Content for  TS 23.288  Word version:  18.3.0

Top   Top   Up   Prev   Next
1…   4…   5…   5A…   6…   6.1.3   6.1.4……   6.1.5…   6.1A…   6.1B…   6.1B.2.3…   6.1C   6.2…   6.2.3…   6.2.6………   6.2.7…   6.2.8…   6.2.9…   6.2.13…   6.2A…   6.2B…   6.2B.3   6.2B.4…   6.2C…   6.2D…   6.2E…   6.2F…   6.3…   6.4…   6.5…   6.6…   6.7…   6.7.3…   6.7.4…   6.7.5…   6.8…   6.9…   6.10…   6.11…   6.12…   6.13…   6.14…   6.16…   6.17…   6.18…   6.19…   6.20…   6.21…   7…   7.4…   7.7…   7.9…   8…   9…   10…


6.1B.2.3  Prepared analytics subscription transferp. 64

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.
Reproduction of 3GPP TS 23.288, Fig. 6.1B.2.3-1: Prepared analytics transfer
Step 0.
Analytics consumer subscribes to the source NWDAF for certain analytics as specified in clause 6.1.1.
Step 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.
Step 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.
Step 1.
The source NWDAF determines that it needs to prepare to transfer analytics to another NWDAF instance.
Step 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. 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.
Step 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.
Step 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.
Step 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).
Step 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 Transferp. 67

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.
Reproduction of 3GPP TS 23.288, Fig. 6.1B.3-1: Analytics Context Transfer
The procedure of analytics context information transfer comprises the following steps:
Step 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.
  • 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 Contextp. 68

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 context 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.
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 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;
      • 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.

Up   Top   ToC