NF type(s) of data sources from which the FL Client NWDAF is able to collect data for local ML Model training, Time Period of Interest, ML Model Interoperability Indicator. 7. FL Server NWDAF sends Federated Learning preparation request to the FL Client NWDAF(s), using Nnwdaf_MLModelTraining_Subscribe or Nnwdaf_MLModelTrainingInfo_Request service with the ML Preparation Flag, to check if the FL Client NWDAF(s) can meet the ML Model training requirement (e.g. Analytics ID, ML Model Interoperability information, Data Availability requirement, FL Availability time requirement (time span needed for the FL process), etc.). Data Availability requirement includes a list of Event IDs of the local data for training, and may also include the dataset statistical properties, the time window of the data samples and the minimum number of data samples. NOTE: Federated Learning preparation procedure (i.e. steps 7-9) can be skipped if the FL Server NWDAF can decide that the FL Client NWDAF(s) supports the FL procedure to be performed, e.g. based on information acquired from previous FL procedures or from the NRF, or based on local configuration. 8. FL Client NWDAF(s) checks if it can meet the ML Model training requirement and/or can successfully download the model if the model information is provided in the request and decides whether to join the Federated Learning process based on operator policy (e.g. pre-configured list of ML Models) and/or implementation. Example criteria used by FL Client NWDAF(s) may be based on its data availability and time availability, computation and communication capability and ML Model Interoperability information. 9. FL Client NWDAF(s) invokes Nnwdaf_MLModelTraining_Notify or Nnwdaf_MLModelTraining_Subscribe response or Nnwdaf_MLModelTrainingInfo_Request response service operation to indicate to the FL Server NWDAF whether it will join the FL procedure and may include the reason in the response message if it cannot join the FL process. 10. FL Server NWDAF determines the FL Client NWDAFs to be involved in the FL procedures based on the information received in step 6 and other information received in step 9 (if available). 6.2C.2.2 General procedure for Federated Learning among Multiple NWDAF Instances Figure 6.2C.2.2-1: General procedure for Federated Learning among Multiple NWDAF 0. The consumer (NWDAF containing AnLF or NWDAF containing MTLF) sends a subscription request to FL server NWDAF to retrieve an ML Model, using Nnwdaf_MLModelProvision including Analytics ID, ML Model Monitoring information as defined in clause 7.5.2, desired ML Model metric (e.g. ML Model Accuracy). NOTE 1: The ML Model Accuracy threshold can be used to indicate the target ML Model Accuracy of the training process and the FL server NWDAF may stop the training process when the ML Model Accuracy threshold is achieved during the training process. If the consumer (i.e. the NWDAF containing AnLF or NWDAF containing MTLF) provides the Time when the ML Model is needed, the FL Server NWDAF can take this information into account to decide the maximum response time for its FL Client NWDAF(s). 1. FL Server NWDAF selects NWDAF(s) containing MTLF (FL Client NWDAF(s)) as described in clause 6.2C.2.1. 2. FL Server NWDAF sends a Nnwdaf_MLModelTraining_Subscribe or Nnwdaf_MLModelTrainingInfo_Request to the selected NWDAF(s) containing MTLF (FL Client NWDAF(s)), which participate in the Federated learning to perform the local model training and determine the interim local ML Model information based on the input parameter in the request from FL Server NWDAF. The request includes the desired ML Model metric and initial ML Model and also includes the maximum response time, the FL Client NWDAF has to report the interim local ML Model information to the FL Server NWDAF before the maximum response time elapses. 3. [Optional] Each FL Client NWDAF collects its local data by using the current mechanism in clause 6.2 if the Client NWDAF has not local data available already. 4. During Federated Learning training procedure, each FL Client NWDAF further trains the ML Model provided by the FL Server NWDAF based on its local data and reports the interim local ML Model information to the FL Server NWDAF in Nnwdaf_MLModelTraining_Notify or Nnwdaf_MLModelTrainingInfo_Request response. The Nnwdaf_MLModelTraining_Notify or Nnwdaf_MLModelTrainingInfo_Request response may also include the Status report of FL training that includes local ML Model metric value (and optionally the used metric) computed by the FL Client NWDAF and Training Input Data Information (e.g. areas covered by the data set, sampling ratio, maximum/minimum of value of each dimension of data, etc.) in the FL Client NWDAF. The Nnwdaf_MLModelTraining_Notify or Nnwdaf_MLModelTrainingInfo_Response also includes the global ML Model metric (value (and optionally the used metric) when the ML Model Accuracy Check Flag was included in the Nnwdaf_MLModelTraining_Subscribe or Nnwdaf_MLModelTrainingInfo_Request (as described in step 7), the global ML Model metric value is calculated by the FL Client NWDAF using the local training data as the testing dataset. NOTE 2: The parameters in characteristics of local training dataset are up to the implementation. The local ML Model, which is sent from the FL Client NWDAF(s) to the FL Server NWDAF during the FL training process, is the information needed by the FL Server NWDAF to build the aggregated model. If the FL Client NWDAF is not able to complete the training of the interim local ML Model within the maximum response time provided by the FL Server NWDAF, the FL Client NWDAF shall send the Delay Event Notification that include the delay event indication, an optional cause code (e.g. local ML Model training failure, more time necessary for local ML Model training) and the expected time to complete the training if available to the FL Server NWDAF before the maximum response time elapses. 4a. [Optional]If FL Server NWDAF receives notification/response that the FL Client NWDAF is not able to complete the training within the maximum response time, the FL Server NWDAF may send to the FL Client NWDAF a new maximum response time in Nnwdaf_MLModelTraining_Subscribe or Nnwdaf_MLModelTrainingInfo_Request, before which the FL Client NWDAF has to report the interim local ML Model information to the FL Server NWDAF. Otherwise, the FL Server NWDAF may indicate FL Client NWDAF to skip reporting for this iteration. FL Server NWDAF includes the current iteration round ID in the message to indicate that the request is to modify the training parameters of the current iteration round. Alternatively, the FL Server NWDAF may inform the FL Client NWDAF to cease the ML Model training by sending termination request and to report back the current local ML Model updates. 5. The FL Server NWDAF aggregates all the local ML Model information retrieved at step 4, to update the global ML Model. The FL Server NWDAF may also compute the global ML Model metric value, e.g. based on the local ML Model metric value(s) provided by the FL Client NWDAF(s) or by applying the global model on the validation dataset (if available). The FL Server NWDAF may update the global ML Model each time a FL Client NWDAF provides updated local ML Model information, or the FL Server NWDAF may decide to wait for local ML Model information from all FL Client NWDAFs before updating the global ML Model. If the FL Server NWDAF provides the maximum response time for the FL Client NWDAF(s) to provide the interim local ML Model information in step 2, or the new maximum response time in step 4a, the FL Server NWDAF decides either to wait for the FL Client NWDAF(s) which have not yet provided their interim local ML Model within the new maximum response time or to aggregate only the retrieved local ML Model information instances to update global ML Model. The FL Server NWDAF makes this decision, considering the notification/response from the FL Client NWDAF or, if the notification is not received, based on local configuration. 6a. [Optional] Based on the consumer request in step 0, the FL Server NWDAF sends a Nnwdaf_MLModelProvision_Notify message to update the global ML Model metric value to the consumer periodically (e.g. a certain number of training rounds or every 10 min) or dynamically when some pre-determined status is achieved (e.g. the ML Model Accuracy threshold is achieved or training time expires). 6b. [Optional] The consumer decides whether the current model can fulfil the requirement, e.g. global ML Model metric value is satisfactory for the consumer and determines to stop or continue the training process. The consumer re-invokes Nnwdaf_MLModelProvision_Subscribe service operation as used in step 0 to continue the training process or invokes Nnwdaf_MLModelProvision_Unsubscribe service operation to stop the training process. 6c. [Optional] Based on the subscription request sent from the consumer in step 6b, the FL Server NWDAF updates or terminates the current FL training process. If the FL Server NWDAF received a request in step 6b to stop the Federated Training process, steps 7 and 8 are skipped. 7. If the FL procedure continues, FL Server NWDAF may determine FL Client NWDAF as described in clause 6.2C.2.3 and sends Nnwdaf_MLModelTraining_Subscribe or Nnwdaf_MLModelTrainingInfo_Request that includes the aggregated ML Model information to selected FL Client NWDAF(s) for next round of Federated Training. The request may also include the ML Model Accuracy Check Flag, that indicates the FL Client NWDAF(s) to use the local training data as the testing dataset to calculate the Model Accuracy of the global ML Model provided by the FL Server NWDAF. 8. Each FL Client NWDAF updates its own ML Model based on the aggregated ML Model information distributed by the FL Server NWDAF at step 7. NOTE 3: The steps 3-8 should be repeated until the training termination condition (e.g. maximum number of iterations, or the result of loss function is lower than a threshold) is reached. When the Federated Training procedure is complete, the FL Server NWDAF requests the FL client NWDAF(s) to terminate the FL procedure by invoking Nnwdaf_MLModelTraining_Unsubscribe service with a cause code that the FL process has finished and optionally with the final aggregated ML Model information. Then the FL client NWDAF(s) terminate the local model training and if the final aggregated ML Model information is received from the FL server NWDAF, the FL client NWDAF(s) can store it for further use. 9. After the training process is complete, the FL Server NWDAF may send Nnwdaf_MLModelProvision_Notify that includes the globally optimal ML Model information to the consumer. 6.2C.2.3 Procedures for Maintaining Federated Learning Processes This clause specifies how to maintain a Federation Learning process in FL execution phase, including FL Server NWDAF triggers reselection, addition, or removal of FL Client NWDAF(s), discovers new FL Client NWDAF(s) via NRF and FL Client NWDAF(s) joins or leaves Federated Learning process dynamically. In Federated Learning execution phase, FL Server NWDAF monitors the status changes of FL Client NWDAF(s) and may reselects FL Client NWDAF(s) based on the updated status, availability and/or capability, etc. NOTE 1: FL Server NWDAF checks if there is a need to carry on the FL execution phase and then reselects FL members for the next iteration if needed. Figure 6.2C.2.3-1: Procedure of FL Server NWDAF reselects FL Client NWDAF(s), FL Client NWDAF(s) Join or Leave Federated Learning Process Dynamically in Federated Learning execution phase The procedure for FL Server NWDAF reselecting FL Client NWDAF(s), FL Client NWDAF(s) joining or leaving Federated Learning process dynamically is as follows: 1a. FL Server NWDAF may get the updated status of current FL Client NWDAF(s) via NRF by using Nnrf_NFManagement service (as in clause 5.2.7.2 of TS 23.502 [3]) in the Federated Learning execution phase. FL Server NWDAF may subscribe to NRF for notifications of status changes of the current NWDAF(s) (FL Client NWDAFs 1…N) by invoking an Nnrf_NFManagement_NFStatusSubscribe service operation. NRF notifies the FL Server NWDAF the status changes of the current FL Client NWDAF(s) by invoking Nnrf_NFManagement_NFStatusNotify service operation(s). The status of a current FL Client NWDAF could be availability changes, capability changes (e.g. it will not support FL anymore, etc.). 1b. The current FL Client NWDAF(s) may inform FL Server NWDAF that it is leaving the Federated Learning process by invoking Nnwdaf_MLModelTraining_Notify service operation with Termination Request and cause code (reason for leaving, e.g. high NF load, time availability changes). 1c. FL Server NWDAF may get the information of the new FL Client NWDAF(s) dynamically via NRF by subscribing to the event that a new FL Client NWDAF registers (Nnrf_NFManagement_NFStatusSubscribe service as in clause 5.2.7.2 of TS 23.502 [3]). 1d. NWDAF may subscribe for NF load analytics of the FL Client NWDAF(s). 1e. FL Client NWDAF(s) may send Status report of FL training and Global ML Model Accuracy Information by invoking Nnwdaf_MLModelTraining_Notify service. 2. FL Server NWDAF checks FL Client NWDAF(s) status based on the received information and may determine whether reselection of FL Client NWDAF(s) for the next round(s) of Federated Learning is needed based on the received information from step 1. NOTE 2: Several examples of the factors that the FL Server NWDAF can consider to reselect the FL Client NWDAF(s) are updated status of FL Client NWDAF reported by NRF is different than the criteria were initially used for selecting the client; characteristics of local training dataset is different than global validation dataset owned by FL Server NWDAF and/or the metric value of the global model calculated using the local training dataset is much different from that calculated by other FL Client NWDAFs; the metric value of the global model calculated using the local training dataset is lower than the metric value calculated using the global validation dataset owned by FL Server NWDAF; the metric value of the global model calculated using the local training dataset is lower than ML Model metric value received in Nnwdaf_MLModelMonitor_Notify when FL Server NWDAF using AnLF-assisted MTLF ML Models Accuracy Monitoring; the load of the FL Client NWDAF (from the NF load Analytics or from the FL Client NWDAF directly) is high; the FL Server NWDAF receives leave request from the FL Client NWDAF; the FL Client NWDAF did not report the status of FL Training within the maximum response time. 3. [If re-selection is needed as judged in step 2] If step 1c is not performed, FL Server NWDAF may discover new candidate FL Client NWDAF(s) via NRF by using Nnrf_NFDiscovery services as in clause 5.2.7.3 of TS 23.502 [3]. FL Server NWDAF reselects FL Client NWDAF(s) from the current FL Client NWDAF(s) and the new candidate FL Client NWDAF(s) (found in steps 1c or 3). For the new candidate FL Client NWDAF(s), the interaction between FL Server NWDAF and FL Client NWDAF(s) is same as the selection procedure described in clause 6.2C.2.1. The adding / deleting FL Client NWDAF(s) may happen at the end of each iteration. 4. FL Server NWDAF sends termination request by invoking Nnwdaf_MLModelTraining_Unsubscribe service operation or Nnwdaf_MLModelTrainingInfo_Request service operation with Termination Request to the FL Client NWDAF(s), optionally indicating the reason, e.g. FL Client NWDAF is unselected by the FL Server NWDAF for the FL process, or the FL process is suspended, etc. And FL server may also send the updated global ML Model information to the unselected FL client NWDAF. FL Client NWDAF(s) terminates operations for the Federated Learning process if receive termination request from the FL Server NWDAF and may perform further action to be qualified in participation of FL training in the next cycles. NOTE 3: In the case of high load, the FL Client NWDAF can e.g. decline new service request. In the case of low accuracy, the FL Client NWDAF can gather new local training data. 6.2D AnLF Analytics Accuracy Monitoring Procedures 6.2D.1 General The Analytics Accuracy Information comprises a set of parameters as defined in clause 6.1.3. When multiple NWDAFs are deployed, some NWDAFs may be specialized with the analytics accuracy checking capability. When an NWDAF containing AnLF has the analytics accuracy checking capability, such an NWDAF is able to: - Receive a subscription or a request for analytics IDs via Nnwdaf_AnalyticsSubscription_Subscribe or Nnwdaf_AnalyticsInfo_Request service operation with the indication for activating the mechanisms for checking the accuracy of such analytics ID as defined in clause 6.1.3. - Provide the accuracy information to the consumer via Nnwdaf_AnalyticsSubscription_Notify or Nnwdaf_AnalyticsInfo_Request response service operation. NOTE 1: In this version of the specification, NWDAF containing AnLF can provide accuracy information to an NF consumer that subscribes or requests for the analytics. NOTE 2: When receiving a subscription from an NF consumer that includes a request for accuracy information, the analytics output and the accuracy information can be provided by NWDAF containing AnLF in a single notification or via separate notifications. NOTE 3: When receiving a request from an NF consumer that includes a request for accuracy information, the analytics and the accuracy information can be provided by NWDAF containing AnLF within the single response. NOTE 4: In this version of the specification, the subscription or request for accuracy information independently from requesting an analytics output is not supported. Based on the triggers described in clause 5C.1, NWDAF containing AnLF starts the accuracy monitoring and generation of Analytics Accuracy Information for an Analytics ID. The Analytics Accuracy Information may be requested per Analytics ID and scoped using the same parameters as those defined in the Target of Analytics Reporting as defined in clause 6.1.3 and Analytics Filter Information (e.g. for a specific area, specific slice) of the requested Analytics ID. When the analytics accuracy checking is activated in an NWDAF containing the AnLF, the NWDAF may store for a period of time the necessary information to determine the analytics accuracy and provide the accuracy information to consumers when requested or use it for its internal processes. NWDAF containing AnLF generates the accuracy information as described in clause 5C.1. 6.2D.2 Procedures for Analytics Accuracy Information Subscription This procedure is used by NF consumers of analytics ID to subscribe to receive analytics output and Analytics Accuracy Information related to the requested analytics ID for NF consumer. Figure 6.2D.2-1 shows the procedure for accuracy information subscription and provisioning. Figure 6.2D.2-1: Analytics Accuracy Information Subscribe 1. The NWDAF service consumer selects the appropriated NWDAF containing AnLF according to clause 5.2 and subscribes or modifies the subscription for Analytics Accuracy Information by invoking the Nnwdaf_AnalyticsSubscription_Subscribe service operation. The parameters that can be included in the subscription to trigger the accuracy information checking and provisioning are listed in clause 6.1.3. If the subscription is not the initial subscription request, it may include Analytics Feedback Information as described in clause 6.1.1. 2. When a subscription request is received, the NWDAF containing AnLF verifies the parameters of the Analytics Accuracy Request information received from the NWDAF service consumer in step 1. The NWDAF containing AnLF starts the Analytics Accuracy Monitoring and generation of the Analytics Accuracy Information related to the analytics ID indicated in the subscription according to the parameters defined in Analytics Accuracy Request Information in clause 6.1.3. The NWDAF containing AnLF is to compute Analytics Accuracy Information according to the methods in clause 6.2D.1. If the NWDAF containing AnLF does not have enough necessary data, it will perform step 3b to collect ground truth data before computing Analytics Accuracy Information. The NWDAF containing AnLF may have started to perform the Analytics Accuracy Monitoring and Analytics Accuracy Information generation, triggered by other NWDAF service consumer(s) before. Upon receiving a new request from the NWDAF service consumer, the NWDAF containing AnLF determines whether new data collection is needed for Analytics Accuracy Information generation based on the corresponding analytics subscription. In addition to the received request from the NWDAF service consumer, based on local policy, the NWDAF containing AnLF may determine to start the Analytics Accuracy Monitoring and Analytics Accuracy Information generation. 3a. The NWDAF containing AnLF performs the data collection for the subscribed analytics ID and generates the analytics output. 3b. The NWDAF containing AnLF performs the data collection (e.g., ground truth data collection) for accuracy information generation for the subscribed analytics ID and generates the associated Analytics Accuracy Information. If Analytics Feedback Information is included in step 1, the NWDAF containing AnLF may take it into account and determine whether it affects the ground truth data by the internal logic to generate Analytics Accuracy Information. 4a. The NWDAF containing AnLF provides the analytics output according to the parameters defined in Analytics Reporting Information included in the subscription request when there is no Analytics Accuracy Request Information included in the subscription in step 1. NOTE: Steps 3b and 4a can occur in any order. 4b. The NWDAF containing AnLF provides the Analytics Accuracy Information together with the analytics output for the analytics ID according to the parameters defined in the Analytics Accuracy Request Information included in the subscription request. 4c. The NWDAF containing AnLF provides only the Analytics Accuracy Information for the analytics ID according to the parameters defined in the Analytics Accuracy Request Information included in the subscription request. The Analytics Accuracy Information is provided in a separated notification when the periodicity for providing the Analytics Accuracy Information indicated in the Analytics Accuracy Request Information is different from the periodicity for providing the analytics output indicated in the subscription request, or the accuracy value is under the analytics accuracy threshold which is indicated in the subscription request or locally configured. 5. When determining the low or insufficient accuracy for an analytics ID, i.e. the deviation of the analytics output using the trained ML Model from the ground truth data (which are collected from Data Producer NF corresponding to the requested analytic ID at the time which the prediction refers to) does not meet analytics accuracy requirement, which indicates the accuracy value is under the analytics accuracy threshold(s) (which are locally configured or received in the subscribe request), the NWDAF containing AnLF may notify the NWDAF Service consumer with the Stop Analytics Output Consumption indication and the Stop Analytics Output Consumption time window. 6. (Optional) The NWDAF Service Consumer may decide to stop the consumption of analytics output without unsubscribing to the analytics ID, based on its own logic or based on a received notification from NWDAF with the Stop Analytics Output Consumption indication. The NWDAF Service Consumer invokes the Nnwdaf_AnalyticsSubscription_Subscribe service operation including the Subscription Correlation ID to modify an existing subscription and provides the parameter Pause analytics consumption flag in the Analytics Accuracy Request Information. 7. When the NWDAF determines an improvement in the accuracy of an analytics ID (e.g. meet the accuracy requirement of the analytics consumer) or when the Stop Analytics Output Consumption time window set by itself is finished, the NWDAF notifies the NWDAF Service Consumer of the accuracy information for the analytics ID to resume the consumption of the analytics output, therefore reactivating an existing analytics ID subscription that has been previously stopped. 8. (Optional) The NWDAF Service Consumer based on its own logic can notify the NWDAF to resume the notification of analytics output, therefore reactivating an existing subscription to analytics ID that has been paused either by NWDAF Service Consumer request (step 6) or by NWDAF indication (step 5). The NWDAF Service Consumer invokes the Nnwdaf_AnalyticsSubscription_Subscribe service operation including the Subscription Correlation ID to modify an existing subscription and provides the parameter Resume Analytics Subscription request in the Analytics Accuracy Request Information. 6.2D.3 Procedures for Analytics Accuracy Information Request This procedure is used by NF consumers of analytics ID to request Analytics Accuracy Information related to the requested analytics ID for NF consumer. Figure 6.2D.3-1 shows the procedure for accuracy information request and response. Figure 6.2D.3-1: Analytics Accuracy Information Request 1. The NWDAF service consumer selects the appropriated NWDAF containing AnLF according to Clause 5.2 and requests for Analytics Accuracy Information by invoking the Nnwdaf_AnalyticsInfo_Request service operation. The parameters that can be included in the request to trigger the accuracy information checking and provisioning are listed in clause 6.1.3. 2. When a request is received, the NWDAF containing AnLF determines whether the request is only for analytics output generation or if it includes the Analytics Accuracy request. If the Analytics Accuracy request is included, the NWDAF containing AnLF starts the Analytics Accuracy Monitoring and generation of the Analytics Accuracy Information related to the analytics ID indicated in the request and according to the parameters defined in Analytics Accuracy Request Information in clause 6.1.3. The NWDAF containing AnLF is to compute Analytics Accuracy Information according to the methods in Clause 6.2D.1. If the NWDAF containing AnLF does not have enough necessary data, it will perform step 3b to collect ground truth data before computing Analytics Accuracy Information. The NWDAF containing AnLF may have started to perform the Analytics Accuracy Monitoring and Analytics Accuracy Information generation, triggered by other NWDAF service consumer(s) before. Upon receiving a new request from the NWDAF service consumer, the NWDAF containing AnLF determines whether new data collection is needed for Analytics Accuracy Information generation based on the corresponding analytics request. In addition to the received request from the NWDAF service consumer, based on local policy, the NWDAF containing AnLF may determine to start the Analytics Accuracy Monitoring and Analytics Accuracy Information generation. 3a. The NWDAF containing AnLF performs the data collection for the requested analytics ID and generates the analytics output. 3b. The NWDAF containing AnLF performs the data collection (e.g., ground truth data collection) for accuracy information generation for the requested analytics ID and generates the associated Analytics Accuracy Information. 4a. The NWDAF containing AnLF provides the analytics output according to the parameters defined in Analytics Reporting Information included in the request, when no Analytics Accuracy Request Information is included in the request in step 1. NOTE: Step 3b, 4a can occur in any order. 4b. The NWDAF containing AnLF provides the requested analytics output and Analytics Accuracy Information for the analytics ID according to the parameters defined in the Analytics Accuracy Request Information included in the request. 6.2E MTLF-based ML Model Accuracy Monitoring 6.2E.1 General MTLF-based ML Model Accuracy Monitoring procedure is where an NWDAF containing MTLF determines ML Model degradation based on newly collected test data and retrain or reprovisioning the existing ML Model. 6.2E.2 Procedure for MTLF-based ML Model Accuracy Monitoring Figure 6.2E.2-1 illustrates the procedure for monitoring the accuracy of the provisioned ML Model using newly collected data. NWDAF containing AnLF may provide inference data to NWDAF containing MTLF for model accuracy monitoring and the NWDAF containing MTLF determines retraining or re-provisioning of the ML Model. Figure 6.2E.2-1: Procedure for MTLF-based ML Model Accuracy Monitoring 1. An analytics consumer initiates a subscription for analytics exposure services towards an NWDAF containing AnLF. 2. The NWDAF containing AnLF requests an ML Model from the appropriate NWDAF containing MTLF, using the Nnwdaf_MLModelProvision_Subscribe service operation. The NWDAF containing AnLF may include an ML Model accuracy threshold which is used as an indicator to execute the accuracy monitoring operations as defined in clause 6.2A.2. NWDAF containing AnLF may include a DataSetTag (see clause 6.2B.1) and/or ADRF ID, which is used to store and fetch the inference data (including input data, prediction and the ground truth data at the time which the prediction refers to) from ADRF which are relevant for the accuracy monitoring and re-training/re-provisioning of ML Model. If the NWDAF containing AnLF receives ML Model(s), the NWDAF containing AnLF sends set of tuples (unique ML Model identifier and a DataSetTag and/or ADRF ID) to the NWDAF containing MTLF by invoking Nnwdaf_MLModelProvision_Subscribe service operation for subscription modification. 