Figure 9.13.2-1 illustrate the procedure for supporting adaptive data transmission. The SEALDD facilitates to transmit data between the VAL client and VAL server adaptively by using application enablement layer capabilities of the AIML Enablement server defined in TS 23.482. This procedure can be used to support adaptive data transmission for XR services.
Pre-conditions:
The AIMLE Server is known by the consumer and the SEALDD Server.
The AI/ML services provided by the AIMLE Server is allowed for exposure to the consumer and the SEALDD Server.
The consumer decides on the application which requires split image/video processing (distribute image/video processing tasks to multiple split endpoints for processing).
Up to requirements of high layer application, SEALDD Server triggers interaction with AIMLE Server for assistance information.
The consumer (e.g. VAL server, VAL client via SEALDD client) sends adaptive data transmission request to the SEALDD server (e.g. for supporting XR service to deliver/distribute data to the split AI/ML endpoints (i.e. VAL client(s))). The request message contains message as defined in Table 9.13.3.1-1.
Upon receiving the request, the SEALDD server performs an authorization check. If authorization is successful, the SEALDD server responses to the VAL server. The SEALDD server derives information for the request, for example, based on the message in the request to decide whether assistance from 5GC and/or AIMLE server is needed or not, and determine downstream entities and services needed based on the information derived.
Then the SEALDD server sends request to the AIMLE server for split operation assistance information for the data transmission. For example, for split AI/ML image/video processing, the request contains the information of split AI/ML endpoints (from the request in step 1), data information (e.g. data type, data size, range of data size) to be delivered/distributed to the split AI/ML endpoints, may contain QoS requirements on the data transmission, e.g. desired data transmission time. The information flow for the request reuses the mechanisms specified in TS 23.482.
The SEALDD server may request and receive assistance from 5GC, as described in step 2, e.g. assistance information from NWDAF (or via NEF) about analytics on E2E data volume transfer time. In the request from SEALDD server to NWDAF for E2E data volume transfer time analytics, the split AI/ML endpoints information and a pre-configured data size (or data size range) may be used (the pre-configured data size may be the data size provided in step 1 or be decided by the SEALDD server according to the data size range in step 1 and based on its local policy. The data size range is provided in step 1.). The analytics subscription can be modified when the actual data size changed. The actual data size may be obtained via measurement described in Sdd_TransmissionQuality Management API.
The SEALDD server triggers to establish connection among the consumer, SEALDD server, SEALDD client(s), and with the split AI/ML endpoints (i.e. VAL client(s)) for the data transmission. The split operation assistance information received in steps 2 and 3 can be used by the SEALDD server separately or jointly based on need and local policy.
According to the assistance information, the SEALDD server may adjust the data volume to be transmission to each of the split AI/ML endpoints (if adaptive assign of date volume to the split AI/ML endpoints is allowed), or adjust the time point/time window for data transmission to the split endpoints.
Operations are performed at the consumer, SEALDD server, SEALDD client(s), and VAL client(s) for the data transmission, e.g. image/video/multi-modal data transmission from the consumer to the split AI/ML endpoints.
Table 9.13.3.1-1 describes the information flow from the consumer to the SEALDD server for requesting or updating request for an adaptive data transmission.