3. The NWDAF containing MTLF provides trained ML Model(s) to the NWDAF containing AnLF as specified in clause 6.2A and clause 6.2B. The NWDAF containing MTLF may include an accuracy information which is used to indicate the accuracy of ML Model during the training. When the step 2 is for a subscription modification (i.e. including Subscription Correlation ID) and contains the set of tuples (unique ML Model identifier and a DataSetTag and/or ADRF ID), the NWDAF containing MTLF determines the relationship between the ML Model and the DataSetTag. 4. The NWDAF containing AnLF registers the use of the ML Model with the NWDAF containing MTLF to indicate its capability of sending Analytics Feedback Information and/or Analytics Accuracy for the ML Model as described in clause 6.2E.3. NOTE 1: The NWDAF containing AnLF receives the Analytics Feedback information from the Analytics consumer. 5. Due to the registration in the previous step, the NWDAF containing MTLF may subscribe to the NWDAF containing AnLF to get Analytics Feedback Information and/or Analytics Accuracy for the provisioned ML Model by invoking Nnwdaf_MLModelMonitor_Subscribe service operation, if the service operation is supported by the NWDAF containing AnLF. 6. The Analytics consumer may send Analytics Feedback Information in an Nnwdaf_AnalyticsSubscription_Subscribe message as described in clause 6.1.1. 7. The NWDAF containing AnLF may send the Analytics Feedback Information and/or Analytics Accuracy for the provisioned ML Model by invoking Nnwdaf_MLModelMonitor_Notify service operation as requested in step 5. When the NWDAF containing MTLF receives Analytics Feedback Information and/or Analytics Accuracy of the ML Model, the NWDAF containing MTLF may trigger step from 8 to 13 to enhance the ML Model accuracy. 8a-8f. The NWDAF containing MTLF, based on the request(s) from one or more NWDAF containing AnLF or its local policy, determines whether to perform ML Model Accuracy Monitoring and re-training/re-provisioning of ML Model by collecting new data from various data sources: - The NWDAF containing MTLF may collect new data for ML Model Accuracy monitoring, re-training and re-provisioning from the data source NFs and DCCF by invoking Nnf_EventExposure_Subscribe and Ndccf_DataManagement_Susbscribe service operation, respectively. - When ADRF ID and/or DataSetTag is given by step 2, the NWDAF containing MTLF may retrieve historical data from the ADRF indicated by the NWDAF containing AnLF at step 2. by invoking Nadrf_DataManagementRetrievalRequest or Nadrf_DataManagementRetrieval_Subscribe service operation. Otherwise, the NWDAF containing MTLF may retrieves the historical data from the DCCF or the NWDAF containing AnLF by invoking Ndccf_DataManagement_Subscribe or Nnwdaf_DataManagement_Subscribe service operation, respectively. - If the NWDAF containing AnLF does not include a DataSetTag with ADRF ID at step 2, the NWDAF containing MLTF may request ADRF to subscribe for the collection of the analytics and data that correspond to the analytics generated by the ML Model provisioned in step 3, using the procedures defined in clause 6.2B.3. - The NWDAF containing MTLF may subscribe to UDM to get notification on change in the subscription data for Target of ML Model Reporting by invoking Nudm_SDM_Subscribe service operation and the UDM subscribes to the UDR to get notifications of the modification on UE subscription data by invoking Nudr_DM_Subscribe service operation. - The NWDAF containing MTLF may consider the data quality into the accuracy monitoring by collecting fault prediction analytics data from MDAS to determine the status of Data Source NFs, using MDA Request. If the NWDAF containing MTLF has already collected new test data and performed ML Model Accuracy Monitoring and retraining which is triggered by other NWDAF containing AnLF(s) (for ML Model Accuracy Monitoring and retraining), the NWDAF containing MTLF, based on its internal logic, determines whether to use the data for the subscription or not. 9a-9f. The NWDAF containing MTLF receives the requested data from various sources as requested in steps 8a-8f. 10. Based on the collected analytics and data from steps 9a-9f, the NWDAF containing MTLF computes the accuracy using the methods described in clause 5C.1. The NWDAF containing MTLF may discard data from data sources if it detects the data quality of that source is not good. The NWDAF containing MTLF may generate prediction with the collected input data to calculate the accuracy if only input data and ground truth data are available. NOTE 2: How the NWDAF containing MTLF determines whether the data from the data source is of good quality or needs to be discarded is up to the NWDAF implementation and configuration. 11. An accuracy report is sent to the NWDAF containing AnLF, e.g. when the reporting threshold is met by invoking Nnwdaf_MLModelProvision_Notify service operation. 12. Based on the computed accuracy, the NWDAF containing MTLF may decide to re-train/re-provision the ML Model. 13. When the newly generated or re-trained ML Model is ready, the NWDAF containing MTLF sends new or re-trained ML Model to the NWDAF containing AnLF by invoking Nnwdaf_MLModelProvision_Notify service operation. The NWDAF containing MTLF may send the accuracy report of the new or re-trained ML Model to the NWDAF containing AnLF. 6.2E.3 Procedure for AnLF-assisted MTLF ML Models Accuracy Monitoring 6.2E.3.1 General The procedures described in this clause enable the following functionality: - An NWDAF containing AnLF may register with an NWDAF containing MTLF when it starts using an ML Model and monitoring the accuracy of analytics generated by that ML Model for a given Analytics ID. It is assumed that the NWDAF containing AnLF obtained the ML Model in a previous interaction with the NWDAF containing MTLF, e.g. using the Nnwdaf_MLModelInfo_Request or Nnwdaf_MLModelProvision_Subscribe services. If the AnLF receives ML Model provider information in the Nnwdaf_MLModelInfo_Request response or Nnwdaf_MLModelProvision_Subscribe Notify message, it registers to the NWDAF containing MTLF indicated by the ML Model provider information. This registration enables the NWDAF containing MTLF to become aware of NWDAF containing AnLF that are using a given ML Model for certain Analytics ID and that the NWDAF containing AnLF supports the capability of monitoring the accuracy of the corresponding analytics. - An NWDAF containing MTLF may subscribe to an NWDAF containing AnLF where an existing Nnwdaf_MLModelMonitor service is established for receiving notifications of the accuracy of analytics generated by a given ML Model for a certain Analytics ID. NWDAF containing AnLF can generate the accuracy information in many ways: e.g. comparing predictions of ML Model and its corresponding ground truth data, comparing changes in internal configuration for the analytics ID generation, previous existent records of Analytics Accuracy Information, etc. 6.2E.3.2 Procedures for registering the monitoring of the analytics accuracy of an ML Model When an NWDAF containing AnLF starts making use of an ML Model and it has the ability either to monitor the analytics accuracy of the ML Model, or to deliver Analytics Feedback Information for the analytics generated by the ML Model, it registers with the NWDAF containing MTLF, that is responsible for training/updating this ML Model. When the NWDAF containing AnLF is no longer using the ML Model or monitoring the accuracy of the analytics generated by that ML Model for the Analytics ID, it de-registers it with the responsible NWDAF containing MTLF. Figure 6.2E.3.2-1 illustrates the procedure by which an NWDAF containing AnLF registers with an NWDAF containing MTLF that it is starting to make use and monitor the analytics accuracy of an ML Model. A new Nnwdaf_MLModelMonitor_Register service operation is used for that purpose. Figure 6.2E.3.2-1: Procedure for ML Model monitoring registration An NWDAF containing AnLF may start monitoring the accuracy of an ML Model based on local policy or request from its service consumer. 1-2. The NWDAF containing AnLF sends an Nnwdaf_MLModelMonitor_Register request to an NWDAF containing MTLF (NF ID of the NWDAF containing AnLF, unique identifier of the ML Model, optionally: subscription endpoint of the Nnwdaf_MLModelMonitor_Subscribe service operation at the NWDAF containing AnLF, Analytics ID, Target of Analytics Reporting, Analytics filter). The NWDAF containing MTLF is now aware of the NF ID of the NWDAF containing AnLF that is monitoring the accuracy of that ML Model. If the NWDAF containing AnLF is a target NWDAF in analytics transfer procedure (as defined in clause 6.1B), based on the ML Model Accuracy Information received from source NWDAF containing AnLF, the NWDAF containing AnLF also includes in the Nnwdaf_MLModelMonitor_Register service request the ML Model accuracy transfer indication, which includes the original Subscription Correlation ID for the ML Model Accuracy Information provided by the source NWDAF containing AnLF and the source NF ID of the NWDAF containing AnLF. NOTE 1: These parameters support the NWDAF containing MTLF to map the registration of a new NWDAF containing AnLF with an existing subscription for consumption of ML Model Accuracy Information from a previous NWDAF containing AnLF (i.e. source NWDAF containing AnLF which as described in steps 3-4 may provide a termination indication), enabling NWDAF containing MTLF to reassociate the data from the previous subscription to the new the subscription for ML Model accuracy provisioning at the new NWDAF containing AnLF. 3-4. When the NWDAF containing AnLF is no longer using the ML Model, it sends an Nnwdaf_MLModelMontior_Deregister service operation. If NWDAF containing AnLF is registered with a NWDAF containing MTLF, is a source NWDAF containing AnLF in an analytics transfer procedure (as defined in clause 6.1B) and is no longer using the ML Model, the NWDAF containing AnLF sends Nnwdaf_MLModelMontior_Deregister service operation request including the ML Model accuracy provisioning termination information, which includes: a termination indication, the termination cause set to analytics transfer and optionally the NWDAF containing AnLF NF ID of the target NWDAF. NOTE 2: The ML Model accuracy termination information is used by the NWDAF containing MTLF to determine whether the termination request is from the source NWDAF containing AnLF. If so, the NWDAF containing MTLF will not delete any data immediately upon receiving of a de-registration request. Then the NWDAF containing MTLF is able to associate the data from the source NWDAF containing AnLF to the target NWDAF containing AnLF. 6.2E.3.3 Procedures for monitoring the analytics accuracy of an ML Model An NWDAF containing MTLF, due to the registration of monitoring of the analytics accuracy of an ML Model received from NWDAF containing AnLF and local policies, subscribes to the NWDAF containing AnLF for receiving notifications of either the accuracy of the ML Model, or Analytics Feedback Information of the ML Model. The NWDAF containing MTLF may get the Subscription endpoint address of the NWDAF containing AnLF from the information received in a previous registration or through a service discovery procedure at the NRF. Figure 6.2E.3.3-1 illustrates the procedure either for monitoring the analytics accuracy of an ML Model or for delivery of Analytics Feedback Information of an ML Model. Nnwdaf_MLModelMonitor_Subscribe and Nnwdaf_MLModelMonitor_Notify service operations are used for the purposes. A service consumer, i.e. an NWDAF containing MTLF, subscribes at a service producer, i.e. an NWDAF containing AnLF, to be notified when either the analytics accuracy of the previously provisioned ML Model is not sufficient, or Analytics Feedback Information is retrieved from analytics consumer NF. Figure 6.2E.3.3-1: Procedure for monitoring the analytics accuracy of an ML Model 0. Upon the reception of an Nnwdaf_MLModelMonitor_Register request and based on local policy, the NWDAF containing MTLF determines to subscribe to the Analytics Accuracy Monitoring for the ML Model as defined in clause 5C.1. 1. The NWDAF containing MTLF sends an Nnwdaf_MLModelMonitor_Subscribe request (Analytics ID(s), unique identifier(s) of the ML Model(s) to be monitored, desired accuracy metrics to be monitored, optionally Reporting Threshold(s), Analytics ID, Target of Analytics Reporting and Analytics filter for each ML Model identifier or Reporting Period) to an NWDAF containing AnLF subscription endpoint. When the NWDAF containing MTLF determines during the registration process described in clause 6.2E.3.2 that a subscription request for ML Model Accuracy Monitoring to an NWDAF containing AnLF is related to a previous subscription for ML Model Accuracy Information to a different NWDAF containing AnLF (due to changes in the provider of the ML accuracy monitoring for a given ML Model, as an effect of analytics transfer among NWDAFs containing AnLF), the NWDAF containing MTLF may use as base for the new subscription request at the new NWDAF containing AnLF the parameters associated with the original subscription identification for the ML Model Accuracy Information that was received in the registration request of the new NWDAF containing AnLF, as described in steps 1-2 of clause 6.2E.3.2. 2. The NWDAF containing AnLF sends a response to the NWDAF containing MTLF. 3. The analytics consumer NF may send Analytics Feedback Information to the NWDAF containing AnLF as described in clause 6.1.1. 4. When step 1 is triggered, the NWDAF containing AnLF may start monitoring the analytics accuracy of the ML Model(s), if it not started yet. NOTE 1: The NWDAF containing AnLF can monitor the analytics accuracy in many ways: e.g. comparing predictions of ML Model and its corresponding ground truth data, comparing changes in internal configuration for the analytics ID generation, previous existent records of Analytics Accuracy Information, etc. 5. The NWDAF containing AnLF determines whether the analytics accuracy of the ML Model is insufficient, i.e. deviation of the output analytics using the trained ML Model from ground truth data (which are collected from Data Producer NF corresponding to analytic ID requested at the time which the prediction refers to) does not meet the analytics accuracy requirement, which indicates the accuracy value is under the Reporting Threshold(s) (which are locally configured or received in the Subscribe request), or the Reporting Period indicated in the Subscribe request is reached. 6. Either the Analytics Feedback Information is retrieved at step 3 or the NWDAF containing AnLF detects the analytics accuracy of ML Model is insufficient at step 5, the NWDAF containing AnLF sends an Nnwdaf_MLModelMonitor_Notify request to the notification endpoint (e.g. the NWDAF containing MTLF). The Notify request includes either Analytics Feedback Information, or the monitored accuracy information of the ML Model (e.g. unique identifier(s) of the ML Model(s) to be monitored, Analytics ID, Target of Analytics Reporting and Analytics filter for each ML Model identifier, a deviation value which indicates the deviation of the predictions generated using the ML Model(s) from the ground truth data and the network data when the deviation occurs (which can be used by the NWDAF containing MTLF for possible ML Model retraining) and the number of inferences that were performed during the time interval between Nnwdaf_MLModelMonitor_Register request and the Notify request or between the time of last Notification message and the time of the current Notification message) and optionally an indication that the analytics accuracy of the ML Model does not meet the requirement of accuracy for the ML Model. 7. The NWDAF containing MTLF sends a response. 8. The NWDAF containing MTLF determines whether the ML Model is degraded or not based on the notification at step 6. If the notification contains Analytics Feedback Information, the NWDAF containing MTLF may determine ML Model degradation based on the procedures as described in clause 6.2E.2. Otherwise when the NWDAF containing MTLF has received the multiple Analytics Accuracy Information, from one or more NWDAFs containing AnLF, it may consider that the ML Model is degraded/to be updated (i.e. enough number Analytics Accuracy Information received from one or more NWDAFs containing AnLF, indicating insufficient analytics accuracy). NOTE 2: The actual mechanism for the NWDAF containing MTLF for determining the degradation of the ML Model degradation is an internal procedure of the NWDAF containing MTLF, e.g. the NWDAF containing MTLF calculate a global accuracy based on the Analytics Accuracy Information and the number of inferences received from multiple NWDAFs containing AnLF. 9. When an ML Model is considered degraded / to be updated at step 8, the NWDAF containing MTLF re-trains the existing ML Model or selects a new ML Model. If the network data was not included in the Nnwdaf_MLModelMonitor_Notify request of step 6, the NWDAF containing MTLF may request data from the NWDAF containing AnLF, ADRF and/or other 5GS entities as specified in clause 6.2 and use the collected data for ML Model retraining. The NWDAF containing MTLF notifies the NWDAF(s) containing AnLF with the updated trained ML Model Information by invoking Nnwdaf_MLModelProvision_Notify service operation, as described in clause 6.2A. 6.2E.4 Procedure for MTLF-based AI/ML model performance monitoring for LMF-based AI/ML Positioning NWDAF containing MTLF may support to perform the performance monitoring for ML model used by LMF for LMF-based AI/ML positioning. The LMF may request the NWDAF to monitor the ML model performance by including the ML Model identifier of the ML model, for which the performance needs to be monitored, and an ML Model Accuracy Threshold, which is used as an indication to execute the accuracy monitoring operations, in the Nnwdaf_MLModelProvision_Subscribe or Nnwdaf_MLModelInfo_Request service operation as described in clause 6.2A. When the NWDAF containing MTLF monitors the ML Model performance, he NWDAF requests the LMF to provide the following as input data using the procedure of input data collection as specified in clause 6.22.4 of TS 232.273 [39]: - Optionally, Location measurement data from the PRU(s)/UE(s) and/or the NG-RAN. - Ground truth data from PRU(s)/UE(s). - Optionally, the corresponding inference output, i.e. location estimation of the PRU(s)/UE(s) using LMF-based AI/ML positioning. Before collecting the input data from UE(s), LMF needs to check the user consent in UDM during the data collection procedure. The NWDAF containing MTLF evaluates the ML Model performance by comparing the ground truth data against the corresponding location estimation of the PRU(s)/UE(s), which is derived by the NWDAF containing MTLF using its local ML model and the location measurement data from the PRU(s)/UE(s) and/or the NG-RAN, or derived by the LMF using the ML model and location measurement data from the PRU(s)/UE(s) and/or the NG-RAN. If the ML Model Accuracy threshold requested by the LMF is not satisfied, the NWDAF containing MTLF retrains the ML model, or the NWDAF containing MTLF sends to LMF an ML Model degradation indicator as specified in clause 6.2A.2, which may trigger the LMF to change the positioning method, e.g. from LMF-based AI/ML Positioning to legacy positioning. Figure 6.2E.4-1 illustrates the procedure by which the NWDAF containing MTLF monitors the ML model performance for LMF-based AI/ML positioning. Figure 6.2E.4-1: Monitoring performance of ML Model used for LMF-based AI/ML positioning 1. The LMF may request an ML Model for LMF-based AI/ML positioning by invoking the Nnwdaf_MLModelProvision_Subscribe or Nnwdaf_MLModelInfo_Request service operation. The LMF may include an ML Model identifier and an ML Model Accuracy Threshold for ML model performance monitoring. 2. The NWDAF containing MTLF trains the ML Model and, based on the ML Model Monitoring information (i.e. ML Model identifier, ML Model Accuracy Threshold) or its local policy, decides to start monitoring the accuracy of the ML model. 3. The NWDAF containing MTLF requests input data from the LMF for ML model monitoring as per clause 6.22.4 of TS 23.273 [39]. 4. The NWDAF containing MTLF determines, by comparing the ground truth data against the corresponding location estimation of the PRU(s)/UE(s), whether the ML model is degraded or not. 5. If the ML model performance cannot meet the requirement (e.g. lower than the ML Model Accuracy threshold), the NWDAF containing MTLF may retrain the ML model. 6. The NWDAF containing MTLF provides a new or retrained ML model to the LMF, or notifies the LMF that the ML model is degraded by invoking the Nnwdaf_MLModelProvision_Notify or Nnwdaf_MLModelInfo_Request response service operation. 6.2F Procedure for ML Model Training 6.2F.1 ML Model Training Subscribe/Unsubscribe The procedure in Figure 6.2F.1-1 is used by an NWDAF service consumer, i.e. an NWDAF containing MTLF to subscribe to another NWDAF, i.e. an NWDAF containing MTLF, for a trained ML Model based on the ML Model file or ML Model information as described in clause 6.2F.2 provided by the NWDAF service consumer. The service may be used by an NWDAF containing MTLF to enable e.g. Federated Learning or to update ML Model. The service is also used by an NWDAF service consumer to request an NWDAF containing MTLF to prepare training ML Model or modify existing ML Model training subscription. Figure 6.2F.1-1: Procedure for ML Model Training subscribe/unsubscribe 1. The NWDAF service consumer may subscribe or unsubscribe for training an ML Model by invoking the Nnwdaf_MLModelTraining_Subscribe/ Nnwdaf_MLModelTraining_Unsubscribe service operation. The parameters that can be provided by the NWDAF service consumer are listed in clause 6.2F.2. In order to enable Federated Learning, NWDAF Service consumer act as FL Server NWDAF can subscribe to multiple NWDAFs containing MTLF act as FL Client NWDAFs, which are selected by the FL Server NWDAF. The FL server NWDAF may use the request to check if an NWDAF can meet the ML Model training requirement (e.g. ML Model Interoperability information, Analytics ID, Serving Area and/or availability of data and time). In such case, the FL server NWDAF includes an ML Preparation Flag. When the ML Preparation Flag presents in the request, the service provider NWDAF only checks if it can meet the ML Model training requirement (e.g. ML Model Interoperability information, Analytics ID, Serving Area and/or availability of data and time) and / or can successfully download the model if the model information is provided. The FL server NWDAF may use the request to get the Model Accuracy information of the global ML Model calculated by the FL Client NWDAFs. In such cases, the service consumer NWDAF includes a Model Accuracy Check Flag. When the Model Accuracy Check Flag is present in the request, the service provider NWDAF uses the local training data as the testing dataset to calculate the Model Accuracy information of the ML Model provided by the service consumer NWDAF. When NWDAF service consumer determine to further update the ML Model, NWDAF service consumer modifies the subscription by invoking Nnwdaf_MLModelTraining_Subscribe service operation including Subscription Correlation ID with ML Model Information (as defined in clause 6.2A.2). 2. The NWDAF containing MTLF trains ML Model provided at step 2 by collecting new data or re-use the data that it owns. If the ML Model file is not provided in step 1, the NWDAF containing MTLF shall first get the ML Model using the information indicated at step 1. 3. When the NWDAF containing MTLF completes ML Model training, the NWDAF containing MTLF notifies the NWDAF service consumer with ML Model Information (as defined in clause 6.2A.2) of updated ML Model by invoking the Nnwdaf_MLModelTraining_Notify service operation. The parameters that can be provided by the NWDAF containing MTLF as service provider is specified in clause 6.2F.2. If the NWDAF containing MTLF determines to terminate the ML Model training, i.e. NWDAF containing MTLF will not provide further notifications related to this request, then the NWDAF containing MTLF may notify the NWDAF Service consumer a Terminate Request indication with cause code (e.g. NWDAF overload, not available for the FL process anymore, etc.) by invoking the Nnwdaf_MLModelTraining_Notify service operation. In order to enable Federated Learning, NWDAF containing MTLF acting as FL Client NWDAF can notify NWDAF Service consumer acting as FL Server NWDAF the local ML Model information and status report of FL training including accuracy information of local model and Training Input Data Information (e.g. areas covered by the data set, sampling ratio, maximum/minimum of value of each dimension, etc.). If the Model Accuracy Check Flag is present in the Nnwdaf_MLModelTraining_Subscribe, the service provider NWDAF acting as FL Client NWDAF may notify the NWDAF Service consumer acting as FL Server NWDAF the Model Accuracy information of the global ML Model. 6.2F.2 Contents of ML Model Training The consumers of the ML Model training services (i.e. an NWDAF containing MTLF) may provide the input parameters in Nnwdaf_MLModelTraining_Subscribe or Nnwdaf_MLModelTrainingInfo_Request service operations as listed below: - Analytics ID: identifies the analytics for which the ML Model is requested to be trained. - ML Model Interoperability Information as defined in clause 6.2A.2. - (Only for Nnwdaf_MLModelTraining_Subscribe) A Notification Target Address (+ Notification Correlation ID) as defined in TS 23.502 [3] clause 4.15.1, allowing to correlate notifications received from the NWDAF containing MTLF with the subscription. - [OPTIONAL] ML Model Information (as defined in clause 6.2A.2). - [OPTIONAL] ML Model file. NOTE 1: It is up to NWDAF implementation to determine whether to include ML Model file in input parameters considering ML Model file size, etc. - [OPTIONAL] ML Model identifier: identifies the provided ML Model. - [OPTIONAL] ML Preparation Flag: identifies whether the request is for preparing Federated Learning or executing Federated Learning. - [OPTIONAL] ML Model Accuracy Check Flag: identifies that the request is for using the local training data as the testing dataset to calculate the Model Accuracy of the global ML Model provided by the NWDAF service consumer acting as the FL Server NWDAF. - [OPTIONAL] ML Correlation ID: identifies the Federated Learning procedure for training the ML Model. This parameter is included when the service is used for Federated Learning. - [OPTIONAL] Data Availability requirement. This is the requirement on data availability for the ML Model training. e.g. FL Server NWDAF sends the requirement in preparation request to a FL Client NWDAF for selecting the FL Client NWDAF which can meet the data availability requirement. The following may be included: - Event ID list to be collected for local model training. - Dataset statistical properties as defined in clause 6.1.3. - Time window of the data samples. - Minimum number of data samples. - [OPTIONAL] FL Availability time requirement. This is the requirement on availability time for the ML Model training, e.g. FL Server NWDAF sends the requirement in preparation request to FL Client NWDAF for selecting the FL Client NWDAF which is available in the required time for training ML Model. - [OPTIONAL] Training Filter Information: enables to select which data for the ML Model training is requested, e.g. S-NSSAI, Area of Interest. Parameter types in the Training Filter Information are the same as or subset of parameter types in the ML Model Filter Information which are defined in clause 6.2A.2. - [OPTIONAL] Target of Training Reporting: indicates the object(s) for which data for ML Model training is requested, i.e. group of UEs identified by a list of Internal-Group-Ids or any UE (i.e. all UEs). - [OPTIONAL] Use case context: indicates the context of use of ML Model. - [OPTIONAL] Training Reporting Information with the following parameters: - Maximum response time: indicates maximum time for waiting notifications (i.e. model training results). - [OPTIONAL] Iteration round ID: indicates the iteration round number of current ML Model training. - [OPTIONAL] Expiry time. - [OPTIONAL] Indication of skipping the current FL round. The NWDAF containing MTLF provides to the consumer of the ML Model training service operations as described in clause 7.10 and clause 7.11, the output information in notification or response as listed below: - (Only for Nnwdaf_MLModelTraining_Notify) The Notification Correlation Information. - [OPTIONAL] ML Model Information (as defined in clause 6.2A.2). - [OPTIONAL] ML Model identifier: identifies the provisioned ML Model. - [OPTIONAL] Global ML Model Accuracy information: The model metric value of the global ML Model and optionally the used metric, which is calculate by the FL Client NWDAF using the local training data as the testing dataset. [OPTIONAL] Status report of FL training: Accuracy information of local model and Training Input Data Information (e.g. areas covered by the data set, sampling ratio, maximum/minimum of value of each dimension, etc.), which are generated by the FL Client NWDAF during FL procedure. NOTE 2: The parameters in Training Input Data Information are up to the implementation. - [OPTIONAL] ML Correlation ID. This parameter may be included when the service is used for Federated Learning. - [OPTIONAL] Iteration round ID: indicates the iteration round number of ML Model training indicated by the FL Server NWDAF. - [OPTIONAL] Delay Event Notification with the following parameters: - delay event indication: this parameter indicates that FL Client NWDAF is not able to complete the training of the interim local ML Model within the maximum response time provided by the FL Server NWDAF. - [OPTIONAL] cause code (e.g. local ML Model training failure, more time necessary for local ML Model training, etc.). - [OPTIONAL] Expected time to complete the training: Indicates to the FL Server NWDAF that expected remaining training time and may be provided with Delay Event Notification. 6.2F.3 ML Model Training Information Request The procedure in Figure 6.2F.3-1 is used by an NWDAF service consumer, i.e., an NWDAF containing MTLF to request another NWDAF, i.e., an NWDAF containing MTLF, for the information about ML Model training based on the ML Model file or ML Model information as described in clause 6.2F.2 provided by the NWDAF service consumer. The service may be used by an NWDAF containing MTLF to enable e.g. Federated Learning. Figure 6.2F.3-1: Procedure for ML Model Training Information Request 1. The NWDAF service consumer may request the NWDAF containing MTLF to get the information about the ML Model training based on the ML Model file or ML Model information as described in clause 6.2F.2 provided by the service consumer by invoking the Nnwdaf_MLModelTrainingInfo_Request service operation. The parameters that can be provided by the NWDAF service consumer are listed in clause 6.2F.2. In order to enable Federated Learning, NWDAF Service consumer acting as FL Server NWDAF requests to get ML Model Training Information from multiple NWDAF containing MTLF acting as FL Client NWDAFs, which are selected by the FL Server NWDAF. The details are specified in clause 6.2C. The NWDAF service consumer may use the request to check if an NWDAF can meet the ML Model training requirements (e.g. ML Model Interoperability information, Analytics ID, Service Area/DNAI and/or availability of data and time). In such cases, the NWDAF service consumer includes an ML Preparation Flag. The NWDAF service consumer may use the request to get the Model Accuracy of the ML Model provided by the service consumer using local training data in the NWDAF containing MTLF as the testing dataset. In such cases, the service consumer NWDAF includes a Model Accuracy Check Flag. 2. When the ML Preparation Flag is present in the request, the NWDAF containing MTLF only checks whether it can meet the ML Model training requirement and/or can successfully download the model if the model information is provided. Based on the check result, the NWDAF containing MTLF gets a successful return code or failure cause code (e.g. NWDAF does not meet the ML training requirements) as the information about the ML Model training. When the Model Accuracy Check Flag is present in the request, the NWDAF containing MTLF uses the local training data as the testing dataset to calculate the Model Accuracy information of the ML Model provided by the service consumer. The NWDAF containing MTLF includes the Model Accuracy information into the information about the ML Model training. When the NWDAF containing MTLF is ongoing ML Model training based on the ML Model file or ML Model information as described in clause 6.2F.2 provided by the NWDAF service consumer, the NWDAF containing MTLF gets a failure cause code (e.g. ML training is not complete) as the information about the ML Model training. When the NWDAF containing MTLF completes ML Model training based on the ML Model file or ML Model information as described in clause 6.2F.2 provided by the NWDAF service consumer, the NWDAF containing MTLF gets a successful return code and the ML Model Information of the trained ML Model as the information about the ML Model training. 3. The NWDAF containing MTLF replies to the NWDAF service consumer with the information about the ML Model training by invoking the Nnwdaf_MLModelTrainingInfo_Request response service operation. 6.2G Void 6.2H Vertical Federated Learning among NWDAFs and AFs 6.2H.1 General This clause specifies procedures for Vertical Federated learning where AFs and NWDAF can can either act as VFL servers or VFL clients. Procedures for registration and discovery, for VFL training preparation, for VFL training, and for VFL inference are covered. Both the VFL server and VFL client store the VFL model after finishing the VFL training process and use the same VFL local model to perform the VFL inference later based on the VFL correlation ID. The differences between the VFL training and inference are that for the inference there is no check of the labels, nor any server intermediate model training information are sent and as such only client intermediate results are sent to the server. 6.2H.2 Procedures 6.2H.2.1 Registration and Discovery procedure for Vertical Federated Learning 6.2H.2.1.1 Registration and Discovery procedure for Vertical Federated Learning when NWDAF or trusted AF is acting as the VFL server Figure 6.2H.2.1.1-1: Registration and Discovery procedure for Vertical Federated Learning when NWDAF or trusted AF is acting as VFL server Steps 1 to 3 are the NWDAF and AF Registration procedures when the VFL server is NWDAF or a trusted AF. 1a. VFL Server NWDAF/trusted AF registers to NRF with its NF profile, which includes NF Type (i.e. NWDAF type or AF type), Analytics ID(s), service area if available, VFL capability information per analytics ID and VFL interoperability indicator(s) and optional supported feature ID(s) and other parameters, as described in clause 5.2. 1b. NWDAF as VFL client registers to NRF with its NF profile, which includes NF Type (i.e. NWDAF type), Analytics ID(s), service area if available, VFL capability information per analytics ID and VFL interoperability indicator(s) and optional supported feature ID(s), as defined in clause 5.2. 1c. When untrusted AF as VFL client, it shall register to the NEF via OAM configuration: Analyics ID(s) and its VFL capability information per supported analytics ID and VFL interoperability indicator(s) and optional supported feature ID(s) and other parameters, as defined in clause 5.5. Then NEF updates NEF profile to NRF including associated AF ID and AF's parameters described in clause 5.5. 1d. When trusted AF as VFL client, it registers to NRF with its NF profile, which includes NF Type (i.e. AF type), analytics ID(s), service area if available, VFL capability information per analytics ID and VFL interoperability indicator and optional supported feature IDs and other parameters, as defined in clause 5.5. 2. The NRF receives the registrations from VFL server and VFL client(s), and stores their NF profile. 3. The NRF sends registration response to VFL server and VFL client(s). Steps 4 to 6 are the NWDAF, AF and NEF Discovery procedures when the VFL server is NWDAF or trusted AF. 4-6. The VFL server determines that the ML Model requires VFL based on e.g. operator policy, Analytics ID, VFL interoperability indicator(s) and Service Area. NOTE: Step 4 in Figure 6.2H.2.1.1-1 may be triggered at the VFL Server by itself or a request from a consumer. If an NWDAF can not perform as VFL Server, it first discovers and selects another VFL Server from NRF by invoking the Nnrf_NFDiscovery_Request service operation. The following criteria might be used: Analytics ID, VFL capability type as VFL server, Time Period of Interest and optional Service Area. Once the VFL Server is determined, the VFL Server discovers other NWDAF(s) and/or AF(s) as VFL Client from NRF by invoking the Nnrf_NFDiscovery_Request service operation. The following criteria might be used: NF type(s) (i.e. NWDAF type, AF type, or NEF type), Analytics ID, VFL capability type (i.e.VFL client), VFL interoperability indicator(s), Time Period of Interest, optional feature ID(s) and optional Service Area. The AF(s) discovered by the VFL Server may be trusted and/or untrusted, the NF type may be AF type when the AF(s) as VFL Client are trusted, and the NF type may be NEF type when the AF(s) as VFL Client are untrusted and the NF type may contain both AF type and NEF type if both trusted AF and untrusted AF are involved as VFL clients. When an AF is acting as a VFL server, only NWDAF(s) are discovered as VFL clients. 6.2H.2.1.2 Registration and Discovery procedure for Vertical Federated Learning when untrusted AF is acting as the VFL server The procedure below shows registration and discovery for VFL training and inference when the untrusted AF is the VFL server. There can be multiple NWDAFs as VFL clients. NOTE 1: For a deployment scenario that untrusted AF is the VFL server and only one NWDAF will be an VFL client, it is assumed that VFL client information for an Analytic ID is configured in NEF and step 6 and 7 may be skipped in Figure 6.2H.2.1.2-1. Figure 6.2H.2.1.2-1: Registration and Discovery procedure for Vertical Federated Learning when an untrusted AF is acting as VFL server and NWDAF(s) are the VFL clients Steps 1 to 3 are the NWDAF and AF Registration procedures when the VFL server is untrusted AF. 1. Same as the step 1b, in clause 6.2H.2.1.1, NWDAF as VFL client registers to NRF with its NF profile, it may include those analytics IDs on which it supports to do VFL with AF as VFL server. NOTE 2: The AF can use non-standardized values of the Analytics ID. The non-standardized values can be used by the AF as the VFL server to initiate the VFL training or VFL inference and these values need to be supported by NWDAFs acting as VFL clients. It is the operator´s responsibility to guarantee that non-standardized Analytics ID values within a PLMN are unique. 2-3. Same as the steps 2-3, in clause 6.2H.2.1.1. Based on operator local policies, the AF may register its capability as VFL server for an Analyticcs ID to NEF via OAM configuration, this is used to make it possible for NWDAF to trigger the AF to train a model. The AF may also register an AnalyticsID in NRF e.g. when a trained model is available. Steps 4-10 are the NWDAF Discovery procedures when the VFL server is an untrusted AF. 4. Untrusted AF acting as the VFL server determines that VFL operations are required and the NWDAF(s) as VFL client(s) are required. The AF sends a Nnef_NFDiscovery_Request for the VFL client(s) to the NEF and for each client provides selection criteria: Analytics ID, required NF type (i.e. NWDAF type), VFL capability type (i.e. VFL client), VFL interoperability indicator(s), Time Period of Interest, optional required feature ID(s), and optional Service Area. NOTE 3: Step 4 in Figure 6.2H.2.1.2-1 may be triggered at the VFL Server by itself or when requested from a consumer. 5. The NEF checks based on configured policies whether the AF is entitled to request single or multiple a VFL client(s) for the Analytics ID. 6-7. The NEF discovers VFL client(s) (i.e. NWDAF(s)) on behalf of the AF from the NRF by invoking the Nnrf_NFDiscovery_Request using the selection criteria provided by the AF as defined in step 4. 8. The NEF selects NWDAF(s), e.g. by using the NWDAF(s) NF profile and parameters including load/priority/capacity, that are capable of acting as VFL client(s) and matching the received selection criteria, such as location information. The NEF anonymizes NWDAF instances ID(s) and assigns external NWDAF ID(s) for each selected NWDAF instance as VFL client. The NEF stores the external NWDAF ID(s) together with information how to reach the NWDAFs. NOTE 4: The external NWDAF ID assigned by NEF is temporary and can be released if needed. 9. The NEF sends Nnef_NFDiscovery_Request response only including the external NWDAF ID(s) as selected in step 8 and information how they relate to corresponding selection criteria provided by the AF, VFL interoperability indicator(s), Time Period of Interest, optional feature ID(s) and optional Service Area to the untrusted AF. 10. The AF stores the received external NWDAF ID(s) and uses it subsequent interactions with the NEF to indicate the target VFL client. NOTE 5: If the VFL client NWDAFs are statically preconfigured to the NEF and untrusted AF and not released, the discovery procedure can be skipped. The AF may indicate to the NEF that the AF will no longer use the VFL process identified by VFL Correlation ID (external NWDAF IDs are no longer required). If the NEF receives this indication, it shall remove the stored external NWDAF ID(s) associated with the VFL correlation ID allocated to the AF. Then, the NEF responds to the AF. 6.2H.2.2 Preparation procedure for Vertical Federated Learning 6.2H.2.2.0 General The preparation procedure is used to check if the VFL Client(s) can meet the ML Model training requirement. The procedure includes the negotiation, between server and client(s) to enable interoperability, sample alignment and may include feature negotiation if the VFL Server did not learn the supported FeatureIDs from each VFL Client using the discovery phase, alternatively the VFL Server may know the supported FeatureIDs by a VFL Client based on configuration. The Vertical Federated Learning preparation procedure can be skipped if the VFL Server can decide which VFL Client(s) support the VFL procedure to be performed, e.g. based on local configuration or offline procedures. 6.2H.2.2.1 Preparation procedure for Vertical Federated Learning when NWDAF/trusted AF is the VFL Server Figure 6.2H.2.2-1: Preparation procedure for Vertical Federated Learning when NWDAF/Trusted AF is the VFL Server Editor´s note: For UEs as samples, additional discussion is needed on whether UE needs to be registered or not and whether the NWDAF as VFL server can check whether UEs are registered before interacting with VFL clients. 1. An NWDAF as VFL Server may send a Vertical Federated Learning preparation request including the Analytics ID to each of the NWDAF VFL Client(s), using Nnwdaf_VFLTraining_Request and to each of the AF VFL Clients(s), using Naf_VFLTraining_Request possibly via NEF when the VFL Client is an untrusted AF. An AF as VFL Server may send a Vertical Federated Learning preparation request including the Analytics ID to each of the NWDAF VFL Client(s), using Nnwdaf_VFLTraining_Request. The NWDAF or trusted AF as a VFL Server may also provide, the suggested VFL Interoperability Information to negotiate the intermediate results that will be used in VFL process (e.g, VFL training and/or VFL inference), the suggested list of sample IDs (e.g, SUPI(s), GPSI(s)) expected to be used in the VFL process. Optionally the feature ID(s) together with VFL Interoperability indicator expected to be used in VFL process for each VFL Client, and as additional criteria for sample alignment, time window of the data samples, and required minimum sample size. When a Trusted AF is acting as a VFL Server, the VFL Client can only be an NWDAF. NOTE: In step 1, the feature ID(s) will only be provided to the VFL Client(s) when the VFL Server received the supported feature ID(s) of VFL Client(s) during the VFL Client(s) discovery procedure as defined in clause 6.2H.2.1. 2. Each VFL Client checks if it can meet the ML Model training requirement. Each VFL Client checks the list of sample IDs and required criteria for sample alignment suggested by the VFL Server and then provides to the VFL Server the list of sample IDs that it can accept within the sample IDs suggested by the VFL Server and satisfying the required criteria for sample alignment. If the suggested list of sample IDs is not provided by the VFL Server, the VFL Client provides the list of supported sample IDs to the VFL Server. Each VFL Client checks the VFL Interoperability Information and determines which VFL Interoperability information that the VFL Client supports. The VFL Client determines whether it supports the feature ID(s) that will be used in VFL process when the feature ID(s) was provided by the VFL Server; alternatively, the VFL Clients may provide the list of supported feature ID(s), which is associated to the VFL Interoperability information, to the VFL Server, or VFL Server may know the supported featureID(s) for a VFL Client based on configuration. 3. Each NWDAF VFL Client invokes Nnwdaf_VFLTraining_Response or and each AF VFL Client invokes Naf_VFLTrainingRequest_Response, possibly via NEF when the AF is untrusted, to indicate to the VFL Server whether it accepts the ML Model training requirements, the VFL Client can also indicate that it cannot join the FL process. When NEF is involved in the procedure, if the NEF receives the SUPI as sample ID(s) from NWDAF, the NEF shall map the SUPI(s) to GPSI(s) before sending to the AF. 4. The VFL Server determines the final list of samples ID(s) that all selected VFL Client(s) support as the samples to be used in the VFL process. If the VFL Server is NWDAF, it may map the sample ID(s) in different types (e.g, SUPI, GPSI) from the VFL Client(s) into same format (i.e SUPI) as described in clause 6.2.8.2.4.4 before determining the final list of sample ID(s). The VFL Server also determines the feature ID(s) per VFL Client and VFL Interoperability Information to be used in VFL process and provides them to the selected VFL Client(s) at the start of the training phase, as described in clause 6.2H.2.3.1. 6.2H.2.2.2 Preparation procedure for Vertical Federated Learning when untrusted AF is the VFL server This clause specifies the preparation (including sample alignment) procedure for untrusted AF-initiated VFL scenarios between AF and NWDAF(s) within a single PLMN. Figure 6.2H.2.2-1: Preparation procedure for Vertical Federated Learning when untrusted AF is the VFL Server Editor´s note: For the VFL preparation procedure, whether and how NEF does pre-work of sample IDs intersection before the VFL server determines the final sample IDs is FFS. The untrusted AF as VFL server first determines the VFL client(s) of VFL process(es), NEF ID and external NWDAF ID(s) of all selected VFL Client(s) for each VFL process and uses it subsequent interactions with the NEF. 1- The untrusted AF as VFL server sends VFL preparation requests to each of the candidate VFL client(s), using Nnef_VFLTrainingRequest_Request to check if the VFL client(s) can meet the ML Model training requirement (which includes the Analytics ID, list of sample IDs, and optionally Feature ID, etc.). The external NWDAF IDs obtained in the discovery procedure (see clause 6.2H.2.1.1) is included to indicate the target NWDAFs. The suggested VFL Interoperability Information to negotiate the intermediate resuls that will be used in training. 2- The NEF maps the external NWDAF and GPSI(s) to the internal NWDAF and SUPI(s). The NEF send VFL preparation request to the corresponding candidate internal NWDAF ID using Nnwdaf_VFLTraining_Request service with the same information as provided in step 1 in clause 6.2H.2.2.1. 3- Same as step 2 in clause 6.2H.2.2-1. 4- Same as step 3 in clause 6.2H.2.2-1. 5- The NEF maps the internal NWDAF and SUPI(s) to the external NWDAF and GPSI(s). The NEF sends the Nnef_VFLTraining_Request Response to the VFL Server with the same information as provided in step 4. 6- Same as step 4 in clause 6.2H.2.2-1. 6.2H.2.3 Training Procedure for Vertical Federated Learning 6.2H.2.3.1 Training Procedure for Vertical Federated Learning when NWDAF or trusted AF is acting as VFL server The figure 6.2H.2.3.1-1 below shows the training procedure for Vertical Federated Learning when NWDAF is acting as VFL server. Figure 6.2H.2.3.1-1: Training procedure for Vertical Federated Learning when NWDAF or trusted AF is acting as VFL server Editor's note: How the NEF assists the VFL training process as well as whether the service operations going via NEF is using the existing or new service operation are FFS. 0. [OPTIONAL] VFL training is triggered in the VFL server for the following cases: - Based on local configuration and agreement among vendors and/or application providers participating in the same group for specific VFL task(s). - Triggered by either Analytics consumer, AnLF via ML model provisioning or MTLF internal (VFL Server is a trusted AF): Step 0a (case A). If the NWDAF containing AnLF that wants to perform inference and does not have a model, it discovers an NWDAF containing MTLF from NRF and sends a subscription request for a model to the NWDAF containing MTLF using Nnwdaf_MLModelProvision_Subscribe. Step 0b. (case A). If the discovered NWDAF containing MTLF decides to use VFL for the subscription and realizes it cannot be VFL server but wants an AF to act as VFL server, the NWDAF containing MTLF discovers an AF as VFL server according to 6.2H.2.1, and may send a request to the VFL server AF using Naf_Training_Subscribe including Analytics ID, optionally Notification target address, the AF as VFL server starts VFL Training according to step 2. The NWDAF containing MTLF may either send in the service subscription response that no ML model available due to VFL model to be used, the NWDAF containing AnLF may request the inference output using the requested Analytics ID as described in step 1 of clause 6.2H.2.4.1. Alternatively, the NWDAF containing MTLF may indicate that training is ongoing, then step 1 follows. NOTE 1: The AF may have registered in NRF for the particular Analytics ID that it can perform Inference for it. In this case no training will be performed. - Triggered by either Analytics consumer, AnLF via ML model provisioning (VFL Server is a trusted AF): Step 0c. (case B). Same as step 0a, in addition if the discovered NWDAF containing MTLF decides to use VFL for the subscription and it can be VFL server, the VFL server starts VFL Training according to step 2. If decision to start VFL training is triggered on ML model provisioning from an NWDAF in step 0b or 0c, the VFL server sends in the service subscription response that no ML model available due to VFL model to be used and that the subscription is terminated (i.e. as new cause code). The VFL server may send training is done to the first NWDAF containing AnLF. When VFL training is done, the first NWDAF contain AnLF will be needing to perform VFL Inference using the requested Analytics ID to retrieve an VFL Inference output from the VFL Server (AF or NWDAF) using the requested Analytics ID as described in step 1 of clause 6.2H.2.4.1. - Triggered by either Analytics consumer, AnLF via ML model provisioning (VFL Server is a trusted AF): Step 0d. (case C). If the NWDAF containing AnLF that wants to perform inference and does not have a model, it discovers an NWDAF containing MTLF from NRF and sends a subscription request for a model to the NWDAF containing MTLF using Nnwdaf_MLModelProvision_Subscribe. If the discovered NWDAF containing MTLF does not have a model, it sends a response to the AnLF that no Model is available due to VFL used and that the subscription is terminated (i.e. as new cause code), and may also provide VFL server ID. If no VFL Server ID is sent, the NWDAF it discovers VLF Server from NRF that supports the Analytics ID, then requests to train a ML Model to the VFL Server. Alternatively, the NWDAF containing AnLF knows by configuration that the AnalyticsID is trained using VFL, then the AnLF discovers the VFL Server and requests the VFL Server to train the ML Model. - Triggered by analytics consumer. (VFL Server is either an AF or an NWDAF): Step 0e. (case 5). NF consumer needs analytics output for a specific Analytics ID, it discovers an NWDAF from NRF to provide the analytics ID. If the discovered NWDAF is a VFL server which supports the analytics ID, then to generate the analytics output via VFL inference as defined in clause 6.2H.2.4, the VFL server may trigger VFL training if no corresponding VFL training has been performed. NOTE 2: In the case when AF is server the clients can only be NWDAFs. NOTE 3: VFL server can also decide to initiate VFL training based on operator policy or internal configuration. 1. The NWDAF acting as VFL server determines the VFL clients that participate in VFL procedure in the VFL clients discovery and preparation phase as described in the clause 6.2H.2.1 and clause 6.2H.2.2. NOTE 4: VFL Server can determine to start the training based on local configuration and agreement among vendors and/or application providers participating in the same group for specific VFL task(s). Steps 2-6 are repeated until the training termination condition is reached. 2. To start the VFL training, the VFL server sends a request to start the training to all selected VFL clients. The request includes VFL correlation ID, at least the parameters negotiated during the preparation phase and theVFL training iteration number set to 0, and a Notification Correlation ID. Optionally, the VFL Server includes parameters according to clause 6.2H.3. If the VFL procedure continues in subsequent iterations, the VFL server sends a request for a new VFL training iteration containing an incremented a VFL training iteration number and intermediate model training information to each of the VFL clients for next round of VFL training. The VFL Server, based on internal logic, may provide checkpoint information according to clause 6.2H.3. 2a. The VFL server sends a Nnwdaf_VFLTraining_Subscribe to the selected NWDAF VFL clients(s). 2b. The VFL server sends a Naf_VFLTraining_Subscribe to the selected trusted AF VFL clients(s). 2c. For each selected untrusted AF VFL clients, the VFL server sends a Nnef_VFLTraining_AFClient_Subscribe to the NEF handling that AF. 2d. For each selected untrusted AF VFL client, the NEF sends a Naf_VFLTraining_Subscribe to that AF. The NEF may also translate the analytic filter information if needed, e.g. TAIs into geographical area. 3. [Optional] Each VFL client collects its local data by using the current mechanism if the VFL client has no local data already available. The data used by each VFL Client is collected as per alignment information. 4. During VFL training procedure, each VFL client further trains the local ML model associated with the same VFL Correlation ID based on their own collected or available data and when applicable (e.g. after the first round of training) and possible intermediate model training information distributed by the VFL server in the previous training iteration. Each VFL Client computes and reports the client intermediate training result of the local ML model to the VFL server. VFL client(s) may also report a delta from the initial list of samples according to clause 6.2H.3. Based on internal logic, the VFL server may indicate if the training has to resume from a previous checkpoint by using its training iteration round ID. NOTE 5: The intermediate model training information and intermediate training result are constructed in per sample granularity. NOTE 6: The contents of intermediate training result depend on the type of ML Model or algorithm used in VFL training and are up to implementation. 5. Each VFL client reports the computed client intermediate training result of the local ML model to the VFL server. The Notification Correlation ID and a VFL training iteration number. A client may indicate in message for a request to leave the VFL. NOTE 7: If the VFL Server deems the training can continue, it responds back by informing the FL Client to cease the ML Model training by performing step 8 for this client. 5a. A NWDAF VFL client sends a Nnwdaf_VFLTraining_Notify. 5b. A trusted AF VFL client sends a Naf_VFLTraining_Notify to the VFL server. 5c. An untrusted AF VFL client sends a Naf_VFLTraining_Notify to the NEF. 5d. For each untrusted AF VFL client, the NEF converts any external identifiers to internal identifiers and sends a Nnef_VFLTraining_AFClient_Notify to the VFL server. 6. The VFL server may collect the local data and generate its own local intermediate training result. The VFL Server computes the intermediate model training information (e.g. gradient information or loss information that may contain loss function or loss value) based on the VFL Client(s) intermediate training result(s) received in step 4, its own local intermediate results and the label. The intermediate model training information is used for updating the models of VFL clients and/or the model of the VFL Server. Different intermediate model training information may be computed for different VFL clients and for the VFL Server itself. NOTE 8: The contents of loss information and gradient information depend on the type of ML Model or algorithm used in VFL training and are determined by implementation. The VFL server may locally compute contribution weights for each VFL client participating in the VFL pocess and store them for subsequent use, e.g. during inference. How the VFL computes the VFL contribution weights is up to implementation. The VFL server may also decide to compute the global ML model metric (e.g. ML model accuracy) based on all the intermediate training result received from VFL clients and the label at any time of VFL training. When determining to generate the ML model metric, the VFL server includes an indication to indicate to the VFL clients to use the sample ID of the dataset for accuracy monitoring to calculate intermediate training results in Nnwdaf_VFLTraining_Subscribe service operation. Corresponding to the sample IDs for accuracy monitoring, each VFL client collects local data as input data for local ML model trained in step 4, then the VFL client generates the intermediate training result based on the trained local ML model and the input data, and it provides the intermediate training result to VFL server. The VFL server computes the ML model metric (e.g. VFL accuracy) based on the received intermediate training result and the label, which are both corresponding to the sample IDs for accuracy monitoring. The VFL server may optionally take account the local ML model accuracy monitoring information received from VFL client(s) when computing global ML Model metric. Editor's note: Whether weight of the VFL Client is computed by VFL server is FFS. Editor's note: Whether VFL server and VFL clients share feature information is FFS. 7. [Optional] The VFL server evaluates (e.g. based on the convergence of a loss function or loss value and/or if the pre-set iteration number is reached and/or global ML model metric is stable) whether VFL Training process converged. If the VFL Server evaluates the VFL Training process did not converge, the VFL Server determines another round of VFL training is required and repeats step 2 - 6. If the VFL Server evaluates the VFL training process converged, it determines the VFL Training is completed. In this case, the VFL Server terminates the current VFL training process via step 7. The VFL training termination decision may be also made as follows: Step 7b, 7c, 7f and 7h. Based on the consumer request, the VFL server notifies VFL status report to the consumer, to update the metric value to the consumer periodically (e.g. a certain number of training rounds or at fixed periods) or dynamically when some pre-determined status is achieved (e.g. the ML Model Accuracy threshold is achieved or training time expires).The status report may include global model metric (e.g. ML model accuracy). If NWDAF including MTLF has forwarded a subscription request to VFL server AF in step 0b, and it did not add Notification target set to the first NWDAF containing AnLF, it adds the AF ID, in the Notification to the first NWDAF containing AnLF. Steps 7d, 7e, 7g and 7i. The consumer decides whether the current model can fulfil the requirement, e.g. ML model metric is satisfactory for the consumer and determines to either unsubscribe or continue the training process. Based on the subscription request sent from the consumer, the VFL server updates or terminates the current VFL training process. 8. The VFL server sends VFL training termination message to VFL Client if it decides to terminate the VFL training process, the termination message contains VFL Correlation ID and may contain the intermediate model training information to each of the VFL clients. 8a. The VFL server sends a Nnwdaf_VFLTraining_Unsubscribe t to the selected NWDAF VFL clients(s). 8b. The VFL server sends a Naf_VFLTraining_Unsubscribe to the selected trusted AF VFL clients(s). 8c. For each selected untrusted AF VFL clients, the VFL server sends a Nnef_VFLTraining_AFClient_Unubscribe to the NEF handling that AF. 8d. For each selected untrusted AF VFL clients, the NEF sends a Naf_VFLTraining_Unsubscribe to that AF. 9. The VFL Server, stores VFL correlation ID, the local trained ML Model, the mapping information of the VFL correlation ID to the following parameters: Analytics ID related to the VFL training process, locally trained Model. Additionally, the VFL server stores the VFL client information (including the NF ID of the VFL Client), which may be used to determine associated VFL client in the VFL inference. Each VFL client updates local ML model based on the intermediate model training information, if received and stores VFL correlation ID, the locally trained ML Model, the mapping information of the VFL correlation ID to locally trained Model NOTE 9: The VFL correlation ID and the stored mapping information are used later for inference as described in Clause 6.2H.2.4.1. NOTE 10: If untrusted AF is involved in VFL Clients, the message between NWDAF acting as VFL Server and the untrusted AF is via NEF. 6.2H.2.3.2 Training Procedure for Vertical Federated Learning untrusted AF is acting as VFL server Figure 6.2H.2.3.2-1: Training procedure for Vertical Federated Learning when untrusted AF is acting as VFL server Editor's note: The ENs listed in clause 6.2H.2.3.1 are also applied to this clause. 0. [CONDITIONAL] Same as in step 0 in clause 6.2H.2.3.1, when the AF is the VFL Server, but AF is replaced with untrusted AF, and Nnef_Training_Subscribe offered by NEF is used. NEF forwards the subscription request to AF using Naf_Training_Subscribe. In addition: - Either based on the information received or internal configuration, VFL server decides to initiate VFL training procedure, or - case E) same as step 0d, but the Analytics consumer contacts the NWDAF containing AnLF that wants to perform inference and does not have a model, it discovers VLF Server from NRF that is a VFL Server, then request the VFL Server via NEF to perform inference. The VFL server may trigger VFL training if no corresponding VFL training has been performed. 1. Same as step 1 in Figure 6.2H.2.3.1-1. Steps 2-7 are repeated until the training termination condition is reached. 2. To start VFL training, the VFL server do same as in step 2 in Figure 6.2H.2.3.1-1, using Nnef_VFLTraining_Subcribe. 2c. If a NWDAF VFL client is selected to aggregate the intermediate training results of other VFL client clients, thisNWDAF VFL client may send Nnwdaf_VFLTraining_Subscribe to other one or more indirect NWDAF VFL client as configured by the VFL server from which it desires to receive the client intermediate training results. NOTE: To support the intermediate results sharing between VFL clients in step 2c, the VFL clients might be selected based on the VFL Interoperability Indicators and/or the parameters in VFL Interoperability Information (e.g. gradient dimension of local model, split point of the preconfigured initial model, etc.). 3. [Optional] Same as step 3 in Figure 6.2H.2.3.1-1. 4. Same as step 4 in Figure 6.2H.2.3.1-1. 5. Same as step 5 in Figure 6.2H.2.3.1-1. 5a. A NWDAF VFL client sends a Nnwdaf_VFLTraining_Notify. 5b. For an untrusted AF acting as VFL server, the NEF converts any internal identifiers to external identifiers, provides the external NWDAF ID and sends a Nnef_VFLTrainingNotify to the VFL server. 5c-5d. The NWDAF VFL clients share the client intermediate training results with the NWDAF VFL client from which it received subscription request in step 2c. This NWDAF VFL client aggregates the received client intermediate training results from the NWDAF VFL clients, performs local computation using the aggregated intermediate inference results, then sends one notification to the NEF by including its client intermediate training result. 6. [Optional] Same as step 6 in Figure 6.2H.2.3.1-1. 7. Same as step 7 in Figure 6.2H.2.3.1-1. Same as in step 0 in clause 6.2H.2.3.1, when the AF is the VFL Server, but AF is replaced with untrusted AF, and Nnef_Training_Notify offered by NEF is used. NEF forwards the subscription request to AF using Naf_Training_Notify. 8. Same as step 8 of Figure 6.2H.2.3.1-1. However, sub steps in that figure are not applicable. 8a. For each NWDAF VFL client, the untrusted AF as VFL server sends a Nnef_VFLTraining Unsubscribe to the NEF handling that AF. The untrusted AF identifies the VFL client using the external NWDAF ID assigned in the discovery procedure (see clause 6.2H.2.1.1). 8b. The NEF sends an Nnwdaf_VFLTraining_Unsubscribe to the NWDAF VFL client indicated by the received external NWDAF ID. 8c. An NWDAF VFL client may send Nnwdaf_VFLTraining_Unsubscribe to other one or more NWDAF VFL clients in step 2c. 9. Same as step 9 of Figure 6.2H.2.3.1-1. 6.2H.2.4 Inference procedure for vertical federated learning 6.2H.2.4.1 Inference procedure for vertical federated learning when NWDAF or Trusted AF is acting as VFL server Figure 6.2H.2.4.2-1: Inference procedure for vertical federated learning when NWDAF or Trusted AF is acting as VFL server The inference procedure when trusted AF is acting as VFL server may be triggered by a request or subscription from a 5GC consumer NF or internal service logic of the AF acting as VFL server. If triggered by internal service logic of the AF acting as VFL server, the steps 0, 1,7 and 8 are skipped. 0. The analytics consumer NF sends an Analytics request/subscribe (Analytics ID, Target of Analytics Reporting= e.g. UE IDs and optionally both Analytics Reporting Information=Analytics target period and Analytics Filter, Analytics Accuracy Request if the analytics consumer NF requests Analytics Accuracy Monitoring) to NWDAF containing AnLF by invoking a Nnwdaf_AnalyticsInfo_Request or a Nnwdaf_AnalyticsSubscription_Subscribe., providing parameters as defined in clause 6.1.3. 1. If the NWDAF containing AnLF can be the VFL server to generate the VFL inference results for the requested analytics ID, then step 1 is skipped. If the NWDAF containing AnLF can not generate the analytics output, the NWDAF containing AnLF determines the VFL Server AF for the requested analytics. If the VFL server is trusted AF, the NWDAF containing AnLF sends a request to trusted AF VFL server using Naf_Inference_subscribe/request including Analytics ID, Target of Analytics Reporting = e.g. UE IDs and optionally Analytics Reporting Information=Analytics target period Analytics Filter and Analytics Accuracy Request. 2. Based on the information received in the step 0 or 1, VFL server decides to initiate the VFL inference procedure with the VFL clients. Before the VFL server initiate the VFL inference procedure, the VFL server may initiate the VFL training procedure if no VFL model is already trained as described in the clause 6.2H.2.3 based on the information received in the step 0 or 1 or VFL server local configuration. VFL Server selects clients(s) using information stored in the VFL training process. The server may select some or no clients, e.g. depending on the accuracy of the VFL model, the contribution to the training result and the current status of the VFL clients. When no VFL Clients are selected, the VFL server may generates the VFL inference results based only on its local trained ML model associated with the determined VFL correlation ID, skipping the steps 2 - 6 (and 7 if step 1 was also skipped). NOTE 1: If the server does not select VFL clients for all features, it can internally store information about the features considered when deriving the analytics results together with a timestamp to identify the analytics results, to enable to subsequently trace the origin of analytics results and to interpret possible feedback on analytics accuracy. NOTE 2: If a client is not available for inference, e.g. depending on the contribution to the training result, the server can decide to select another available VFL client with a well-trained model, or the VFL server can start a new training procedure for a new VFL client. VFL server NWDAF or trusted AF determines and sends a VFL Inference request/subscription to each VFL client including the Target of VFL inference = e.g. UE IDs, VFL correlation ID to indicate the VFL client which previously well-trained VFL local model associated with this ID will be used and optionally VFL inference filter. Data time window: Time when intermediate local result is needed. Editor's note: It is FFS whether additional parameters are needed to send from the VFL server to VFL client, e.g. parameters used in training phase and parameters from Analytics request. 2a. For each NWDAF VFL client, the VFL Server NWDAF or trusted AF sends an Nnwdaf_VFLInference_Subscribe or Nnwdaf_VFLInference_Request to the VFL client. 2b. For each trusted AF VFL client, the VFL Server NWDAF sends an Naf_VFLInference_Subscribe or Naf_VFLInference_Request to the VFL client. 2c. For each untrusted AF VFL client, the VFL Server NWDAF sends an Nnef_VFLInference_Subscribe or Nnef_VFLInference_Request to the NEF serving the AF. 2d. For each untrusted AF VFL client, the NEF converts any internal identifiers to external identifiers and sends an Naf_VFLInference_Subscribe or Naf_VFLInference_Request to the untrusted AF VFL client. NOTE 3: For the trusted AF is acting as VFL server, the VFL client can only be the NWDAF. 3. Each VFL Client collects its local data by using the current mechanism if the VFL Client does not have local data available already. 4. Based on the VFL correlation ID, each VFL Client determines the VFL local model to generate the intermediate local inference results. NOTE 4: In this Release, it is assumed that local intermediate inference is shared between VFL server and VFL client. 5. VFL Client sends the client intermediate local results to the VFL server. The intermediate local results, which are sent from the VFL Client to the VFL Server during the VFL inference process, are the information for the VFL Server to combine and generate the VFL inference results. If the VFL server used an inference subscription in step 2, step 5 may be repeated. 5a. Each NWDAF VFL client sends an Nnwdaf_VFLInference_Notify or Nnwdaf_VFLInference_Request response to the VFL Server NWDAF or trusted AF. 5b. Each trusted AF VFL client sends a Naf_VFLInference_Notify or Naf_VFLInference_Request response to the VFL Server NWDAF. 5c. Each untrusted AF VFL client sends a Naf_VFLInference_Notify or Naf_VFLInference_Request response to the NEF. 5d. For each untrusted AF VFL client, the NEF converts any external to internal identifiers and sends an Nnef_VFLInference_Notify or Nnef_VFLInference_Request response to the NWDAF VFL server. 6. The VFL server may collect its local data and generate the intermediate local inference results. When the VFL Server selected VFL clients to participate in the VFL Inference process, it combines all the intermediate local results to generate the VFL inference results based on the VFL correlation ID. The VFL server takes into account the participation of each VFL client during the ML training process and the importance of the intermediate local results when generates the combined inference output. The VFL server may compute the VFL accuracy based on all the intermediate local results received from VFL clients and the label, if it receives Analytics accuracy request in step 0. 7. Depending on request, the NWDAF or trusted AF VFL server sends Nnwdaf_AnalyticsInfo_Response or Nnwdaf_AnalyticsSubscription_Notify or Naf_Inference_Response/Notify to the consumer (i.e. NWDAF containing AnLF) including the VFL inference results, optionally, VFL accuracy. 8. The NWDAF containing AnLF provides the analytics output to the analytics consumer NF based on the VFL inference results by means of either Nnwdaf_AnalyticsInfo_Response or Nnwdaf_AnalyticsSubscription_Notify, depending on the service used in step 0. The NWDAF containing AnLF may provide VFL accuracy if the consumer NF provides Analytics Accuracy request in step 0. NOTE 3: There may be some time delay between the time when the AnLF provides the analytics output to consumer NF and the time when the AnLF provides VFL accuracy, as the labels are collected after making predictions. The analytics result and VFL accuracy may be send in different response messages. 6.2H.2.4.2 Inference procedure for vertical federated learning when untrusted AF is acting as VFL server Figure 6.2H.2.4.2-1: Inference procedure for vertical federated learning when untrusted AF is acting as VFL server The inference procedure when untrusted AF is acting as VFL server may be triggered by a request or subscription from a 5GC consumer NF or internal service logic of the AF acting as VFL server. If triggerd by internal service logic of the AF acting as VFL server, the steps 0, 1,7 and 8 are skipped. 0. Same as step 0 in clause 6.2H.2.4.1. 1. Same as step 1 in clause 6.2H.2.4.1, to NEF using Nnef_Inference_subscribe/request. NEF forwards the subscription request to AF using Naf_Inference_subscribe/request. Editor's note: When the AnLF determine the VFL server AF is FFS. For example, during the training phase. 2. Same as step 2 in clause 6.2H.2.4.1, to NEF using Nnef_VFLInference_subscribe/request. An untrusted AF includes the external NWDAF ID and sends the request to the NEF. NEF converts any received external identifiers to internal identifiers and forwards the subscription request to NWDAF using Nnwdaf_VFLInference_subscribe/request. 2c An NWDAF VFL client may send Nnwdaf_VFLInference_Subscribe or Nnwdaf_VFLInference_Request to the VFL client to other one or more NWDAF VFL client as configured from which it desires to receive the client intermediate inference results. NOTE 1: To support the intermediate results sharing between VFL clients in step 2c, the VFL clients might be selected based on the VFL Interoperability Indicators and/or the parameters in VFL Interoperability Information (e.g. gradient dimension of local model, split point of the preconfigured initial model, etc.). 3. Same as step 3 in clause 6.2H.2.4.1. 4. Same as step 4 in clause 6.2H.2.4.1. NOTE 2: In this Release, it is assumed that local intermediate inference is shared between VFL server and VFL client. 5. Same as step 5 in clause 6.2H.2.4.1, to NEF using Nnwdaf_VFLInference_Notify/Request Response. NEF converts any received internal identifiers to external identifiers and forwards the subscription notify to the untrusted AF using Nnef_VFLInference_Notify/Request Response. 5c-5d. The NWDAF VFL clients share the client intermediate inference results with the NWDAF VFL client from which it received request in step 2c. This NWDAF VFL client aggregates the received client intermediate inference results, performs local computation using the aggregated intermediate inference results, then sends one notify message to the NEF by including its client intermediate inference result. If the VFL server used an inference subscription in step 2, step 5 may be repeated. 6. Same as step 6 in clause 6.2H.2.4.1. 7. Same as step 7 in clause 6.2H.2.4.1, to NEF using Naf_Inference_Notify/Request Response. NEF converts any received internal identifiers to external identifiers and forwards the subscription Notify to NWDAF using Nnef_Inference_Notify/Request Response. 8. Same as step 8 in clause 6.2H.2.4.1. 6.2H.3 Contents of ML Model Training service for Vertical Federated Learning The consumers of the ML Model training services may provide the input parameters in Nnwdaf_VFLTraining service or Naf_VFLTraining service or in Nnefe_VFLTraining service as listed below: - Analytics ID: identifies the analytics for which the ML Model is requested to be trained. - At start of Training phase VFL server sends VFL Correlation ID: identifies a VFL process to be executed among the candidate VFL participants. - [OPTIONAL] VFL Interoperability Information that indicates the intermediate model training information that the VFL Server supports and asks the client to verify its support for (e.g. activations, gradients, type of loss), the content of the VFL Interoperability Information is not standardized in this Release. - [OPTIONAL] In training phase, Maximum response time (i.e. the maximum time between VFL clients receive intermediate model training information and send back intermediate training result). - [OPTIONAL] Training Filter Information: enables to select which data for the ML Model training is requested, e.g. S-NSSAI, Area of Interest. Parameter types in the Training Filter Information are the same as or subset of parameter types in the ML Model Filter Information which are defined in clause 6.2A.2. - In training, the Intermediate model training information includes what the Server sends to each client which contains what has been agreed in the preparation phase. The content is not standardized in this Release. - ML model accuracy check flag: request ML model accuracy monitoring information which assists VFL server to perform ML model accuracy monitoring. - In the preparation phase, the initial list of samples selected by the VFL Server. - [OPTIONAL] At start of Training phase, sample ID of the dataset for accuracy monitoring are selected from the sample ID list decided in the preparation phase, identifying the data to be used for generating intermediate training result by VFL client and supporting the VFL Server to determine the VFL ML model accuracy. NOTE 1: How the VFL server determine the sample IDs of dataset for accuracy monitoring is up to implementation. - [OPTIONAL] In later steps in training, the VFL server may provide delta from the sample list sent in first step of training procedure to indicate which samples aligned in the preparation phase will not be part of the rest of training procedure. NOTE 2: The delta list can only remove samples from the initial sample list. - [OPTIONAL] Checkpoint information: indicates whether model status at the current iteration should be saved as a training checkpoint. The VFL client provides to the consumer of the ML Model training service operations the output information in as listed below: - VFL Interoperability information supported by each VFL Clients. - In the preparation phase, the list of samples accepted by the VFL client. - [OPTIONAL] In training the VFL client may provide delta from the sample list received in first step of training procedure to indicate which samples will not be part of the rest of training procedure. NOTE 3: The delta list can only remove samples from the initial sample list. - [OPTIONAL] In preparation phase, the Feature ID(s) that the VFL client supports. A Feature ID indicates what features the VFL client can use for an Analytics ID, the values of the Feature ID are not standardized. - [OPTIONAL] VFL Interoperability Information that indicates the Intermediate model training information that the client supports in response to what the server sent in the subscription request. - In training, the Intermediate training result includes what the Client sends to the server which contains what has been agreed in the preparation phase. The content is not standardized in this Release. - [OPTIONAL] Local ML model accuracy monitoring information. Editor's note: Which parameters are optional, or mandatory is FFS, the list of parameters also needs to be extended.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.3 Slice load level related network data analytics
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.3.1 General	The NWDAF provides slice load level information to a consumer NF on a Network Slice level or a Network Slice instance level or both. The NWDAF is not required to be aware of the current subscribers using the slice. The NWDAF notifies slice specific network status analytics information to the consumer NF that is subscribed to it. A consumer NF may collect directly slice specific network status analytics information from NWDAF. This information is not subscriber specific. The NWDAF services as defined in the clause 7.2 and clause 7.3 are used to expose slice load level analytics from the NWDAF to the consumer NF (e.g. PCF, NSSF, AMF or AF, possibly via NEF). The consumer of these analytics shall indicate in the request or subscription: - Analytics ID = "Load level information"; - Analytics Filter Information: - S-NSSAI and NSI ID; NOTE 1: The use of NSI ID in the network is optional and depends on the deployment choices of the operator. If used, the NSI ID is associated with S-NSSAI. NSI ID is only applicable when the consumer of analytics is NSSF or AMF. - optionally, the list of analytics subsets that are requested among those specified in clause 6.3.3A; - optionally, for analytics exposure in roaming case (see clause 6.1.5), the PLMN ID identifying the target PLMN (i.e. PLMN of which the roaming analytics is requested); and - optionally, for analytics exposure in roaming case (see clause 6.1.5), mapped S-NSSAI of the HPLMN if the consumer NF is in the VPLMN. NOTE 2: The terms "HPLMN" and "VPLMN" here refer to a roaming case in which at least one UE served by the NWDAF analytics consumer is involved. - an optional Area of Interest (i.e. list of TAs or cells); NOTE 3: When the AF provides geographical area, the NEF can map it into list of TAs or cells. - an optional list of NF types; - optionally, Load Level Threshold value; - optionally, "maximum number of objects" indicating the maximum number of Network Slice instances expected in output, when the Analytics Filter Information does not indicate an NSI ID; and - an Analytics target period indicating the time period over which the statistics or predictions are requested.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.3.2 Void	6.3.2A Input data The detailed information collected by the NWDAF is listed in Table 6.3.2A-1 and Table 6.3.2A-2. Table 6.3.2A-1: OAM Input data for slice load analytics Information Source Description UE registered in a Network Slice/Network Slice instance OAM Mean number of UEs registered in a NW slice or NW slice instance as defined in TS 28.552 [8]. (NOTE 1). PDU Session established on a Network Slice/Network Slice instance OAM Mean number of established PDU Sessions in a NW slice or NW slice instance as defined in TS 28.552 [8]. (NOTE 1). Load of NFs associated to Network Slice instance OAM Resource utilization information of a Network Slice instance obtained from its constituent NF instances. NF instance load input data collection is described in clause 6.5, Table 6.5.2-1. NOTE 1: 5GC performance measurements can be provided per S-NSSAI by OAM as defined in TS 28.552 [8]. Table 6.3.2A-2: 5GC NF Input data for slice load analytics Information Source Description Timestamps 5GC NF A time stamp associated with the collected information. UE registers/de-registers to a Network Slice/Network Slice instance AMF(s) AMF reports that a UE registered or deregistered to a S-NSSAI or to a S-NSSAI and NSI ID. Number of UEs served by the AMF AMF(s) AMF reports the total number of UEs served by the AMF per S-NSSAI or per S-NSSAI and NSI ID. (NOTE 1) PDU Session established/released on a Network Slice SMF(s) SMF reports that a PDU Session is established or released per S-NSSAI or per S-NSSAI and NSI ID. Current number of UEs registered in a NW slice NSACF NSACF reports the number of UE registered at the S-NSSAI. Current number of PDU Sessions established in a NW slice NSACF NSACF reports the number of PDU Sessions established at the S-NSSAI. Load of NFs associated to Network Slice instance NRF Resource utilization information of a Network Slice instance obtained from its constituent NF instances. NF instance load input data collection is described in clause 6.5, Table 6.5.2-1. NOTE 1: AMF reports the total number of registered UE in the AMF at each associated time stamp. NOTE 2: SMF reports multiple PDU Sessions when establishment or release happened at the same time, indicated by the time stamp. NOTE 3: Based on the internal logic, the NWDAF determines the source for the data collection. NWDAF collects input data on the number of UEs registered in a S-NSSAI or S-NSSAI and NSI ID combination using one of the following options: - Total number of UE registered to a S-NSSAI or to a S-NSSAI and NSI ID from each AMF(s) and/or from NSCAF serving the slice: - Namf_EventExposure_Subscribe (Target for Event Reporting = "any UE", Event ID = "Number of UEs served by the AMF and located in "Area of Interest"", Event Filter information = S-NSSAI(s) or one or more of the tuple (S-NSSAI, NSI ID), Event reporting mode = periodic along with periodicity) as defined in clause 5.2.2.3.1 of TS 23.502 [3]; or - Nnsacf_SliceEventExposure_Subscribe (EventID = "Number of UE registered", EventFilter = "S-NSSAI", Event reporting mode = periodic along with periodicity) as defined in clause 5.2.21.4.2 of TS 23.502 [3]. - Individual UE registration/deregistration to a S-NSSAI or to a S-NSSAI and NSI ID reported by AMF(s): - Namf_EventExposure_Subscribe (Target for Event Reporting = "any UE", Event ID = "UE moving in or out of a subscribed "Area of Interest", Event Filter information = S-NSSAI(s) or one or more of the tuples (S-NSSAI, NSI ID), Event reporting mode = reporting to a maximum number or a maximum duration) as defined in clause 5.2.2.3.1 of TS 23.502 [3]. NWDAF collects input data on the number of PDU Sessions established in a S-NSSAI using one of the following options: - Total number of PDU Sessions established in a S-NSSAI from each NSACF serving the slice: - Nnsacf_SliceEventExposure_Subscribe (EventID = "Number of PDU sessions established", EventFilter = "S-NSSAI(s)", Event reporting mode = periodic along with periodicity) as defined in clause 5.2.21.4.2 of TS 23.502 [3]. - Individual PDU Session Established or PDU Session Released in a S-NSSAI from SMF: - Nsmf_EventExposure_Subscribe (Target for Event Reporting = "any UE", Event ID = "PDU Session Establishment and/or PDU Session Release", Event Filter information = S-NSSAI(s), Event reporting mode = reporting to a maximum number or a maximum duration) as defined in clause 5.2.8.3.1 of TS 23.502 [3].
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.3.3 Void	6.3.3A Output analytics The NWDAF services as defined in the clause 7.2 and 7.3 are used to expose the following analytics: - Network Slice instance load statistics information as defined in Table 6.3.3A-1. - Network Slice load statistics information as defined in Table 6.3.3A-2. - Network Slice instance load predictions information as defined in Table 6.3.3A-3. - Network Slice load predictions information as defined in Table 6.3.3A-4. Table 6.3.3A-1: Network Slice instance load statistics Information Description S-NSSAI Identification of the Network Slice. Network Slice instances (1..max) List of Network Slice instance(s) within the S-NSSAI. > NSI ID Identification of the Network Slice instance. > Number of UE Registrations (NOTE 1) Number of UE registrations of the Network Slice instance (average, variance). > Number of PDU Sessions establishment (NOTE 1) Number of PDU Session establishments of the Network Slice instance (average, variance). > Resource usage (NOTE 1) The usage of assigned virtual resources currently in use for the NF instances (mean usage of virtual CPU, memory, disk) as defined in clause 5.7 of TS 28.552 [8], belonging to a particular Network Slice instance. > Resource usage threshold crossings (NOTE 1) Number of times resource usage threshold is met or exceeded or crossed on the Network Slice instance and the time when it happened. It is present if threshold is provided by the consumer as Analytics Filter. > Resource usage threshold crossings time period (1..max) (NOTE 1, NOTE 2) Resource usage threshold crossing vector including time elapsed between times each threshold is met or exceeded or crossed on the Network Slice instance if a threshold value is provided by the consumer as Analytics Filter. > Load Level (NOTE 1) The load level of the Network Slice Instance indicated by the S-NSSAI and the associated NSI ID (if applicable) in the Analytics Filter, it is present if Load Level Threshold is not provided by the consumer as Analytics Filter. > Crossed Load Level Threshold (NOTE 1) An indication on whether the Load Level Threshold is met or exceeded by the statistics value of the Load Level. It is present if the Load Level Threshold is provided by the consumer as Analytics Filter. Spatial validity Area (i.e. list of TAIs) where the network slice instance load statistics applies. Spatial validity may be a subset of the requested Area of Interest provided by the consumer. NOTE 1: Analytics subset that can be used in "list of analytics subsets that are requested". NOTE 2: The time period is a time interval specified by a start time and an end time timestamps within the Analytics target period. Table 6.3.3A-2: Network Slice load statistics Information Description S-NSSAI Identification of the Network Slice. > Number of UE Registrations (NOTE 1) Number of UE registrations at the Network Slice (average, variance). > Number of PDU sessions establishments (NOTE 1) Number of PDU Session establishments at the Network Slice (average, variance). > Load Level (NOTE 1) The load level of the Network Slice Instance indicated by the S-NSSAI and the associated NSI ID (if applicable) in the Analytics Filter, it is present if Load Level Threshold is not provided by the consumer as Analytics Filter. > Crossed Load Level Threshold (NOTE 1) An indication on whether the Load Level Threshold is met or exceeded by the statistics value of the Load Level. It is present if the Load Level Threshold is provided by the consumer as Analytics Filter. Spatial validity Area (i.e. list of TAIs) where the network slice load statistics applies. Spatial validity may be a subset of the requested Area of Interest provided by the consumer. NOTE 1: Analytics subset that can be used in "list of analytics subsets that are requested". Table 6.3.3A-3: Network Slice instance load predictions Information Description S-NSSAI Identification of the Network Slice. Network Slice instances (1..max) List of Network Slice instance(s) within the S-NSSAI. > NSI ID Identification of the Network Slice instance. > Number of UE Registrations (NOTE 1) Number of predicted UE registrations at the Network Slice instance (average, variance). > Number of PDU Sessions establishment (NOTE 1) Number of predicted PDU Session establishments of the Network Slice instance (average, variance). > Resource usage (NOTE 1) The predicted usage of assigned virtual resources for the NF instances (mean usage of virtual CPU, memory, disk) as defined in clause 5.7 of TS 28.552 [8], belonging to a particular Network Slice instance. > Resource usage threshold crossings (NOTE 1) Number of predicted times resource usage threshold is met or exceeded or crossed at the Network Slice instance and the time when it happened. It is present if a threshold value is provided by the consumer as Analytics Filter. > Resource usage threshold crossings time period (1..max) (NOTE 1, NOTE 2) Predicted Resource usage threshold vector including predicted time elapsed between times each threshold is met or exceeded or crossed on the Network Slice instance, it is present if a threshold value is provided by the consumer as Analytics Filter. > Load Level (NOTE 1) The load level of the Network Slice Instance indicated by the S-NSSAI and the associated NSI ID (if applicable) in the Analytics Filter, if Load Level Threshold is not provided by the consumer as Analytics Filter. > Crossed Load Level Threshold (NOTE 1) An indication on whether the Load Level Threshold is met or exceeded by the predicted value of the Load Level. It is present if the Load Level Threshold is provided by the consumer as Analytics Filter. Spatial validity Area (i.e. list of TAIs) where the network slice instance load prediction applies. Spatial validity may be a subset of the requested Area of Interest provided by the consumer. > Confidence Confidence of this prediction. NOTE 1: Analytics subset that can be used in "list of analytics subsets that are requested". NOTE 2: The time period is a time interval specified by a start time and an end time timestamps within the Analytics target period. Table 6.3.3A-4: Network Slice load predictions Information Description S-NSSAI Identification of the Network Slice. > Number of UE Registrations (NOTE 1) Predicted Number of UE registrations at the Network Slice (average, variance). > Number of PDU sessions establishments (NOTE 1) Predicted Number of PDU Session establishments at the Network Slice (average, variance). > Load Level (NOTE 1) The load level of the Network Slice Instance indicated by the S-NSSAI and the associated NSI ID (if applicable) in the Analytics Filter, if Load Level Threshold is not provided by the consumer as Analytics Filter. > Crossed Load Level Threshold (NOTE 1) An indication of whether the Load Level Threshold is met or exceeded by the predicted value of the Load Level. It is present if the Load Level Threshold is provided by the consumer as Analytics Filter. Spatial validity Area (i.e. list of TAIs) where the network slice load prediction applies. Spatial validity may be a subset of the requested Area of Interest provided by the consumer. > Confidence Confidence of this prediction. NOTE 1: Analytics subset that can be used in "list of analytics subsets that are requested". NOTE: If no NSI ID is provided as Analytics Filter, slice load level related output analytics are provided according to Tables 6.3.3A-2 and 6.3.3A-4. Otherwise slice instance load level related output analytics are provided according to Tables 6.3.3A-1 and 6.3.3A-3. The predictions are provided with a Validity Period, as defined in clause 6.1.3. The Network Slice Load statistics and predictions may be exposed to the AF by NEF under the conditions defined in clause 5.9.2.3 of TS 33.501 [49] NEF security requirements.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.3.4 Procedures	Figure 6.3.4-1: Network Slice load analytics provided by NWDAF Figure 6.3.4-1 shows the procedure for NWDAF to derive slice load analytics. The steps are described as follows: 1. A consumer NF subscribes to/requests a NWDAF using Nnwdaf_AnalyticsSubscription_Subscribe or Nnwdaf_AnalyticsInfo_Request service operation (Analytics ID = Load level information and a set of Analytics Filters (e.g. S-NSSAI, NSI ID, Area of Interest)). 2. [OPTIONAL] If the NWDAF does not have already the slice information, it gains the slice information from OAM (as described in clause 6.2.3) and selects, based on discovery towards NRF, the AMF, SMF and NSSF instance(s) relevant to the Analytics Filters provided in the analytics subscription. 3. [OPTIONAL] If the NSI ID(s) are not provided in the analytics subscription by the consumer NF, the NWDAF invokes Nnssf_NSSelection_Get service operation from NSSF to obtain the NSI ID(s) corresponding to the S-NSSAI in the subscription. NOTE: Step 4a to step 7 are conditional depending on the NWDAF internal logic that determines the source(s) of data collection. 4a. [CONDITIONAL] The NWDAF may subscribe to input data in Table 6.3.2A-1 from the OAM according to the data collection principles from the OAM described in clause 6.2.3. 4b. [CONDITIONAL] The NWDAF may collect input data from the NRF (see clause 6.5) to derive slice instance resource usage statistics and predictions for a Network Slice instance. 5. [CONDITIONAL] The NWDAF may subscribe to the AMF(s) event exposure service to collect data on the number of UEs currently registered on certain Network Slice and, if available, its constituent Network Slice instance(s) as defined in clause 6.3.2A. If required, the NWDAF may also collect the corresponding UE IDs. 6. [CONDITIONAL] The NWDAF may subscribe to the SMF(s) event exposure service to collect data on the number of PDU sessions established and/or released at the SMF on currently registered on certain Network Slice as defined in clause 6.3.2A. NWDAF can then use such collected data to determine the number of PDU sessions established on i) a Network Slice; and ii) if available, on a Network Slice instance by leveraging the data collected in step 5. 7. [CONDITIONAL] The NWDAF may subscribe to one or multiple NSACFs to collect data on either the number of UE registered in a S-NSSAI or the number of PDU sessions established in a S-NSSAI as defined in clause 6.3.2A. When multiple NSACFs are selected by the NWDAF for the S-NSSAI, the NWDAF aggregates the reports from the NSACFs to derive the number of UEs registered in the S-NSSAI or the number of PDU sessions established in the S-NSSAI. 8. The NWDAF derives slice load analytics. 9. The NWDAF delivers analytics to the consumer NF by invoking Nnwdaf_AnalyticsSubscription_Notify or Nnwdaf_AnalyticsInfo_Request response service operations.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.4 Observed Service Experience related network data analytics
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.4.1 General	This clause specifies how NWDAF can provide Observed Service Experience (i.e. average of observed Service MoS and/or variance of observed Service MoS indicating service MOS distribution for services such as audio-visual streaming as well as services that are not audio-visual streaming such as V2X and Web Browsing services) analytics, in the form of statistics or predictions, to a service consumer. The Observed Service Experience analytics may provide one or more of the following outputs: - Service Experience for a Network Slice: Service Experience for a UE or group of UEs or any UE in a Network Slice; - Service Experience for an Application: Service Experience for a UE or group of UEs or any UE in an Application or a set of Applications; - Service Experience for an Edge Application over a UP path: Service experience for a UE or a group UEs or any UE in an Application or a set of Applications over a specific UP path (UPF, DNAI and EC server); - Service Experience for an Application over a RAT Type or Frequency or both: Service experience for a UE or group of UEs in an Application or a set of Applications over a RAT Type or over a Frequency or both as defined in Table 6.4.1-1. - Service Experience for an Application transferring data over a PDU Session: Service experience for a UE or group of UEs or any UE in an Application or a set of Applications transferring data over a PDU Session with PDU Session parameters i.e. S-NSSAI, DNN, PDU Session Type , SSC mode and optionally an Access Type or with combination of PDU Session parameters such as a list of the tuple (PDU Session Type, SSC mode) optionally per Access Type. Therefore, Observed Service experience may be provided as defined in clause 6.4.3. For example, individually per UE or group of UEs, or globally, averaged per Application or averaged across a set of Applications on a Network Slice. The service consumer may be an NF (e.g. PCF, NSSF, AMF, NEF), AF, or the OAM. The consumer of these analytics shall indicate in the request or subscription: - Analytics ID = "Service Experience"; - Target of Analytics Reporting as defined in clause 6.1.3; - Analytics Filter Information as defined in Table 6.4.1-1 and optionally a list of analytics subsets that are requested (see clause 6.4.3); - optionally, maximum number of objects and maximum number of SUPIs; - optionally, preferred level of accuracy of the analytics; - optionally, preferred level of accuracy per analytics subset; - optionally, preferred order of results for the list of Application Service Experiences and/or Slice instance service experiences: "ascending" or "descending"; - optionally, preferred granularity of location information: TA level or cell level or "longitude and latitude level"; - Analytics target period that indicates the time window for which the statistics or predictions are requested; - in a subscription, the Notification Correlation Id and the Notification Target Address; and - optionally, Reporting Thresholds, which apply only for subscriptions and indicate conditions on the Service Experience to be reached in order to be notified by the NWDAF (see Table 6.4.3-1 and Table 6.4.3-2). NOTE: Definition of "longitude and latitude level" is described in clause 6.1.3. Table 6.4.1-1: Analytics Filter Information related to the observed service experience Information Description Mandatory Application Network Slice Edge Applications over a UP path Application over RAT Type and frequency Application transfering data over a PDU Session Application ID (0...max) The identification of the application(s) for which the analytics information is subscribed or requested. Y N Y Y Y S-NSSAI (NOTE 3) When requesting Service Experience for a Network Slice: identifies the Network Slice for which analytics information is subscribed or requested. When requesting Service Experience for an Application: identifies the S-NSSAI used to access the application together with the DNN listed below. N Y N N C NSI ID(s) Identifies the Network Slice instance(s) for which analytics information is subscribed or requested. N N N N N Area of Interest (NOTE 6) Identifies the Area (i.e. set of TAIs), as defined in TS 23.501 [2] for which the analytics information is subscribed or requested. N N N N N DNN (NOTE 3) When requesting Service Experience for an Application, this is the DNN to access the application. N N N N C DNAI (NOTE 1) Identifier of a user plane access to one or more DN(s) where applications are deployed as defined in TS 23.501 [2]. N N Y N N RAT Type (NOTE 2) Identifies the RAT type. N N N Y N Frequency (NOTE 2) Identifies the Frequency value(s) (e.g. high, low). N N N Y N Application Server Addresses (NOTE 1) List of IP addresses/FQDNs of the Application Servers the Target of Analytics Reporting has a communication session for which Service Experience Analytic information is requested. N N Y N N UPF anchor ID (NOTE 1) (NOTE 4) Identifies the UPF where a UE has an associated PDU session N N N N N PDU Session type (NOTE 3) Identifies the type of the associated PDU Session N N N N C SSC Mode (NOTE 3) Identifies the SSC Mode selected for the associated PDU Session N N N N C Access Type (NOTE 3) Identifies the Access type of the associated PDU Session N N N N C Mapped NSSAI (NOTE 5) Identifies the mapped NSSAI in the HPLMN. May be used in VPLMN for analytics exposure in roaming case (see clause 6.1.5). N N Y N N PLMN ID Identifies the target PLMN (i.e. PLMN from which the analytics are requested, for analytics exposure in roaming case (see clause 6.1.5). N N Y N N NOTE 1: These parameters can be provided when a consumer requires analytics for an edge application over a UP path. NOTE 2: A service consumer can provide either a RAT Type or a Frequency or a specific combination of RAT Type and Frequency. A service consumer can also provide multiple instances of RAT Type or multiple instances of Frequency or multiple combinations of RAT type and Frequency. A service consumer can also provide "any" RAT type indication "any" Frequency value indication or "any" indication for all the RAT type and Frequency value the NWDAF has received for the application. NOTE 3: One or more of these parameters can be provided by the consumer when requesting analytics for an application running over a PDU Session(s). NOTE 4: UPF ID is only needed when the target of NWDAF analytics on Service Experience is a specific UPF. NOTE 5: The terms "HPLMN" and "VPLMN" here refer to a roaming case in which at least one UE served by the NWDAF analytics consumer is involved. NOTE 6: If the request is for fine granularity location information (i.e. with a finer granularity than cell), the AOI may be described as shown in clause 5.5 of TS 23.273 [39]. NOTE 1: A service consumer can use the Area of Interest in order to reduce the amount of signalling that the analytics subscription or request generates. The NWDAF shall notify the result of the analytics to the consumer as specified in clause 6.4.3. NWDAF collects the network data from AF (directly or via NEF) and from other 5GC NF(s) in order to calculate and provide statistics and predictions on the observed service experience to a consumer NF or to OAM. When the AF provides Service Experience Information and the Target of Analytics Reporting is one or more UE ID(s), the AF may also provide a Service Experience Contribution Weight with each UE's Service Experience value. The Service Experience Contribution Weight is determined by the AF and indicates the relative importance of each UE's Service Experience. The NWDAF may use the Service Experience Contribution Weight(s) to calculate and provide statistics, confidence values and predictions on the observed service experience to a consumer NF or to OAM. NOTE 2: The relative importance of that is conveyed in the Service Experience Contribution Weights is used to indicate the relative importance of the Service Experience value (i.e. MOS). For example, it might be that one the Service Experience of one UE is not very important because the UE is an infrequent user of the service, or the UE does not use all features that are associated with the service. Whereas another UE might be considered important because the UE is a frequent user of the service or a user who uses many features that are associated with the service. Based on the Analytics Filter information in Table 6.4.1-1 and the Target of Analytics Reporting provided by the service consumer in the analytics subscription or request, NWDAF determines whether service experience analytics should be delivered for: i) Application(s); ii) Network Slice; iii) both Application(s) and Network Slice; iv) Edge Applications over a UP path; v) Application(s) over RAT Type(s) and/or Frequency value(s); vi) Application(s) running over a PDU Session using the following PDU Session parameters or combination of them, i.e. S-NSSAI, DNN, PDU Session type, SSC Mode and optionally per Access Type. If NWDAF is unable to differentiate based on the analytics subscription or request, it provides service experience analytics for both Application(s) and Network Slice. If service experience for both Application(s) and Network Slice is desired but the Target of Analytics Reporting or Analytics Filter information values (e.g. Area of Interest) need to be different, separate subscriptions/requests may be provided by the service consumer.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.4.2 Input Data	The service data and performance data collected from the AF (including the service data collected from the UE through the AF), the network data from other 5GC NFs and the network data from OAM and MDAS/MDAF for observed service experience are defined in Table 6.4.2-1, 6.4.2-1a, Table 6.4.2-2, Table 6.4.2-3, Table 6.4.2-4 and Table 6.4.2-5 respectively. Table 6.4.2-1: Service Data from AF related to the observed service experience Information Source Description Application ID AF To identify the service and support analytics per type of service (the desired level of service) IP filter information AF Identify a service flow of the UE for the application Locations of Application AF/NEF Locations of application represented by a list of DNAIs. The NEF may map the AF-Service-Identifier information to a list of DNAI when the DNAI(s) being used by the application are statically defined. Service Experience AF Refers to the QoE per service flow as established in the SLA and during on boarding. It can be either e.g. MOS or video MOS as specified in ITU-T P.1203.3 [11] or a customized MOS for any kind of service including those not related to video or voice. UE ID AF The list of UE IDs that are associated with the Service Experience value(s). When the AF is untrusted, GPSIs will be provided. When the AF is trusted SUPIs will be provided. Service Experience Contribution Weight AF The list of Service Experience Contribution Weights that are associated with each of the provided UE IDs. QoE metrics UE (via AF) QoE metrics observed at the UE(s). QoE metrics and measurement as described in TS 26.114 [27], TS 26.247 [28], TS 26.118 [29], TS 26.346 [30], TS 26.512 [31] or ASP specific QoE metrics in TS 26.512 [31], as agreed in the SLA with the MNO, may be used. Timestamp AF A time stamp associated to the Service Experience provided by the AF, mandatory if the Service Experience is provided by the ASP. Application Server Instance AF The IP address or FQDN of the Application Server that the UE had a communication session when the measurement was made. NWDAF subscribes to the service data from AF in the Table 6.4.2-1 either directly for trusted AFs by invoking Naf_EventExposure_Subscribe service (Event ID = Service Experience information, Event Filter information = Area of Interest, Application ID) as defined in TS 23.502 [3], or indirectly for untrusted AFs via NEF by invoking Nnef_EventExposure_Subscribe service (Event ID = Service Experience information, Event Filter information = Area of Interest, Application ID) where NEF translates the Area of Interest into geographic zone identifier(s). For the information whose source is UE (via AF), the AF collects data from the UE as defined in clause 6.2.8. NOTE: When the Service Experience is expressed as a customized MOS, the customized MOS might be defined by the content provider or by the MNO and might be based on the nature of the targeted service type (e.g. web browsing, gaming, augmented reality, V2X, SMS). Table 6.4.2-1a: Performance Data from AF Information Source Description UE identifier AF IP address of the UE at the time the measurements was made. UE location AF The location of the UE when the performance measurement was made. Application ID AF To identify the service and support analytics per type of service (the desired level of service). IP filter information AF Identify a service flow of the UE for the application. Locations of Application AF/NEF Locations of application represented by a list of DNAIs. The NEF may map the AF-Service-Identifier information to a list of DNAIs when the DNAIs being used by the application are statically defined. Application Server Instance address AF/NEF The IP address/FQDN of the Application Server that the UE had a communication session when the measurement was made. Performance Data AF The performance associated with the communication session of the UE with an Application Server that includes: Average Packet Delay, Average Loss Rate and Throughput. Timestamp AF A time stamp associated to the Performance Data provided by the AF. NWDAF subscribes to the performance data from AF in the Table 6.4.2-1a either directly for trusted AFs by invoking Naf_EventExposure_Subscribe service (Event ID = Performance Data, Event Filter information = Area of Interest, Application ID) as defined in TS 23.502 [3], or indirectly for untrusted AFs via NEF by invoking Nnef_EventExposure_Subscribe service (Event ID = Performance Data, Event Filter information = Area of Interest, Application ID) where NEF translates the Area of Interest into geographic zone identifier(s). Table 6.4.2-1b: QoE measurements from OAM related to specific service type Information Source Description UE ID/Area scope OAM The UE Identity (i.e., qoETarget ("IMSI" or "SUPI") or areaScope as defined in TS 28.405 [55]) that are associated with the QoE measurements. Service Type OAM Identify the service type for the QoE measurements, i.e., Dynamic Adaptive Streaming over HTTP (DASH), Multimedia Telephony Service for IMS (MTSI) or Virtual Reality (VR). (NOTE) Area scope OAM Areas of the UEs with the QoE measurements. QoE metrics OAM QoE metrics observed at the UE(s). QoE metrics and measurement as described in TS 28.406 [56]. Timestamp OAM A time stamp associated to the QoE measurements data collection provided by the UE. NOTE: The NWDAF maps the Application ID into a Service Type that is understood by OAM. NWDAF subscribes the network data from OAM in the Table 6.4.2-1b by using the services provided by OAM as described in clause 6.2.3. Table 6.4.2-2: QoS flow level Network Data from 5GC NF related to the QoS profile assigned for a particular service (identified by an Application Id or IP filter information) Information Source Description Timestamp 5GC NF A time stamp associated with the collected information. Location AMF The UE location information, e.g. cell ID or TAI. Finer granularity location (1...max) GMLC UE positions. ....>UE location GAD shape or location coordinates (see TS 23.032 [34]). ....>Timestamp A time stamp when the location was measured. ....>LCS QoS The accuracy of the measurement. UE ID AMF List of SUPIs. If UE IDs are not provided as Target of Analytics Reporting for slice service experience, AMF returns the UE IDs matching the AMF event filters. DNN SMF DNN for the PDU Session which contains the QoS flow. S-NSSAI SMF S-NSSAI for the PDU Session which contains the QoS flow. Application ID SMF Used by NWDAF to identify the application service provider and application for the QoS flow. DNAI SMF Identifies the access to DN to which the PDN session connects. PDU Session type SMF Type of the PDU Session. SSC Mode SMF SSC Mode selected for the PDU Session. Access Type SMF List of Access Types used for the PDU Session. IP filter information SMF Provided by the SMF, which is used by NWDAF to identify the service data flow for policy control and/or differentiated charging for the QoS flow. QFI SMF QoS Flow Identifier. QoS flow Bit Rate UPF The observed bit rate for UL direction; and The observed bit rate for DL direction. QoS flow Packet Delay UPF The observed Packet delay for UL direction; and The observed Packet delay for the DL direction. Packet transmission UPF The observed number of packet transmission. Packet retransmission UPF or AF The observed number of packet retransmission. NOTE: Care needs to be taken with regards to load and major signalling caused when requesting Any UE. This could be achieved via utilization of some event filters (e.g. Area of Interest for AMF), Analytics Reporting Information (e.g. SUPImax), or sampling ratio as part of Event Reporting Information. NWDAF subscribes to the network data from 5GC NF(s) in the Table 6.4.2-2 by invoking Nnf_EventExposure_Subscribe service operation with the following Event IDs as input parameters: - AMF Source: Namf_EventExposure_Subscribe (Event IDs = Location Changes, Area of Interest). - SMF Source: Nsmf_EventExposure_Subscribe (Event ID = QFI allocation). Table 6.4.2-3: UE level Network Data from OAM related to the QoS profile Information Source Description Timestamp OAM A time stamp associated with the collected information. Reference Signal Received Power OAM (see NOTE 1) The per UE measurement of the received power level in a network cell, including SS-RSRP, CSI-RSRP as specified in clause 5.5 of TS 38.331 [14] and E-UTRA RSRP as specified in clause 5.5.5 of TS 36.331 [15]. Reference Signal Received Quality OAM (see NOTE 1) The per UE measurement of the received quality in a network cell, including SS-RSRQ, CSI-RSRQ as specified in clause 5.5 of TS 38.331 [14] and E-UTRA RSRQ as specified in clause 5.5.5 of TS 36.331 [15]. Signal-to-noise and interference ratio OAM (see NOTE 1) The per UE measurement of the received signal to noise and interference ratio in a network cell, including SS-SINR, CSI-SINR, E-UTRA RS-SINR, as specified in clause 5.1 of TS 38.215 [12]. RAN Throughput for DL and UL OAM (see NOTE 1) The per UE measurement of the throughput for DL and UL as specified in clauses 5.2.1.1 and 5.4.1.1 of TS 37.320 [20]. RAN Packet delay for DL and UL OAM (see NOTE 1) The per UE measurement of the packet delay for DL and UL, including per QCI per UE packet delay as specified in clause 5.2.1.1 of TS 37.320 [20] and per DRB per UE packet delay as specified in clause 5.4.1.1 of TS 37.320 [20]. RAN Packet loss rate for DL and UL OAM (see NOTE 1) The per UE measurement of the packet loss rate for DL and UL, including the per QCI per UE packet loss rate as specified in clause 5.2.1.1 of TS 37.320 [20] and the per DRB per UE packet loss rate as specified in clause 5.4.1.1 of TS 37.320 [20]. The mapping information between cell ID and frequency OAM The mapping information between cell ID and frequency (NOTE 2). Cell Energy Saving State OAM List of the cells which are within the area of interest and are in energy saving state, as specified in clauses 3.1 and 6.2 of TS 28.310 [24]. NOTE 1: Per UE measurement for a specific UE from OAM (via MDT), is as captured in clause 6.2.3.1. NOTE 2: The MDT measurement report provides the cell identity and carrier frequency information for UE's serving cell and neighbour cell(s). The NWDAF can get the mapping information between cell ID and frequency using OAM service as described in clause 6.2.3. NWDAF subscribes the network data from OAM in the Table 6.4.2-3 by using the services provided by OAM as described in clause 6.2.3. Table 6.4.2-4: UE level Network Data from 5G NF related to the Service Experience Information Source Description Timestamp 5GC NF A time stamp associated with the collected information. Location AMF The UE location information, e.g. cell ID or TAI. Finer granularity location GMLC UE positions. UE ID AMF List of SUPIs. RAT Type SMF The RAT type the UE camps on. The Event Filters for the service data collection from SMF, AMF and AF are defined in TS 23.502 [3]. The timestamps are provided by each NF to allow correlation of QoS and traffic KPIs. The clock reference is able to know the accuracy of the time and correlate the time series of the data retrieved from each NF. The NWDAF collects the following MDAF analysis result listed in Table 6.4.2-5, as defined in clauses 8.4.2.1.3 and 8.4.4.1.3 of TS 28.104 [45]. Table 6.4.2-5: Data collection from MDAS/MDAF of service experience and energy saving state analysis Information Source Description ServiceExperienceIssueType MDAF Indication of the service experience issue type. The allowed value is one of the enumerated values: RAN issue, CN issue, both. AffectedObjects MDAF The managed object instances where the service experience is applicable, e.g. SubNetwork Instance, NetworkSlice Instance, S-NSSAI. ServiceExperienceStatistics MDAF The statistics of the level of service experience for a service in a certain time period, e.g. there are five levels which are represented by 1, 2, 3, 4, 5 where level 1 represents the users are enduring bad experience while level 5 represents the users' requirements are perfectly satisfied. ServiceExperiencePredictions MDAF The predictions of the level of service experience for a service in a certain time period. StatisticsOfCellsEsState MDAF The statistic result of current energy saving state of the cells at a certain time.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.4.3 Output Analytics	The NWDAF services as defined in the clause 7.2 and 7.3 are used to expose the analytics. - Service Experience statistics information is defined in Table 6.4.3-1. - Service Experience predictions information is defined in Table 6.4.3-2. Table 6.4.3-1: Service Experience statistics Information Description Slice instance service experiences (0..max) List of observed service experience information for each Network Slice instance. > S-NSSAI Identifies the Network Slice > NSI ID (NOTE 2) Identifies the Network Slice instance within the Network Slice. > Network Slice instance service experience Service experience across Applications on a Network Slice instance over the Analytics target period (average, variance). > SUPI list (0..SUPImax) (NOTE 3) List of SUPI(s) for which the slice instance service experience applies. > Ratio (NOTE 3) Estimated percentage of UEs with similar service experience (in the group, or among all UEs). > Spatial validity (NOTE 6) Area where the Network Slice service experience analytics applies. > Validity period Validity period for the Network Slice service experience analytics as defined in clause 6.1.3. Application service experiences (0..max) List of observed service experience information for each Application. > S-NSSAI Identifies the Network Slice used to access the Application. > Application ID Identification of the Application. > Service Experience Type Type of Service Experience analytics, e.g. on voice, video, other. > UE location (NOTE 1, NOTE 5) Indicating the UE location information (e.g. TAI list, gNB ID, or location coordinates, etc) when the UE service is delivered. > UPF Info (NOTE 4) Indicating UPF serving the UE. > DNAI Indicating which DNAI the UE service uses/camps on. > DNN DNN for the PDU Session which contains the QoS flow. > Application Server Instance Address Identifies the Application Server Instance (IP address of the Application Server) or FQDN of Application Server. > Service Experience Service Experience over the Analytics target period (average, variance). > SUPI list (0..SUPImax) (NOTE 3) List of SUPI(s) with the same application service experience. > Ratio (NOTE 3) Estimated percentage of UEs with similar service experience (in the group, or among all UEs). > Spatial validity (NOTE 6) Area where the Application service experience analytics applies. > Validity period Validity period for the Application service experience analytics as defined in clause 6.1.3. > RAT Type (NOTE 7) Indicating the list of RAT type(s) for which the application service experience analytics applies. > Frequency (NOTE 7) Indicating the list of carrier frequency value(s) of UE's serving cell(s) where the application service experience analytics applies. > SSC Mode SSC Mode selected for the PDU Session used to associate with the application. > PDU Session Type Type of PDU Session used to associate with the application. > Access Type List of Access Type(s) used for the PDU Session for the application. NOTE 1: This information element is an Analytics subset that can be used in "list of analytics subsets that are requested" and "Preferred level of accuracy per analytics subset". NOTE 2: The NSI ID is an optional parameter. If not provided the Slice instance service experience indicates the service experience for the S-NSSAI. NOTE 3: The SUPI list and Ratio in the service experience information for an application can be omitted, if the corresponding parameter(s) is/are provided and are assigned with the same value(s) in the service experience information for the slice instance which the application belongs to. Otherwise, the SUPI list and Ratio are mandatory to be provided for an application service experience. NOTE 4: If the consumer NF is an AF, the "UPF info" shall not be included. NOTE 5: When possible and applicable to the access type, UE location is provided according to the preferred granularity of location information. UE location shall only be included if the Consumer analytics request is for single UE or a list of UEs. Inclusion of UE location requires user consent. NOTE 6: The Spatial validity is present in the output parameters if the consumer provided the Area of Interest as defined in Table 6.4.1-1. NOTE 7: When "any" value has been provided in the request (e.g. "any" RAT type, "any" frequency, or "any" for all the RAT type and frequency indication), the NWDAF provides an instance of the Application service experience per combination of RAT Type(s) and/or Frequency value(s) having the same Service Experience. Table 6.4.3-2: Service Experience predictions Information Description Slice instance service experiences (0..max) List of observed service experience information for each Network Slice instance. > S-NSSAI Identifies the Network Slice > NSI ID (NOTE 2) Identifies the Network Slice instance within the Network Slice. > Network Slice instance service experience Service experience across Applications on a Network Slice instance over the Analytics target period (average, variance). > SUPI list (0..SUPImax) (NOTE 3) List of SUPI(s) for which the slice instance service experience applies. > Ratio (NOTE 3) Estimated percentage of UEs with similar service experience (in the group, or among all UEs). > Spatial validity (NOTE 6) Area where the Network Slice service experience analytics applies. > Validity period Validity period for the Network Slice service experience analytics as defined in clause 6.1.3. > Confidence Confidence of this prediction. Application service experiences (0..max) List of predicted service experience information for each Application. > S-NSSAI Identifies the Network Slice used to access the Application. > Application ID Identification of the Application. > Service Experience Type Type of Service Experience analytics, e.g. on voice, video, other. > UE location (NOTE 1, NOTE 5) Indicating the UE location information (e.g. TAI list, gNB ID, or location coordinates, etc.) when the UE service is delivered. > UPF Info (NOTE 4) Indicating UPF serving the UE. > DNAI Indicating which DNAI the UE service uses/camps on. > DNN DNN for the PDU Session which contains the QoS flow. > Application Server Instance Address Identifies the Application Server Instance (IP address of the Application Server) or FQDN of Application Server. > Service Experience Service Experience over the Analytics target period (average, variance). > SUPI list (0..SUPImax) (NOTE 3) List of SUPI(s) with the same application service experience. > Ratio (NOTE 3) Estimated percentage of UEs with similar service experience (in the group, or among all UEs). > Spatial validity (NOTE 6) Area where the Application service experience analytics applies. > Validity period Validity period for the Application service experience analytics as defined in clause 6.1.3. > Confidence Confidence of this prediction. > RAT Type (NOTE 7) Indicating the list of RAT type(s) for which the application service experience analytics applies. > Frequency (NOTE 7) Indicating the list of carrier frequency value(s) of UE's serving cell(s) where the application service experience analytics applies. > SSC Mode SSC Mode selected for the PDU Session used to associate with the application. > PDU Session Type Type of PDU Session used to associate with the application. > Access Type List of Access Type(s) used for the PDU Session for the application. NOTE 1: This information element is an Analytics subset that can be used in "list of analytics subsets that are requested" and "Preferred level of accuracy per analytics subset". NOTE 2: The NSI ID is an optional parameter. If not provided the Slice instance service experience indicates the service experience for the S-NSSAI. NOTE 3: The SUPI list and Ratio in the service experience information for an application can be omitted, if the corresponding parameter(s) is/are provided and are assigned with the same value(s) in the service experience information for the slice instance which the application belongs to. Otherwise, the SUPI list and Ratio are mandatory to be provided for an application service experience. NOTE 4: If the consumer NF is an AF, the "UPF info" shall not be included. NOTE 5: When possible and applicable to the access type, UE location is provided according to the preferred granularity of location information. UE location shall only be included if the Consumer analytics request is for single UE or a list of UEs. Inclusion of UE location requires user consent. NOTE 6: The Spatial validity is present in the output parameters if the consumer provided the Area of Interest as defined in Table 6.4.1-1. NOTE 7: When "any" value has been provided in the request (e.g. "any" RAT type, "any" frequency, or "any" for all the RAT type and frequency indication), the NWDAF provides an instance of the Application service experience per combination of RAT Type(s) and/or Frequency value(s) having the same Service Experience. The number of Service Experiences and SUPIs are limited respectively by the maximum number of objects and the Maximum number of SUPIs provided as part of Analytics Reporting Information by the NWDAF Service Consumer.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.4.4 Procedures to request Service Experience for an Application	Figure 6.4.4-1: Procedure for NWDAF providing Service Experience for an Application This procedure allows the consumer to request Analytics ID "Service Experience" for a particular Application. The consumer includes both the Application ID for which the Service Experience is requested and indicates that the Target of Analytics Reporting is "any UE". If the Target for Analytics Reporting is either a SUPI or an Internal-Group-Id the procedure in clause 6.4.6 applies. At the same time, for an Application ID, a set of initial QoS parameter combinations per service experience window (e.g. one is for 3<Service MOS<4 and another is for 4<Service MOS<5) is defined in PCF (e.g. by configuration of operator policies) that may be updated based on the Service Experience reported by NWDAF. 1. Consumer NF sends an Analytics request/subscribe (Analytics ID = Service Experience, Target of Analytics Reporting = any UE, Analytics Filter Information that may include one or more of the following as defined in Table 6.4.1-1 (Application ID, S-NSSAI, DNN, Application Server Address(es), Area of Interest, RAT type(s), Frequency value(s)), Analytics Reporting Information=Analytics target period) to NWDAF by invoking a Nnwdaf_AnalyticsInfo_Request or a Nnwdaf_AnalyticsSubscription_Subscribe. 2a. The NWDAF may subscribe to the service data from AF in the Table 6.4.2-1 by invoking Nnef_EventExposure_Subscribe or Naf_EventExposure_Subscribe service (Event ID = Service Experience information, Application ID, Event Filter information, Target of Event Reporting = Any UE) as defined in TS 23.502 [3]. NOTE 1: In the case of trusted AF, NWDAF provides the Area of Interest as a list of TAIs to AF. In the case of untrusted AF, NEF translates the requested Area of Interest provided as event filter by NWDAF into geographic zone identifier(s) that act as event filter for AF. 2b. If the NWDAF maps the provided Application ID is DASH, MTSI or VR in step 1, the NWDAF may act as an MnS Consumer to collect the QoE measurements data for all the UEs in the interested area from OAM as defined in TS 28.404 [54], TS 28.405 [55] and TS 28.406 [56]. 2c. The NWDAF subscribes the network data from 5GC NF(s) in the Table 6.4.2-2 by invoking Nnf_EventExposure_Subscribe service operation. 2d. With these data, the NWDAF estimates the Service experience for the application. NOTE 2: Care needs to be taken to avoid end to end excessive signalling for collecting the QoE measurements from the applications in the UE. 3. The NWDAF provides the data analytics, i.e. the observed Service Experience (which can be a range of values) to the consumer NF by means of either Nnwdaf_AnalyticsInfo_Request response or Nnwdaf_AnalyticsSubscription_Notify, depending on the service used in step 1, indicating how well the used QoS Parameters satisfy the Service MoS agreed between the MNO and the end user or between the MNO and the external ASP. NOTE 3: The call flow only shows a request-response model for the interaction of NWDAF and consumer NF for simplicity instead of both request-response model and subscription-notification model. NOTE 4: The non-real time data information from AF includes the service experience data (see Table 6.4.2-1), which indicates the service quality during the service lifetime. If the consumer NF is a PCF and it determines that the application SLA is not satisfied, it may take into account the Observed Service Experience and the operator policies including SLA and required Service Experience (which can be a range of values) to determine new QoS parameters to be applied for the service, as defined in clause 6.1.1.3 and clause 6.2.1.2 of TS 23.503 [4]. If the consumer NF is an AF (e.g. MEC or other Application Server), it may use the Observed Service Experience related network data analytics to determine whether the user experience can be satisfied. If not, the AF may determine to adjust service parameters, e.g. for a video service this may be bit rate, frame rate, codec format, compression parameter, screen size, etc. to better match the network conditions and achieve better user experience. If the consumer NF is SMF, PCF or AF/Application Server, it may take into account the Observed Service Experience analytics per UP path (i.e. UPF and/or DNAI and/or AS instance address as defined in Table 6.4.3-1) to perform the following procedures: - The consumer SMF determines to (re)selects UP paths, including UPF and DNAI, as described in clause 4.3.5 of TS 23.502 [3]. In addition, the SMF may (re)configure traffic steering, updating the UPF regarding the target DNAI with new traffic steering rules. - The consumer AF/Application Server determines to adjust service parameters, e.g. service parameters of video for adjustment may be bit rate, frame rate, codec format, compression parameter, screen size, etc. or service parameters for the AI/ML operations described in clause 6.40 of TS 22.261 [33]. In addition, the AF/ Application Server may provide an updated list of DNAI(s) for SMF to perform relocation when appropriate. - The consumer PCF may provide an updated list of DNAI(s) for SMF to perform relocation upon AF request. If the consumer NF is a NEF, it may take into account the Observed Service Experience analytics to support Member UE selection as detailed in clause 4.15.13 of TS 23.502 [3].
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.4.5 Procedures to request Service Experience for a Network Slice	Figure 6.4.5-1: Procedure for NWDAF providing Service Experience for a UE or a group of UEs in a Network Slice This procedure is similar to the procedure in clause 6.4.4, with the following differences. The consumer needs to request the Analytics ID "Service Experience" for a Target for Analytics Reporting as defined in clause 6.1.3 on a Network Slice, identified by an S-NSSAI. If multiple Network Slice instances of the same Network Slice are deployed, associated NSI ID(s) may be used in addition to S-NSSAI. If 'any UE' is the Target of Analytics Reporting, NWDAF may subscribe to UE mobility event notifications of AMF as described in clause 5.3.4.4 of TS 23.501 [2] using event ID "UE moving in or out of Area of Interest" and Event Filters as described in Table 5.2.2.3.1-1 of TS 23.502 [3] if it is needed to retrieve the list of SUPIs (and GPSIs if available) in the area of interest. The event exposure service request may also include the immediate reporting flag as Event Reporting Information as described in Table 4.15.1-1 of TS 23.502 [3]. In addition, service experience data may need to be collected from multiple Applications. If each Application is hosted in different AFs, NWDAF subscribes the service data in the Table 6.4.2-1 from the different AFs by invoking Nnef_EventExposure_Subscribe or Naf_EventExposure_Subscribe services for each Application (Event ID = Service Experience information, Event Filter information, Application ID) as defined in TS 23.502 [3]. Figure 6.4.5-1 shows an example procedure with two AFs. If one AF provides the service experience data of multiple Applications, the set of Application IDs is provided by NWDAF to the AF with the Naf_EventExposure_Subscribe service operation, as defined TS 23.502 [3]. The Observed Service Experience for a Network Slice when consumed by OAM could be used as described in Annex H of TS 28.550 [7].
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.4.6 Procedures to request Service Experience for a UE	Figure 6.4.6-1 depicts procedure for NWDAF providing Service Experience for an application for a UE or a group of UEs. Figure 6.4.6-1: Procedure for NWDAF providing Service Experience for an application for a UE or a group of UEs The procedure in clause 6.4.4 applies with the following additions. The consumer needs to request the Analytics ID "Service Experience" for a UE identified by a SUPI or a group of UEs identified by a list of Internal Group-Ids. The consumer includes both the Application ID for which their Service Experience is requested and the Target of Analytics Reporting. Analytic Filter Information can be set according to clause 6.4.1. The NWDAF may collect UE location information from the GLMC if the consumer requested fine granularity location information according to clause 6.4.2.1. When NEF is the NF service consumer, the NEF translates a GPSI into a SUPI or an External-Group-Id into an Internal-Group-Id then includes it in the Target of Analytics Reporting.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.5 NF load analytics
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.5.1 General	The clause 6.5 describes how NWDAF can provide NF load analytics, in the form of statistics or predictions or both, to another NF. The service consumer may be an NF, or the OAM. The consumer of these analytics shall indicate in the request: - Analytics ID = "NF load information"; - Target of Analytics Reporting: an optional SUPI or any UE; - Analytics Filter Information: - optional S-NSSAI; - an optional list of NF Instance IDs, NF Set IDs, or NF types; - optional area of interest; - an optional list of analytics subsets that are requested (see clause 6.5.3); - Optional preferred level of accuracy of the analytics; - Optional preferred level of accuracy per analytics subset (see clause 6.5.3); - Optional preferred order of results for the list of resource status: ascending or descending NF load; - Optional Reporting Threshold; the Reporting Threshold is unique for all NFs matching the above Analytics Filter and the reporting applies when the conditions are met for at least one of these NFs; - An Analytics target period indicates the time period over which the statistics or predictions are requested; - In a subscription, the Notification Correlation Id and the Notification Target Address are included. The NWDAF shall notify the result of the analytics to the consumer as indicated in clause 6.5.3. If a list of the NF Instance IDs (or respectively of NF Set IDs) is provided, the NWDAF shall provide the analytics for each designated NF instance (or respectively for each NF instance belonging to each designated NF Set). In such case the Target of Analytics Reporting should be ignored. Otherwise, if a SUPI is provided, the NWDAF shall use the SUPI to determine which NF instances (AMF and SMF) are serving this specific UE, filter them according to the provided S-NSSAI and NF types using data collected from NRF or OAM and provide analytics for these NF instances. NOTE: Only NF instances of type AMF and SMF can be determined using a SUPI.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.5.2 Input data	For the purpose of NF load analytics, the NWDAF may collect the information as listed in Table 6.5.2-1 for the relevant NF instance(s). Table 6.5.2-1: Data collected by NWDAF for NF load analytics Information Source Description NF load NRF The load of specific NF instance(s) in their NF profile as defined per TS 29.510 [18]. NF status NRF The status of a specific NF instance(s) (registered, suspended, undiscoverable) as defined per TS 29.510 [18]. NF resource usage OAM The usage of assigned virtual resources currently in use for specific NF instance(s) (mean usage of virtual CPU, memory, disk) as defined in clause 5.7 of TS 28.552 [8]. NF resource configuration OAM The life cycle changes of specific NF resources (e.g. NF operational or interrupted during virtual/physical resources reconfiguration) as defined in clause 5.2 of TS 28.533 [19]. NOTE 1: The OAM information can be used as a complement to NRF information for some or all of the following aspects: resources utilization, NRF information correlation and alternative source of information if NRF information on load is not available. NOTE 2: NWDAF can request NRF for data related to NF instances, as described in TS 29.510 [18]. NOTE 3: NWDAF can correlate the NF resources configuration with NF resource usage for generating the analytics output. If target NF type is UPF, the NWDAF may collect the information as listed in Table 6.5.2-2, in addition to information listed in Table 6.5.2-1. Table 6.5.2-2: Data collected by NWDAF for UPF load analytics Information Source Description Traffic usage report UPF Report of user plane traffic in the UPF for the accumulated usage of network resources (see clause 5.2.1.3.3 of TS 29.564 [51]). For the purpose of NF load analytics, the NWDAF may collect the information as listed in Table 6.5.2-3 (from OAM via MDT) and Table 6.5.2-5 via the AF (for trusted AF) or NEF (for untrusted AF) in addition to other information described above. Table 6.5.2-3: MDT input data for UE Information Source Description UE Speed OAM (see NOTE 1) UE Speed (see TS 37.320 [20]). UE Orientation OAM (see NOTE 1) UE Orientation (see TS 37.320 [20]). NOTE 1: UE input data collection for a specific UE from OAM (via MDT), is as captured in clause 6.2.3.1. Table 6.5.2-4: Per UE attribute to be collected and processed by the AF Information Source Description Per UE attribute UE Application (see NOTE 1) UE application data to be collected from UE. > Destination Expected final location of UE based on the route planned. > Route Planned path of movement by a UE application (e.g. a navigation app). The format is based on the SLA. > Average Speed Expected speed over the route planned by a UE application. > Time of arrival Expected Time of arrival to destination based on the route planned. NOTE 1: The procedure for data collection from UE Application is as covered in clause 6.2.8. Table 6.5.2-5: AF input data to the NWDAF for Collective Behaviour of UEs Information Source Description Collective Attribute AF / NEF (see NOTE 1, NOTE 2) Characterise collective attribute per set of UEs (see Table 6.5.2-4) within the area of interest. > Number of UEs Total number of UEs that fulfil a collective behaviour within the area of interest. > Timestamp A time stamp of time that the collective attribute derived. > Application ID(s) (see NOTE 3) Identifying the application providing this information > List of UE IDs (see NOTE 4) UE IDs that fulfil a collective behaviour within the area of interest. NOTE 1: For collective behaviour attribute, data processing procedure is as defined in clause 6.2.8. NOTE 2: Per collective attribute, the AF may provide several collective attribute sets, if several sets of UEs with similar behaviour are identified. A similar behaviour can be identified to specific ranges if the AF performs data processing (Data Anonymisation, Aggregation or Normalization) based on NWDAF request. UEs falling in the same range per UE attribute can form a collective attribute set. NOTE 3: The application ID(s) (either external or Internal) is optional. If the application ID(s) is not provided, the relevant application ID(s) can be identified by NWDAF based on the relevant event ID as registered in NRF as covered in clause 6.2.8.2.2. NOTE 4: List of UE IDs is optional and subject to support by the AF when processing the data based on NWDAF request. Based on network configuration, NWDAF may discover the AF from the NRF as defined in 6.2.8.2.2 (based on Collective Behaviour as Event ID or a corresponding Application ID). For AF in trusted domain, the NWDAF invokes step 3a in clause 6.2.8.2.3 by using Naf_EventExposure_Subscribe service (Event ID = Collective Behaviour, Event Filter information, Target of Event Reporting). The collective attribute (see Table 6.5.2-5) can be indicated as part of event filter information as defined in TS 23.502 [3]. Otherwise, the AF notifies for all collective attributes within the area of interest. For AF in untrusted domain, the NWDAF invokes step 3b in clause 6.2.8.2.3 by using Nnef_EventExposure_Subscribe (Event ID = Collective Behaviour, Event Filter information, Target of Event Reporting). The collective attribute (see Table 6.5.2-5) can be indicated as part of event filter information as defined in TS 23.502 [3]. Otherwise, the AF via NEF notifies for all collective attributes within the area of interest. For Collective Behaviour of multiple UEs, NWDAF based on the configuration by MNO may request certain type of data processing from the AF as part of event filter information (e.g. for anonymisation, normalisation, aggregation). The data processing requested by NWDAF is used to anonymise, normalise or aggregate the same UE attribute from multiple UEs at the AF before notifying to the NWDAF. For each UE attribute of a specific UE, whether and how AF is processing the data that is received from the UE depends on the SLA configured in AF (defined in clause 6.2.8.1) and is not known by the NWDAF. To determine NF load (per area of interest), NWDAF may collect and take into account UE trajectory input data from the AF, defined in clause 6.7.2.2, Table 6.7.2.2-2 for UE mobility analytics in addition to MDT input data and /or collective behaviour input data, defined in clause 6.5.2, Table 6.5.2-3 and Table 6.5.2-5, respectively.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.5.3 Output analytics	The NWDAF services as defined in the clause 7.2 and 7.3 are used to expose the analytics. NF load statistics information are defined in Table 6.5.3-1. NF load predictions information are defined in Table 6.5.3-2. Table 6.5.3-1: NF load statistics Information Description List of resource status (1..max) List of observed load information for each NF instance along with the corresponding NF id / NF Set ID (as applicable). > NF type Type of the NF instance. > NF instance ID Identification of the NF instance. > NF status (NOTE 1) The availability status of the NF on the Analytics target period, expressed as a percentage of time per status value (registered, suspended, undiscoverable). > NF resource usage (NOTE 1) The average usage of assigned resources (CPU, memory, disk). > NF load (NOTE 1) The average load of the NF instance over the Analytics target period. > NF peak load (NOTE 1) The maximum load of the NF instance over the Analytics target period. > NF load (per area of interest) (NOTE 1, NOTE 2) The average load of the NF instances over the area of interest. NOTE 1: Analytics subset that can be used in "list of analytics subsets that are requested" and "Preferred level of accuracy per analytics subset". NOTE 2: Applicable only to AMF load based on Input data in clause 6.5.2, Table 6.5.2-3 and Table 6.5.2-5. Table 6.5.3-2: NF load predictions Information Description List of resource status (1..max) List of predicted load information for each NF instance along with the corresponding NF id / NF Set ID (as applicable) > NF type Type of the NF instance > NF instance ID Identification of the NF instance > NF status (NOTE 1) The availability status of the NF on the Analytics target period, expressed as a percentage of time per status value (registered, suspended, undiscoverable) > NF resource usage (NOTE 1) The average usage of assigned resources (CPU, memory, disk) > NF load (NOTE 1) The average load of the NF instance over the Analytics target period > NF peak load (NOTE 1) The maximum load of the NF instance over the Analytics target period > Confidence Confidence of this prediction > NF load (per area of interest) (NOTE 1, NOTE 2) The predicted average load of the NF instances over the area of interest. NOTE 1: Analytics subset that can be used in "list of analytics subsets that are requested" and "Preferred level of accuracy per analytics subset". NOTE 2: Applicable only to AMF load based on Input data in clause 6.5.2, Table 6.5.2-3 and Table 6.5.2-5. NOTE: The variations on per-instance NF load and resource usage could be influenced by the number of running NF instances in addition to the load itself. The predictions are provided with a Validity Period, as defined in clause 6.1.3. The number of resource status is limited by the maximum number of objects provided as part of Analytics Reporting Information.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.5.4 Procedures	The procedure depicted in Figure 6.5.4-1 allows a consumer NF to request analytics to NWDAF for NF load of various NF instances as defined in 6.5.1. Figure 6.5.4-1: NF load analytics provided by NWDAF 1. The NF sends a request to the NWDAF for analytics for NF load for a specific NF, using either the Nnwdaf_AnalyticsInfo or Nnwdaf_AnalyticsSubscription service. The Analytics ID is set to NF load information, the Target of Analytics Reporting and the Analytics Filter Information are set according to clause 6.5.1. The NF can request statistics or predictions or both and can provide a time window. 2-5. If the request is authorized and in order to provide the requested analytics, the NWDAF may need for each NF targeted instance to subscribe to OAM services to retrieve the target NF resource usage and NF resources configuration following steps captured in clause 6.2.3.2 for data collection from OAM. The NWDAF may collect MDT input data per individual UE from OAM (see Table 6.5.2-3). Steps 2-5 may be skipped when e.g. the NWDAF already has the requested analytics. 6. For Collective Behaviour attributes, if the request is authorized and in order to provide the requested analytics, NWDAF may follow the UE Input Data Collection Procedure via the AF as defined in clause 6.2.8 (see Table 6.5.2-4 and Table 6.5.2-5). The NWDAF subscribes to the AF services as above invoking either Nnef_EventExposure_Subscribe or Naf_EventExposure_Subscribe service (Event ID = Collective Behaviour, Event Filter information, Target of Event Reporting) as defined in TS 23.502 [3]. The area of interest is set as part of Event Filter information to specific TAs or AMF region. The UE data is collected from UEs within the area of interest. In the case of trusted AF, the NWDAF provides the Area of Interest as a list of TAIs to the AF. In the case of untrusted AF, NEF translates the requested Area of Interest provided as event filter by the NWDAF into geographic zone identifier(s) that act as event filter for the AF. For collective attributes as defined in Table 6.5.2-5, the AF processes (e.g. anonymize, aggregate and normalize) the data from individual UEs per UE attribute (see Table 6.5.2-4) based on Event Filters indicated by the NWDAF to determine which ones display a collective behaviour within the area of interest before notifying a collective attribute directly (trusted AF) or via NEF (for untrusted AF) to the NWDAF. The AF will provide (per collective attribute) e.g. the number of UEs that fulfil the collective attribute (within an area of interest). NOTE 1: The call flow only shows a subscription/notification model for the simplicity, however both request-response and subscription-notification models should be supported. NOTE 2: If the target NF type is UPF, the NWDAF can collect the information as listed in Table 6.5.2-2. How the NWDAF collects information is defined in clause 5.8.2.17 of TS 23.501 [2] and in clause 4.15.4 of TS 23.502 [3]. 7a. The NWDAF subscribes to changes on the load and status of NF instances registered in NRF and identified by their NF id from NRF using Nnrf_NFManagement_NFStatusSubscribe service operation for each NF instance. 7b. NRF notifies NWDAF of changes on the load and status of the requested NF instances by using Nnrf_NFManagement_NFStatusNotify service operation. 8. The NWDAF derives requested analytics. 9. The NWDAF provides requested NF load analytics to the NF along with the corresponding Validity Period (only for predictions) or area of interest, using either the Nnwdaf_AnalyticsInfo_Request response or Nnwdaf_AnalyticsSubscription_Notify, depending on the service used in step 1. 10-12. If at step 1 the NF has subscribed to receive continuous reporting of NF load analytics, the NWDAF may generate new analytics and, when relevant according to the Analytics target period and Reporting Threshold, provide them along with the corresponding Validity Period (only for predictions) to the NF upon reception of notification of new NF load information from OAM or NRF or UE Input data notification via MDT or the AF (see Table 6.5.2-3 and Table 6.5.2-5). NOTE 3: If the target NF type at step 1 is UPF, the NWDAF can generate new analytics when receiving new information as listed in Table 6.5.2-2.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.6 Network Performance Analytics
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.6.1 General	With Network Performance Analytics, NWDAF provides either statistics or predictions on the gNB status information, gNB resource usage, communication performance and mobility performance in an Area of Interest. In addition, NWDAF can provide statistics or predictions on the number of UEs located in that Area of Interest. The service consumer may be an NF (e.g. PCF, NEF, AF), or the OAM. The consumer of these analytics may indicate in the request: - Analytics ID = "Network Performance"; - Target of Analytics Reporting as defined in clause 6.1.3; - Analytics Filter Information: - Area of Interest (list of TAs or Cell IDs) which restricts the area in focus (mandatory if Target of Analytics Reporting is set to "any UE", optional otherwise); - Optionally, Traffic type of interest (overall traffic, GBR traffic or Delay-critical GBR traffic); NOTE: If Traffic type of interest is not provided, overall traffic is considered. - Optionally, a list of analytics subsets that are requested among those specified in clause 6.6.3; - Optionally, a preferred level of accuracy of the analytics; - Optionally, preferred level of accuracy per analytics subset (see clause 6.6.3); - Optionally, preferred order of results for the list of Network Performance information: - ordering criterion: "number of UEs", "communication performance" or "mobility performance"; - order: ascending or descending; - Optionally, Reporting Thresholds, which apply only for subscriptions and indicate conditions on the level to be reached for respective analytics information (see clause 6.6.3) in order to be notified by the NWDAF; - An Analytics target period indicates the time period over which the statistics or prediction are requested; and - Optionally, maximum number of objects. - In a subscription, the Notification Correlation Id and the Notification Target Address are included. - Optionally, Spatial granularity size (if an Area of Interest is provided) and Temporal granularity size. The NWDAF notifies the result of the analytics to the consumer as indicated in clause 6.6.3.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.6.2 Input Data	The NWDAF collects Load and Performance information in an Area of Interest from the sources listed in Table 6.6.2-1 and number of UEs within Area of Interest from the sources listed in Table 6.6.2-2. Table 6.6.2-1: Load and Performance information collected by NWDAF Load information Source Description Status, load and performance information OAM Statistics on RAN status (up/down), load (i.e. Radio Resource Utilization) and performance per Cell Id for the traffic type of interest and in the Area of Interest as defined in TS 28.552 [8]. NF Load information NRF Load per NF Table 6.6.2-2: Number of UEs in Area of Interest information collected by NWDAF Number of UEs information Source Description Number of UEs AMF Number of UEs in an Area of Interest The NWDAF shall be able to collect UE mobility information as stated in clause 6.7.2.2.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.6.3 Output Analytics	The NWDAF shall be able to provide both statistics and predictions on Network Performance. Network performance statistics are defined in Table 6.6.3-1. Table 6.6.3-1: Network performance statistics Information Description List of network performance information (1..max) Observed statistics during the Analytics target period. > Area subset List of TAs or Cell IDs within the requested area of interest as defined in clause 6.6.1. If a Spatial granularity size was provided in the request or subscription, the number of elements of the list is smaller than or equal to the Spatial granularity size. > Analytics target period subset Time window within the requested Analytics target period as defined in clause 6.6.1. If a Temporal granularity size was provided in the request or subscription, the duration of the Analytics target period subset is greater than or equal to the Temporal granularity size. > gNB status information (NOTE 1) Average ratio of gNBs that have been up and running during the entire Analytics target period in the area subset. > gNB resource usage (NOTE 1) (NOTE 2) Usage of assigned resources (average, peak). > gNB resource usage for GBR traffic (NOTE 1) (NOTE 2) (NOTE 3) Usage of assigned resources for GBR traffic (average, peak). > gNB resource usage for Delay-critical GBR traffic (NOTE 1) (NOTE 2) (NOTE 3) Usage of assigned resources for Delay-critical GBR traffic (average, peak). > Number of UEs (NOTE 1) Average number of UEs observed in the area subset. > Communication performance (NOTE 1) Average ratio of successful setup of PDU Sessions. > Mobility performance (NOTE 1) Average ratio of successful handover. NOTE 1: Analytics subset that can be used in "list of analytics subsets that are requested" and "Preferred level of accuracy per analytics subset". NOTE 2: The average and peak usage of uplink and downlink traffic are provided as percentage. NOTE 3 The resource usage (average, peak) for GBR and Delay-critical GBR traffic types can be computed using the sub-counters of their corresponding 5QI measurements, as defined in clause 5.1.1.2 of TS 28.552 [8]. Network performance predictions are defined in Table 6.6.3-2. Table 6.6.3-2: Network performance predictions Information Description List of network performance information (1..max) Predicted analytics during the Analytics target period > Area subset List of TAs or Cell IDs within the requested area of interest as defined in clause 6.6.1. If a Spatial granularity size was provided in the request or subscription, the number of elements of the list is smaller than or equal to the Spatial granularity size. > Analytics target period subset Time window within the requested Analytics target period as defined in clause 6.6.1. If a Temporal granularity size was provided in the request or subscription, the duration of the Analytics target period subset is greater than or equal to the Temporal granularity size. > gNB status information (NOTE 1) Average ratio of gNBs that will be up and running during the entire Analytics target period in the area subset. > gNB resource usage (NOTE 1) (NOTE 2) Usage of assigned resources (average, peak) > gNB resource usage for GBR traffic (NOTE 1) (NOTE 2) (NOTE 3) Usage of assigned resources for GBR traffic (average, peak). > gNB resource usage for Delay-critical GBR traffic (NOTE 1) (NOTE 2) (NOTE 3) Usage of assigned resources for Delay-critical GBR traffic (average, peak). > Number of UEs (NOTE 1) Average number of UEs predicted in the area subset. > Communication performance (NOTE 1) Average ratio of successful setup of PDU Sessions. > Mobility performance (NOTE 1) Average ratio of successful handover. > Confidence Confidence of this prediction. NOTE 1: Analytics subset that can be used in "list of analytics subsets that are requested" and "Preferred level of accuracy per analytics subset". NOTE 2: The average and peak usage of uplink and downlink traffic are provided as percentage. NOTE 3: The resource usage (average, peak) for GBR and Delay-critical GBR traffic types can be computed using the sub-counters of their corresponding 5QI measurements, as defined in clauses 5.1.1.2 of TS 28.552 [8]. NOTE 1: The predictions are provided with a Validity Period, as defined in clause 6.1.3. NOTE 2: The analytics on number of UEs are related to the information retrieved from the AMFs. The number of network performance information entries is limited by the maximum number of objects provided as part of Analytics Reporting Information. The NWDAF provides Network Performance Analytics to a consumer at the time requested by the consumer in the Analytics target period: - Analytics ID set to "Network Performance". - Notification Target Address including the address of the consumer. - Notification Correlation ID, for the consumer to correlate notifications from NWDAF if subscription applies. - Analytics specific parameters at the time indicated in the Analytics target period.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.6.4 Procedures	Figure 6.6.4-1: Procedure for subscription to network performance analytics 1. The NF sends Nnwdaf_AnalyticsSubscription_Subscribe or Nnwdaf_AnalyticsInfo_Request (Analytics ID="Network Performance", Target of Analytics Reporting, Analytics Filter Information = "Area of Interest", Analytics Reporting Information = ("Reporting Thresholds" and Analytics target Period(s))) to the NWDAF. 2a-2d. The NWDAF discovers from NRF the AMF(s) belonging to the AMF Region(s) that include(s) the Area of Interest and subscribes to NF load and status information from NRF about these AMF(s). 3a-3b. The NWDAF subscribes to OAM services to get the status and load information and the resource usage on the Area of Interest in clause 6.6.2, following the procedure captured in Clause 6.2.3.2. 4a-4b. The NWDAF collects the number of UEs located in the Area of Interest from AMF using Namf_EventExposure_Subscribe service, including the Target of Event Reporting provided as an input parameter (i.e. any UE or Internal Group Identifier). 5. The NWDAF derives the requested analytics. 6. The NWDAF sends Nnwdaf_AnalyticsSubscription_Notify or Nnwdaf_AnalyticsInfo_Request response (Network Performance analytics, Subscription Correlation Id, Confidence). 7-8. A change of network performance information, i.e. change in the gNB status information, gNB resource usage, communication performance and mobility performance in the area of interest at the observed period, is detected by OAM, or a change in the NF load information is reported by NRF and is notified to NWDAF. 9. The NWDAF derives new analytics taking into account the most recent data collected. 10. When relevant according to the Analytics target period and Reporting Thresholds, the NWDAF provides a notification using Nnwdaf_AnalyticsSubscription_Notify (Network Performance analytics, Subscription Correlation Id, Confidence).
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7 UE related analytics
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.1 General	This clause specifies the UE related analytics which can be provided by NWDAF: - UE mobility analytics; - UE communication analytics; - Expected UE behavioural parameters related network data analytics; and - Abnormal behaviour related network data analytics. The NWDAF service consumer may request for these analytics separately, or in a combined way. As an example, an NWDAF service consumer may learn from the NWDAF the expected UE behaviour parameters as defined in clause 4.15.6.3 of TS 23.502 [3], by requesting analytics for both UE mobility (see clause 6.7.2) and for UE communication (see clause 6.7.3). Depending on local regulations, the NWDAF retrieves user consent for the UE with UDM prior to data collection as defined in clause 6.2.2.2 or clause 6.2.2.3. If user consent to collect data is not granted by the UE, the NWDAF rejects/cancels any analytics subscriptions to any of the UE related analytics with target for analytics set to the SUPI or GPSI of that UE. If the target for analytics is either an Internal or External Group Id or a list of SUPIs or "any UE", the NWDAF skips those SUPIs that do not grant user consent for the purpose of analytics or model training. NOTE: Possible uses of such analytics is for the AMF to learn about expected UE behaviour to derive appropriate MICO mode configuration, or for an AF to learn about expected UE behaviour to further provision 5GC with appropriate UE parameters.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.2 UE mobility analytics
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.2.1 General	NWDAF supporting UE mobility statistics or predictions shall be able to collect UE mobility related information from NF, OAM and to perform data analytics to provide UE mobility statistics or predictions. The service consumer may be a NF (e.g. AMF, SMF or AF). The consumer of these analytics may indicate in the request: - Analytics ID = "UE Mobility". - Target of Analytics Reporting: a single UE (SUPI) or group of UEs (i.e. a list of Internal Group Ids); - Analytics Filter Information optionally containing: - Area of Interest (AOI): restricts the scope of the UE mobility analytics to the provided area. If the request is for fine granularity location information (i.e. with a finer granularity than cell), the AOI may be described as shown in clause 5.5 of TS 23.273 [39]; NOTE 1: For LADN service, the consumer (e.g. SMF) provides the LADN DNN to refer the LADN service area as the AOI. - Visited Area(s) of Interest (visited AOI(s)): additional filter to only consider UEs that are currently (i.e. now) in the "AOI" and had previously (i.e. in the "Analytics target period") been in at least one of the Visited AOI(s). If this parameter is provided, the Analytics target period shall be in the past (i.e. supported for statistics only); - Linear distance threshold: An event where the UE moves by more than some predefined straight line distance from a previous location as per TS 23.273 [39]. The consumer can provide more than one value of Linear distance threshold. - an optional list of analytics subsets that are requested (see clause 6.7.2.3); - An Analytics target period indicates the time period over which the statistics or predictions are requested; NOTE 2: For regular analytics scenarios, the Analytics target period is associated with the Analytics Filter Information = AOI, while for the scenario that Analytics ID=UE Mobility and Analytics Filter Information = (AOI and visited AOI(s)), as described in this clause, the Analytics target period is associated with the visited AOI(s) and to obtain the statistics for those UEs that currently reside in the AOI and had previously (i.e. in the "Analytics target period") been in at least one of the Visited AOI(s). - Optionally, maximum number of objects; - Preferred level of accuracy of the analytics; - Optionally, Preferred level of accuracy per analytics subset (see clause 6.7.2.3); - Preferred order of results for the time slot entries: ascending or descending time slot start; - Optionally, preferred granularity of location information: TA level or cell level "longitude and latitude level"; NOTE 3: Definition of "longitude and latitude level" is described in clause 6.1.3. - Optionally, Preferred orientation of location information: ("horizontal", "vertical", "both"); - Optionally, Spatial granularity size and Temporal granularity size; - UE Location order indicator: indicates the NWDAF should derives and provides the UE Mobility analytics for UE Location in time order; and - In a subscription, the Notification Correlation Id and the Notification Target Address are included.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.2.2 Input Data	The NWDAF supporting data analytics on UE mobility shall be able to collect UE mobility information from OAM, 5GC and AFs. The detailed information collected by the NWDAF could be MDT data from OAM, network data from 5GC and/or service data from AFs: - UE mobility information from OAM is UE location carried in MDT data; - Network data related to UE mobility from 5GC is UE location information, UE location trends or UE access behaviour trends, as defined in the Table 6.7.2.2-1; Table 6.7.2.2-1: UE Mobility information collected from 5GC Information Source Description UE ID AMF SUPI. UE locations (1..max) AMF UE positions. >UE location TA or cells that the UE enters (NOTE 1). >Timestamp A time stamp when the AMF detects the UE enters this location. Fine granularity location (1...max) LCS (NOTE 2) UE positions. ....>UE location GAD shape or location coordinates (see TS 23.032 [34]). ....>Timestamp A time stamp when the location was measured. ....>LCS QoS LCS QoS accuracy as defined in clause 4.1b of TS 23.273 [39]. ....>Motion Event Notification) The notification about motion event reporting as described in TS 23.273 [39]. ....>Liner distance threshold NWDAF consumer The distance travelled by the UE before reporting subsequent location as described in TS 23.273 [39]. Type Allocation code (TAC) AMF To indicate the terminal model and vendor information of the UE. The UEs with the same TAC may have similar mobility behaviour. The UE whose mobility behaviour is unlike other UEs with the same TAC may be an abnormal one. Frequent Mobility Registration Update AMF A UE (e.g. a stationary UE) may re-select between neighbour cells due to radio coverage fluctuations. This may lead to multiple Mobility Registration Updates if the cells belong to different registration areas. The number of Mobility Registration Updates N within a period M may be an indication for abnormal ping-pong behaviour, where N and M are operator's configurable parameters. UE access behaviour trends AMF Metrics on UE state transitions (e.g. access, RM and CM states, handover). UE location trends AMF Metrics on UE locations. NOTE 1: UE location includes either the last known location or the current location, under the conditions defined in Table 4.15.3.1-1 in TS 23.502 [3]. NOTE 2: The procedure to collect location data using LCS is described in clause 6.2.12. - Service data related to UE mobility provided by AFs is defined in the Table 6.7.2.2-2; Table 6.7.2.2-2: Service Data from AF related to UE mobility Information Description UE ID Could be external UE ID (i.e. GPSI). Application ID Identifying the application providing this information. UE trajectory (1..max) Timestamped UE positions. >UE location Geographical area that the UE enters. >Timestamp A time stamp when UE enters this area. A list of areas A list of areas used by the AF for the application service. NOTE: The application ID is optional. If the application ID is omitted, the collected UE mobility information can be applicable to all the applications for the UE. Depending on the requested level of accuracy, data collection may be provided on samples (e.g. spatial subsets of UEs or UE group, temporal subsets of UE location information). NOTE: Reporting current UE location can cause AMF to request NG-RAN to report UE location and consequently extra signalling and load in NG-RAN and AMF. The consumer retrieving data from AMF needs to use current location with care to avoid excessive signalling.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.2.3 Output Analytics	The NWDAF supporting data analytics on UE mobility shall be able to provide UE mobility analytics to consumer NFs or AFs. The analytics results provided by the NWDAF could be UE mobility statistics as defined in table 6.7.2.3-1, UE mobility predictions as defined in Table 6.7.2.3-2: Table 6.7.2.3-1: UE mobility statistics Information Description UE group ID or UE ID Identifies the UE(s) for which the statistic applies by a list of SUPIs, or a group of UEs by a list of Internal-Group-Ids defined in clause 5.9.7 of TS 23.501 [2] (see NOTE 1). Time slot entry (1..max) List of time slots during the Analytics target period. > Time slot start Time slot start within the Analytics target period. > Duration Duration of the time slot. If a Temporal granularity size was provided in the request or subscription, the Duration is greater than or equal to the Temporal granularity size. > UE location (1..max) Observed location statistics (see NOTE 2). >> UE location (NOTE 5) TAs or cells which the UE stays or geographical location (longitude and latitude level) (see NOTE 3). >> Ratio (NOTE 5) Percentage of UEs in the group (in the case of a UE group). >> UE's geographical distribution (NOTE 5) The geographical distribution of the UEs among the TAs or cells or location coordinates. >> Requested Linear Distance Threshold (NOTE 4) The linear distance threshold used for UE location reporting. >> Geographical Identifier (NOTE 5) Geographical Identifier as specified in TS 23.228 [47] (see NOTE 6). > UE's direction (NOTE 5) The direction of the UEs in the Area of Interest. NOTE 1: When Target of Analytics Reporting is an individual UE, one UE ID (i.e. SUPI) will be included, the NWDAF will provide the analytics mobility result (i.e. list of (predicted) time slots) to NF service consumer(s) for the UE. NOTE 2: If Visited AOI(s) was provided in the analytics request/subscription, the UE location provides information on the observed location(s) that the UE or group of UEs had been residing during the Analytics Target Period. NOTE 3: When possible and applicable to the access type, UE location is provided according to the preferred granularity of location information and Spatial granularity size. NOTE 4: The requested Linear Distance Threshold is provided only when in the analytic filter information of the analytics request there are multiple linear distance thresholds and the target is a single UE. NOTE 5: Analytics subset that can be used in "list of analytics subsets that are requested" and "Preferred level of accuracy per analytics subset". NOTE 6: It depends on the implementation how the NWDAF collects the geographic identifier. Table 6.7.2.3-2: UE mobility predictions Information Description UE group ID or UE ID Identifies the UE(s) for which the prediction applies by a list of SUPIs, or a group of UEs by a list of Internal-Group-Ids defined in clause 5.9.7 of TS 23.501 [2] (see NOTE 1). Time slot entry (1..max) List of predicted time slots. >Time slot start Time slot start time within the Analytics target period. > Duration Duration of the time slot. If a Temporal granularity size was provided in the request or subscription, the Duration is greater than or equal to the Temporal granularity size. > UE location (1..max) Predicted location prediction during the Analytics target period. >> UE location (NOTE 3) TAs or cells where the UE or UE group may move into or geographical location (longitude and latitude level) (see NOTE 2). >> Confidence Confidence of this prediction. >> Ratio (NOTE 3) Percentage of UEs in the group (in the case of a UE group). >> UE's geographical distribution (NOTE 3) The geographical distribution of the UEs among the TAs or cells or location coordinates. >> Geographical Identifier (NOTE 3) Geographical Identifier as specified in TS 23.228 [47] (see NOTE 4). > UE's direction (NOTE 3) The direction of the UEs in the Area of Interest. NOTE 1: When Target of Analytics Reporting is an individual UE, one UE ID (i.e. SUPI) will be included, the NWDAF will provide the analytics mobility result (i.e. list of (predicted) time slots) to NF service consumer(s) for the UE. NOTE 2: When possible and applicable to the access type, UE location is provided according to the preferred granularity of location information and Spatial granularity size. NOTE 3: Analytics subset that can be used in "list of analytics subsets that are requested" and "Preferred level of accuracy per analytics subset". NOTE 4: It depends on the implementation how the NWDAF collects the geographic identifier. The results for UE groups address the group globally. The ratio is the proportion of UEs in the group at a given location at a given time. The number of time slots and UE locations is limited by the maximum number of objects provided as part of Analytics Reporting Information. The time slots shall be provided by order of time, possibly overlapping. The locations shall be provided by decreasing value of ratio for a given time slot. The sum of all ratios on a given time slot must be equal or less than 100%. Depending on the list size limitation, the least probable locations on a given Analytics target period may not be provided. If a UE Location order indicator is included in the Analytics Reporting information, the NWDAF does not aggregate the UE locations in a long duration but provides the UE locations one by one in their own time period, i.e. the "UE location (1..max)" in the UE Mobility analytics has only one UE location (TA, Cell or a finer granularity UE Location smaller than cell) which indicates the UE is located in this UE location in the duration from the time slot start (i.e. time stamp when the UE enters this location as described in clause 6.7.2.2).
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.2.4 Procedures	The NWDAF can provide UE mobility related analytics, in the form of statistics or predictions or both, directly to another NF. If the NF is an AF and when the AF is untrusted, the AF will request analytics via the NEF and the NEF will then convey the request to NWDAF. NOTE: In the case of untrusted AF the Target of Analytics Reporting can be a GPSI or an External Group Identifier that is mapped in the 5GC to a SUPI or an Internal Group Identifier. Figure 6.7.2.4-1: UE mobility analytics provided to an Analytics Service Consumer 1. The NF sends a request (Analytics ID = UE mobility, Target of Analytics Reporting = UE id or Internal Group ID, Analytics Filter Information = AOI, Analytics Reporting Information= Analytics target period and/or UE Location order indicator) to the serving NWDAF for analytics information on a specific UE or group of UEs, i.e. list of Internal-Group-Ids, using either the Nnwdaf_AnalyticsInfo or Nnwdaf_AnalyticsSubscription service to derive UE mobility information. The NF can request statistics or predictions or both. For LADN service, the NF (i.e. SMF) provides LADN DNN as AOI in the Analytics Filter Information. If NF wants to obtain the aggregated mobility analytics of those UEs, that currently reside in the AOI and had visited at least one of visited AOI(s) during an Analytics target period, the NF may send a request for UE mobility analytics with Analytics ID = UE mobility, Target of Analytics Reporting = UE group ID or UE ID, Analytics Filter Information = (AOI, visited AOI(s)), Analytics Reporting information = Analytics target period. In this case, the requested mobility analytics is a statistics. 2. If the request is authorized and in order to provide the requested analytics, the NWDAF may subscribe to events with all the serving AMFs for the requested UE(s), for notification of location changes. This step may be skipped when e.g. the NWDAF already has the requested analytics available. The NWDAF subscribes the service data for the requested UE(s) from AF(s) in the Table 6.7.2.2-2 by invoking Naf_EventExposure_Subscribe service or Nnef_EventExposure_Subscribe (if via NEF) using event ID "UE Mobility information" as defined in TS 23.502 [3]. The NWDAF collects UE mobility information from OAM for the requested UE(s), following the procedure captured in clause 6.2.3.2. The NWDAF may collect UE location information from the GLMC, which may initiate the UE location service procedure and gets the location of each requested UE(s), if the consumer requested fine granularity location information and/or one or more location requests corresponding to the linear distance threshold values in the analytics request according to clause 6.7.2.1. NOTE 1: The NWDAF determines the serving AMF(s) as described in clause 6.2.2.1. 3. The NWDAF derives requested analytics. If in step 1 the NWDAF receives analytics subscription/request from NF to obtain the aggregated mobility analytics of those UEs, which currently reside in AOI and had visited at least one of visited AOI(s) during an Analytics target period and if visited AOI(s) and AOI are covered by different NWDAFs, in addition to the data collected in the AOI in step 2, the NWDAF can also obtain UE mobility analytics in one of the visited AOI(s) during the Analytics target period from other NWDAF instance(s) for the requested UE(s). Then the NWDAF supporting analytics aggregation capability derives a UE ID list based on the request from the NF in step 1 and the requested aggregated analytics based on the data collected in the AOI in step 2 and UE mobility analytics in one or more of the visited AOI(s) obtained from the other NWDAF instance(s). UE visited locations in visited AOI(s) and AOI will be included in the aggregated UE mobility analytics. NOTE 2: If the visited AOI(s) and AOI are covered by different NWDAFs, then consumer in the AOI firstly discovers a NWDAF supporting analytics aggregation capability in the AOI from the NRF, as defined in clause 6.3.13 of TS 23.501 [2]. 4. The NWDAF provides requested UE mobility analytics to the NF, using either the Nnwdaf_AnalyticsInfo_Request response or Nnwdaf_AnalyticsSubscription_Notify, depending on the service used in step 1. The details for UE mobility analytics provided by NWDAF are defined in clause 6.7.2.3. If in step 1 the NF wants to obtain the aggregated mobility analytics of those UEs, that currently reside in the AOI and had visited at least one of visited AOI(s) during an Analytics target period, the NWDAF will provide the requested aggregated analytics for the UE(s) matching this criteria, i.e. the derived mobility analytics can cover a subset of UEs compared to the Target of Analytics Reporting as provided in step 1. 5-6. If at step 1, the NF has subscribed to receive notifications for UE mobility analytics, after receiving event notification from the AMFs, AFs, GMLC and OAM subscribed by NWDAF in step 2, the NWDAF may generate new analytics and provide them to the NF.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.3 UE Communication Analytics
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.3.1 General	In order to support some optimized operations, e.g. customized mobility management, traffic routing handling, RFSP Index Management, QoS improvement or Inactivity Timer optimization, in 5GS, an NWDAF may perform data analytics on UE communication pattern and user plane traffic and provide the analytics results (i.e. UE communication statistics or prediction) to NFs in the 5GC or an AF. An NWDAF supporting UE Communication Analytics collects per-application communication description from AFs. If consumer NF provides an Application ID, the NWDAF only considers the data from AF, SMF and UPF that corresponds to this application ID. NWDAF may also collect data from AMF. The consumer of these analytics may indicate in the request: - Analytics ID = "UE Communication". - Target of Analytics Reporting: a single UE (SUPI) or group of UEs (a list of Internal-Group-Ids). - Analytics Filter Information optionally including: - S-NSSAI; - DNN; - Application ID; - Area of Interest. - an optional list of analytics subsets that are requested (see clause 6.7.3.3); - An Analytics target period indicates the time period over which the statistics or predictions are requested. - Preferred level of accuracy of the analytics. - Optional Preferred level of accuracy per analytics subset (see clause 6.7.3.3); - Optional preferred order of results for the list of UE Communications: - ordering criterion: "start time" or "duration", - order: ascending or descending; - Optionally, maximum number of objects; - Optionally, Spatial granularity size (if an Area of Interest is provided); and - In a subscription, the Notification Correlation Id and the Notification Target Address are included.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.3.2 Input Data	The NWDAF supporting data analytics on UE communication shall be able to collect communication information for the UE from 5GC. The detailed information collected by the NWDAF includes service data related to UE communication as defined in the Table 6.7.3.2-1. Table 6.7.3.2-1: Service Data from 5GC related to UE communication Information Source Description UE ID SMF, AF SUPI in the case of SMF, external UE ID (i.e. GPSI) in the case of AF Group ID SMF, AF To identify UE group if available Internal Group ID in the case of SMF, External Group ID in the case of AF S-NSSAI SMF Information to identify a Network Slice DNN SMF Data Network Name where PDU connectivity service is provided Application ID SMF, AF Identifying the application providing this information Expected UE Behaviour parameters AF Same as Expected UE Behaviour parameters specified in TS 23.502 [3] UE communication (1..max) UPF, AF Communication description per application >Communication start The time stamp that this communication starts >Communication stop The time stamp that this communication stops >UL data rate UL data rate of this communication >DL data rate DL data rate of this communication >Traffic volume Traffic volume of this communication Type Allocation code (TAC) AMF To indicate the terminal model and vendor information of the UE. The UEs with the same TAC may have similar communication behaviour. The UE whose communication behaviour is unlike other UEs with the same TAC may be an abnormal one. UE locations (1..max) AMF UE positions >UE location TA or cells that the UE enters >Timestamp A time stamp when the AMF detects the UE enters this location UE location trends AMF Metrics on UE locations. PDU Session ID (1..max) SMF Identification of PDU Session. > Inactivity detection time SMF, UPF Value of session inactivity timer. > PDU Session status SMF Status of the PDU Session (activated, deactivated). UE CM state AMF UE connection management state (e.g. CM-IDLE). UE session behaviour trends SMF Metrics on UE state transitions (e.g. "PDU Session Establishment", "PDU Session Release"). UE communication trends SMF Metrics on UE communications. UE access behaviour trends AMF Metrics on UE state transitions (e.g. access, RM and CM states, handover). Depending on the requested level of accuracy, data collection may be provided on samples (e.g. spatial subsets of UEs or UE group, temporal subsets of UE communication information). The application Id is optional. If the application Id is omitted, the collected UE communication information can be applicable to all the applications for the UE.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.3.3 Output Analytics	The NWDAF supporting UE Communication Analytics provides the analytics results to consumer NFs. The analytics results provided by the NWDAF include the UE communication statistics as defined in Table 6.7.3.3-1 or predictions as defined in Table 6.7.3.3-2. Table 6.7.3.3-1: UE Communication Statistics Information Description UE group ID or UE ID Identifies the UE(s) for which the statistic applies by a list of SUPIs, or a group of UEs by a list of Internal-Group-Ids defined in clause 5.9.7 of TS 23.501 [2]. UE communications (1..max) (NOTE 1) List of communication time slots. > Periodic communication indicator (NOTE 1) Identifies whether the UE communicates periodically or not. > Periodic time (NOTE 1) Interval Time of periodic communication (average and variance) if periodic. Example: every hour > Start time (NOTE 1) Start time observed (average and variance) > Duration (NOTE 1) Duration of communication (average and variance). > Traffic characterization S-NSSAI, DNN, ports, other useful information. > Traffic volume (NOTE 1) Volume UL/DL (average and variance). > Ratio Percentage of UEs in the group (in the case of a UE group). Applications (0..max) (NOTE 1) List of applications in use. > Application Id Identification of the application. > Start time Start time of the application. > Duration time Duration interval time of the application. > Occurrence ratio Proportion for the application used by the UE during requested period. > Spatial validity Area where the service behaviour applies. If Area of Interest information was provided in the request or subscription, spatial validity may be a subset of the requested Area of Interest. N4 Session ID (1..max) (NOTE 1) (NOTE 2) Identification of N4 Session. > Inactivity detection time Value of session inactivity timer (average and variance). NOTE 1: Analytics subset that can be used in "list of analytics subsets that are requested" and "Preferred level of accuracy per analytics subset". NOTE 2: This analytics subset shall only be included if the consumer is SMF. Table 6.7.3.3-2: UE Communication Predictions Information Description UE group ID or UE ID Identifies the UE(s) for which the statistic applies by a list of SUPIs, or a group of UEs by a list of Internal-Group-Ids defined in clause 5.9.7 of TS 23.501 [2]. UE communications (1..max) (NOTE 1) List of communication time slots. > Periodic communication indicator (NOTE 1) Identifies whether the UE communicates periodically or not. > Periodic time (NOTE 1) Interval Time of periodic communication (average and variance) if periodic. Example: every hour. > Start time (NOTE 1) Start time predicted (average and variance). > Duration time (NOTE 1) Duration interval time of communication. > Traffic characterization S-NSSAI, DNN, ports, other useful information. > Traffic volume (NOTE 1) Volume UL/DL (average and variance). > Confidence Confidence of the prediction. > Ratio Percentage of UEs in the group (in the case of a UE group). Applications (0..max) (NOTE 1) List of applications in use. > Application Id Identification of the application. > Start time Start time of the application. > Duration time Duration interval time of the application. > Occurrence probability Probability the application will be used by the UE. > Spatial validity Area where the service behaviour applies. If Area of Interest information was provided in the request or subscription, spatial validity may be a subset of the requested Area of Interest. If a Spatial granularity size was provided in the request or subscription, the number of elements of TAs or cells in the area is smaller than or equal to the Spatial granularity size. N4 Session ID (1..max) (NOTE 1) (NOTE 2) Identification of N4 Session. > Inactivity detection time Value of session inactivity timer (average and variance). > Confidence Confidence of the prediction. NOTE 1: Analytics subset that can be used in "list of analytics subsets that are requested" and "Preferred level of accuracy per analytics subset". NOTE 2: This analytics subset shall only be included if the consumer is SMF. NOTE: When Target of Analytics Reporting is an individual UE, one UE ID (i.e. SUPI) will be included, the NWDAF will provide the analytics communication result (i.e. list of (predicted) communication time slots) to NF service consumer(s) for the UE. The results for UE groups address the group globally. The ratio is the proportion of UEs in the group for a given communication at a given time and duration. The number of UE communication entries (1..max) is limited by the maximum number of objects provided as part of Analytics Reporting Information. The communications shall be provided by order of time, possibly overlapping. Depending on the list size limitation, the least probable communications on a given Analytics target period may not be provided.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.3.4 Procedures	The NWDAF can provide UE communication related analytics, in the form of statistics or predictions or both, to a 5GC NF. Figure 6.7.3.4-1: Procedure for UE communication analytics 1. 5GC NF to NWDAF: Nnwdaf_AnalyticsSubscription_Subscribe (Analytics ID = UE communication, Target of Analytics Reporting=SUPI, Analytics Filter Information = (Application ID, Area of Interest, etc.)). 5GC NF sends a request to the NWDAF for analytics on a specific UE(s), using either Nnwdaf_AnalyticsInfo or Nnwdaf_AnalyticsSubscription_Subscribe service. The analytics type indicated by "Analytics ID" is set to "UE communication". The Target of Analytics Reporting is set to SUPI or an Internal Group Identifier and Analytics Filter may include Application ID and Area of Interest. 2a-b. NWDAF to AF (Optional): Naf_EventExposure_Subscribe (Event ID, external UE ID, Application ID, Area of Interest). In order to provide the requested analytics, the NWDAF may subscribe per application communication information, which is identified by Application ID, from AFs for the UE. The Event ID "UE Communication information" as defined in TS 23.502 [3] is used, which indicates communication report for the UE which is requested by the 5GC NF in the step 1. The external UE ID is obtained by the NWDAF based on UE internal ID, i.e. SUPI. In the case of external AF, the NEF translates the requested Area of Interest into a list of geographic zone identifier(s) as described in clause 5.6.7.1 of TS 23.501 [2]. This step is skipped if the NWDAF already has the requested analytics available or has subscribed to the AF. 2c. NWDAF to SMF: Nsmf_EventExposure_Subscribe (Event ID, SUPI, Application ID). In order to provide the requested analytics, the NWDAF subscribes via SMF to UPF information on SUPI, providing e.g. Indication of UPF Event Exposure Service and Target subscription UPF Event Id, Filter Information such as Application ID and/or Area of Interest. This is specified in clause 5.8.2.17 of TS 23.501 [2] and clause 4.15.4 of TS 23.502 [3]. 2d. How SMF subscribes to on UPF is defined in clause 5.8.2.17 of TS 23.501 [2] and in clause 4.15.4 of TS 23.502 [3]. NOTE: The NWDAF request does not trigger any N4 session Establishment/Modification procedure. UPF sends N4 session level reports, including PDU session Inactivity to SMF, according to clause 4.4.2.2 of TS 23.502 [3]. 2f. The UPF provides the requested input data to NWDAF. This is specified in clause 4.15.4 of TS 23.502 [3]. 2g-h. NWDAF to AMF: Namf_EventExposure_Subscribe (Event ID, SUPI, Area of Interest). In order to provide the requested analytics, the NWDAF retrieves one or more of Type Allocation code, UE connection management state, UE access behaviour trends and UE location trends from AMF. NOTE: The NWDAF determines the SMF serving the UE as described in clause 6.2.2.1. 3. The NWDAF derives requested analytics, in the form of UE communication statistics or predictions or both. 4. NWDAF to 5GC NF: Nnwdaf_AnalyticsInfo_Request response or Nnwdaf_AnalyticsSubscription_Notify. The NWDAF provides requested UE communication analytics to the NF, using either Nnwdaf_AnalyticsInfo_Request response or Nnwdaf_AnalyticsSubscription_Notify, depending on the service used in step 1. 5. If the NF subscribed UE communication analytics at step 1, when, based e.g. on new UPF notifications the NWDAF generates new analytics, the NWDAF notifies the new generated analytics to the 5GC NF.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.4 Expected UE behavioural parameters related network data analytics
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.4.1 General	The clause 6.7.4 defines how a service consumer learns from the NWDAF the expected UE behaviour parameters as defined in clause 4.15.6.3 of TS 23.502 [3] for a group of UEs or a specific UE. The service consumer may be an NF (e.g. AMF, AF), or the OAM. The consumer of these analytics shall indicate in the request: - Analytics ID = "UE Mobility" or "UE Communication". - Target of Analytics Reporting: a single UE (SUPI) or group of UEs (i.e. a list of Internal Group Ids). NOTE: In the case of untrusted AF the Target of Analytics Reporting can be a GPSI or an External Group Identifier that is mapped in the 5GC to a SUPI or an Internal Group Identifier - An Analytics target period, which indicates the time period over which the statistics or predictions are requested. - Analytics Filter Information optionally including: - Area of Interest (AOI); - S-NSSAI; - DNN; - Application ID; - an optional list of analytics subsets that are requested (see clause 6.7.3.3). - Optional maximum number of objects. - In a subscription, the Notification Correlation Id and the Notification Target Address are included. The NWDAF shall notify the result of the analytics to the consumer as indicated in clause 6.7.4.3.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.4.2 Input Data	In order to produce "UE Mobility" analytics, the NWDAF collects UE mobility information, UE location trends and/or UE access behaviour trends, as defined in clause 6.7.2.2. In order to produce "UE Communication" analytics, the NWDAF collects UE communication information, UE communication trends, UE session behaviour trends and/or UE access behaviour trends, as defined in clause 6.7.3.2.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.4.3 Output Analytics	The analytics results for "UE Mobility" are specified in Table 6.7.2.3-1 and Table 6.7.2.3-2. The analytics results for "UE Communication" are specified in Table 6.7.3.3-1 and Table 6.7.3.3-2.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.4.4 Procedures
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.4.4.1 NWDAF-assisted expected UE behavioural analytics	Figure 6.7.4.4.1-1: NWDAF assisted expected UE behavioural analytics procedure 1. 5GC NF (e.g. AMF, SMF and AF) to NWDAF: Nnwdaf_AnalyticsInfo_Request (Analytics ID, Target of Analytics Reporting, Analytics Filter Information) or Nnwdaf_AnalyticsSubscription_Subscribe (Analytics ID, Target of Analytics Reporting, Analytics Filter Information). The Analytics ID is set to "UE Mobility" or to "UE Communication"," and the consumer request analytics. 2. If Analytics ID is set to "UE Mobility", the NWDAF collects data from OAM, AMF and/or AF as specified in clause 6.7.2.4 step 2, unless the information is already available. If Analytics ID is set to "UE Communication", the NWDAF collects data from AMF, SMF and/or AF as specified in clause 6.7.3.4 step 2, unless the information is already available. 3. The NWDAF derives requested analytics. 4. NWDAF to 5GC NF: Nnwdaf_AnalyticsInfo_Request response or Nnwdaf_AnalyticsSubscription_Notify. The NWDAF provides requested Expected UE behaviour to the NF, using either Nnwdaf_AnalyticsInfo_Request response or Nnwdaf_AnalyticsSubscription_Notify, depending on the service used in step 1. 5-6. If the NF subscribed to at step 1, when the NWDAF generates new analytics, it provides the new generated analytics to the NF.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.5 Abnormal behaviour related network data analytics
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.5.1 General	This clause defines how to identify a group of UEs or a specific UE with abnormal behaviour, e.g. being misused or hijacked, with the help of NWDAF. NOTE 1: The misused or hijacked UEs are UEs in which there are malicious applications running or UEs which have been stolen. The consumer of this analytics could be a 5GC NF. The 5GC NF subscribes analytics on abnormal behaviour from a NWDAF based on the UE subscription, network configuration or application layer request. The NWDAF performs data analytics on abnormal behaviour if there is a related subscription and returns exception reports that result from the analysis of the correlations between behavioural variables. The exception reports contain an Exception Level expressed in the form of a scalar value, possibly supplemented by additional measurements. The consumer of this analytics shall indicate in the request: - Analytics ID = "Abnormal behaviour"; - Target of Analytics Reporting as defined in clause 6.1.3; - An Analytics target period indicates the time period over which the statistics or predictions are requested; - Analytics Filter Information optionally including: - expected UE behaviour parameters; - expected analytics type or list of Exception IDs with associated thresholds for the Exception Level, where the expected analytics type can be mobility related, communication related or both; - Area of interest; - Application ID; - DNN; - S-NSSAI. NOTE 2: The expected analytics type generally indicates whether mobility or communication related abnormal behaviour analytics or both are expected by the consumer and the list of exception IDs indicates what specific analytics are expected by the consumer. Either the expected analytics type or the list of Exception IDs needs to be indicated, but they are not presented simultaneously. When the expected analytics type is indicated, the NWDAF performs corresponding abnormal behaviour analytics which are supported by the NWDAF. The relation between the expected analytics type and Exception IDs is defined in Table 6.7.5.1-1. - Optionally, maximum number of objects and maximum number of SUPIs; - In a subscription, the Notification Correlation Id and the Notification Target Address are included. Table 6.7.5.1-1: Relation between expected analytics type and Exception IDs Expected analytics type Exception IDs matching the expected analytics type mobility related Unexpected UE location, Ping-ponging across neighbouring cells, Unexpected wakeup, Unexpected radio link failures. communication related Unexpected long-live/large rate flows, Unexpected wakeup, Suspicion of DDoS attack, Wrong destination address, Too frequent Service Access. If the Target of Analytics Reporting is any UE, then the Analytics Filter should at least include: - Area of Interest or S-NSSAI, if the expected analytics type or the list of Exception IDs is mobility related. - Area of Interest, application ID, DNN or S-NSSAI, if the expected analytics type or the list of Exception IDs is communication related. If the Target of Analytics Reporting is any UE, the consumer of this analytics shall request either mobility related only or communication related only abnormal behaviour analytics, but not both at the same time. The expected UE behaviour parameters that the consumer can indicate in the request when known depend on the Exception ID that the consumer expects. They may encompass UE behaviour parameters as defined in clause 4.15.6.3 of TS 23.502 [3] and other parameters. Table 6.7.5.1-2 shows the mapping between each Exception ID and UE behaviour parameters. Table 6.7.5.1-2: Description of Expected UE Behaviour parameters per Exception ID Exception ID UE behaviour parameters to provide Unexpected UE location Expected UE Moving Trajectory Stationary Indication Unexpected long-live/large rate flows Periodic Time Scheduled Communication Time Communication Duration Time Unexpected wakeup Periodic Time Communication Duration Time Scheduled Communication Time Suspicion of DDoS attack Periodic Time Communication Duration Time Scheduled Communication Time Scheduled Communication Type Traffic Profile Expected transaction Dispersion Too frequent Service Access Periodic Time Unexpected radio link failures Expected UE Moving Trajectory Ping-ponging across neighbouring cells Expected UE Moving Trajectory Stationary Indication When the NWDAF detects those UEs that deviate from the expected UE behaviour, e.g. unexpected UE location, abnormal traffic pattern, unexpected transaction dispersion amount, wrong destination address, etc. the NWDAF shall notify the result of the analytics to the consumer as specified in clause 6.7.5.3.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.5.2 Input Data	The Exceptions information from AF is as specified in Table 6.7.5.2-1. On request of the service consumer, the NWDAF shall collect and analyse UE behavioural information from the 5GC NFs (SMF, AMF, AF), or OAM as specified in clauses 6.7.2.2 and 6.7.3.2 and/or expected UE behavioural parameters from UDM as defined in clause 4.15.6.3, TS 23.502 [3], depending on Exception IDs. NOTE: Care needs to be taken with regards to load by avoiding to cause major extra signalling when collecting data for any UE. Table 6.7.5.2-1: Exceptions information from AF Information Description IP address 5-tuple To identify a data flow of a UE via the AF (such as the Firewall or a Threat Intelligence Sharing platform) Exceptions (1..max) (NOTE 1) >Exception ID Indicating the Exception ID (such as Unexpected long-live/large rate flows and Suspicion of DDoS attack as defined in Table 6.7.5.3-2) of the data flow. >Exception Level Scalar value indicating the severity of the abnormal behaviour. >Exception trend Measured trend (up/down/unknown/stable) NOTE 1: The Exceptions information and the UE behavioural information as defined in clauses 6.7.2.2 and 6.7.3.2 could help NWDAF to train an Abnormal classifier, which could be used to classify a UE behaviour data into Normal behaviour or Exception.
e596a2ada7ffa79b8d63b0ea972e234a	23.288	6.7.5.3 Output Analytics	Corresponding to the "abnormal behaviour" Analytics ID, the analytics result provided by the NWDAF is defined in Table 6.7.5.3-1 and Table 6.7.5.3-2. When the level of an exception trespasses above or below the threshold, the NWDAF shall notify the consumer with the exception ID associated with the exception if the exception ID is within the list of exception IDs indicated by the consumer or matches the expected analytics type indicated by the consumer. The NWDAF shall provide the Exception Level and determine which of the other information elements to provide, depending on the observed exception. Abnormal behaviour statistics information is defined in Table 6.7.5.3-1. Table 6.7.5.3-1: Abnormal behaviour statistics Information Description Exceptions (1..max) List of observed exceptions > Exception ID The risk detected by NWDAF > Exception Level Scalar value indicating the severity of the abnormal behaviour > Exception trend Measured trend (up/down/unknown/stable) > UE characteristics Internal Group Identifier, TAC > SUPI list (1..SUPImax) SUPI(s) of the UE(s) affected with the Exception > Ratio Estimated percentage of UEs affected by the Exception within the Target of Analytics Reporting > Amount Estimated number of UEs affected by the Exception (applicable when the Target of Analytics Reporting = "any UE") > Additional measurement Specific information for each risk (see Table 6.7.5.3-3) Abnormal behaviour predictions information is defined in Table 6.7.5.3-2. Table 6.7.5.3-2: Abnormal behaviour predictions Information Description Exceptions (1..max) List of predicted exceptions > Exception ID The risk detected by NWDAF > Exception Level Scalar value indicating the severity of the abnormal behaviour > Exception trend Measured trend (up/down/unknown/stable) > UE characteristics Internal Group Identifier, TAC > SUPI list (1..SUPImax) SUPI(s) of the UE(s) affected with the Exception > Ratio Estimated percentage of UEs affected by the Exception within the Target of Analytics Reporting > Amount Estimated number of UEs affected by the Exception (applicable when the Target of Analytics Reporting = "any UE") > Additional measurement Specific information for each risk (see Table 6.7.5.3-3) > Confidence Confidence of this prediction The predictions are provided with a Validity Period, as defined in clause 6.1.3. The UE characteristics may provide a set of features common to all UEs affected with the exception. The number of exceptions and the length of the SUPI list shall respectively be lower than the parameters maximum number of objects and Maximum number of SUPIs provided as part of Analytics Reporting Information. If PCF subscribes to notifications on "Abnormal behaviour", the NWDAF shall send the PCF notifications about the risk, which may trigger the PCF to update the AM/SM policies. The NWDAF also sends the notification directly to the AMF or SMF, if the AMF or SMF subscribes to the notification, so that the AMF or SMF may, based on operator local policies defined on a per S-NSSAI basis (for AMF) or on a per S-NSSAI, per DNN, or per (DNN,S-NSSAI) basis (for SMF), take actions for risk solving. If the AF subscribes to notifications on "Abnormal behaviour", the NWDAF sends the notifications to the AF so that the AF may take actions for risk solving. The following Table 6.7.5.3-3 gives examples of additional measurement provided by the NWDAF and examples of NF actions for solving each risk. Table 6.7.5.3-3: Examples of additional measurements and NF actions for risk solving Exception ID and description Additional measurement Actions of NFs Unexpected UE location Unexpected UE location (TA or cells which the UE stays) PCF may extend the Service Area Restrictions with current UE location. AMF may extend the mobility restriction with current UE location. Ping-ponging across neighbouring cells Numbers, frequency, time and location information, assumption about the possible circumstances of the ping-ponging If the ping-ponging are per UE, then: 1. the AMF may adjust the UE (e.g. a stationary UE) registration area. 2. the AMF and/or the AF may allow the use of Coverage Enhancement for the affected UE. Unexpected long-live/large rate flows Unexpected flow template (IP address 5 tuple) SMF updates the QoS rule, e.g. decrease the MBR for the related QoS flow. PCF, if dynamic PCC applies for corresponding DNN, S-NSSAI, updates PCC Rules that triggers SMF updates the QoS rule, e.g. decrease the MBR for the related QoS flow. Unexpected wakeup Time of unexpected wake-up AMF applies MM back-off timer to the UE. Suspicion of DDoS attack Victim's address (target IP address list) PCF may request SMF to release the PDU session. SMF may release the PDU session and apply SM back-off timer. Wrong destination address Wrong destination address (target IP address list) PCF updates the packet filter in the PCC Rules that triggers the SMF to update the related QoS flow and configures the UPF. Too frequent Service Access Volume, frequency, time, assumptions about the possible circumstances AF may release the AF session. PCF may request SMF to release the PDU session. SMF may release the PDU session and apply SM back-off timer. Unexpected radio link failures Numbers, frequency, time and location, assumptions about the possible circumstances If the unexpected radio link failures are per UE location bases, the AMF may allow the use of CE (Coverage Enhancement) in the affected location. Also, the Operator may improve the coverage conditions in the affected location. If the unexpected radio link failures are per UE bases, then the AMF and/or the AF may allow the use of CE for the affected UE.

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.