hash
stringlengths
32
32
doc_id
stringlengths
5
12
section
stringlengths
5
1.47k
content
stringlengths
0
6.67M
a968fed150b6156e01512b97375c462a
23.700-82
8.1.4.5 AIML-UU
The interactions related to AIML enablement layer functions between the AIML enablement client and AIML enablement server are supported by AIML-UU reference point. This reference point utilizes Uu reference point as described in 3GPP TS 23.401 [8] and 3GPP TS 23.501 [5].
a968fed150b6156e01512b97375c462a
23.700-82
8.1.4.6 AIML-S
The interactions related to AIML enablement layer functions between the VAL server(s) and the AIML enablement server are supported by AIML-S reference point.
a968fed150b6156e01512b97375c462a
23.700-82
8.1.4.7 AIML-C
The interactions related to AIML enablement layer functions between the VAL client(s) and the AIML enablement client within a VAL UE are supported by AIML-C reference point.
a968fed150b6156e01512b97375c462a
23.700-82
8.1.4.8 AIML-E
The interactions related to AIML enablement functions between different AIML enablement servers are supported by AIML-E reference point (e.g. central and edge AIMLE servers).
a968fed150b6156e01512b97375c462a
23.700-82
8.1.5 Architecture Impacts
The application enabler layer architecture impacts are the following: - Enhance ADAE to consume AIML enablement services for ML-based analytics. - On-network AIML enablement layer functional model is introduced to support application layer AI/ML services.
a968fed150b6156e01512b97375c462a
23.700-82
8.1.6 Corresponding APIs
This solution impacts AIMLE and ADAE APIs. ADAE: - ADAE API is enhanced, including enhancement of ADAE-UU, ADAE-S, and ADAE-C reference points. AIMLE: - New AIMLE API is introduced, including AIML-UU, AIML-S, and AIML-C reference points.
a968fed150b6156e01512b97375c462a
23.700-82
8.1.7 Solution evaluation
This solution addresses Key Issue #1 and introduces AIML enablement layer functional model, the AIML enablement service for supporting application layer AI / ML services and enhance ADAE to consume AIML enablement services for ML-based analytics. The content in Solution#1 can be selected and add to clause 7, for exampl...
a968fed150b6156e01512b97375c462a
23.700-82
8.2 Solution #2: ML client information retrieval
a968fed150b6156e01512b97375c462a
23.700-82
8.2.1 General
The following clauses specify procedures, information flows and APIs for Key issue X to enable the selection of ML clients. Assumptions: 1. The proposed solution is based on the AIML Enabler service producer/consumer architecture. The service producer can be a standalone new entity like an AIML Enabler server or colloc...
a968fed150b6156e01512b97375c462a
23.700-82
8.2.2 Procedures
Figure 8.2.2-1: ML client information retrieval procedure 1. The AIML Enabler service consumer (e.g., VAL server) sends the ML client information retrieval request to the AIML enabler service producer. The request may contain identities to identify the ML service, some filtering criteria to filter for the selection of ...
a968fed150b6156e01512b97375c462a
23.700-82
8.2.3 Architecture Impacts
The current solution is based on the service producer and consumer model, where the service producer and consumer could be AIML enabler server and client. The architecture based on AIML enabler client/server is described in clause 7. The AIML enabler service producer is AIML Enablement server, and AIML enabler service ...
a968fed150b6156e01512b97375c462a
23.700-82
8.2.4 Corresponding APIs
This subclause provides a summary on the corresponding API for solution #2. - ML client information retrieval API (request / response model; API provider: AIMLE server; known consumer: VAL server; corresponding to step 1and step 3 of clause 8.2.2.1).
a968fed150b6156e01512b97375c462a
23.700-82
8.2.5 Solution evaluation
This solution addresses Key Issue #3(bullet #2 - Identify procedures for supporting FL at the application enablement layer, including FL entity discovery, registration, communication, reporting) and Key Issue #7(bullet #1 - How to support the AI/ML for member selection and re-selection (e.g., policies).). This enables ...
a968fed150b6156e01512b97375c462a
23.700-82
8.3 Solution #3: Provision of ML clients to support AI/ML at the application layer
a968fed150b6156e01512b97375c462a
23.700-82
8.3.1 General
The following clauses specify procedures, information flows and APIs for Key issue #3, #7 to provision the ML clients for the AIML Enabler services. Assumptions: 1. The proposed solution is based on the AIML Enabler service producer/consumer architecture. The service producer can be a standalone new entity like an AIML...
a968fed150b6156e01512b97375c462a
23.700-82
8.3.2 Procedures
Figure 8.3.2-1: ML client config provisioning procedure 1. The AIML Enabler service consumer (e.g. VAL server or VAL UE) sends a ML client config provisioning request to the AIML Enabler service producer (e.g. ADAE server). The request may contain UE ML client profile e.g., ML capability, UE QoS. NOTE 1: The UE ML clie...
a968fed150b6156e01512b97375c462a
23.700-82
8.3.3 Architecture Impacts
The current solution is based on the service producer and consumer model, where the service producer and consumer could be AIML enabler server and client. The architecture based on AIML enabler client / server is described in clause 7. The solution does not introduce any architecture impacts. The AIML enabler service p...
a968fed150b6156e01512b97375c462a
23.700-82
8.3.4 Corresponding APIs
This subclause provides a summary on the corresponding API for solution #3. - ML client provisioning request API (request/response model; API provider: AIMLE server; known consumer: AIML consumer (like VAL server, VAL UE); corresponding to step 1and step 2 of clause 8.3.2).
a968fed150b6156e01512b97375c462a
23.700-82
8.3.5 Solution evaluation
This solution (hereafter referred to as evaluated solution) addresses Key Issue #3(bullet 2 - Identify procedures for supporting FL at the application enablement layer, including FL entity discovery, registration, communication, reporting) and KI #7(bullet 2- How to support the AI/ML member participation configurations...
a968fed150b6156e01512b97375c462a
23.700-82
8.4 Solution #4: Support for ML-enabled ADAE analytics
a968fed150b6156e01512b97375c462a
23.700-82
8.4.1 Solution description
This solution addresses Key Issue #2. This solution introduces the selection and configuration of the ML model entities to support a certain ADAE layer analytics event. This solution aims the enhancement of analytics services provided by the analytics enablement functionality, and in particular analytics related to VAL...
a968fed150b6156e01512b97375c462a
23.700-82
8.4.2 Architecture Impacts
The application enabler layer architecture impacts are the following: - AIMLE server is introduced to provide support the training and/or inference for the analytics event. - ADAES is enhanced to select and configure the entity (-ies) to serve as ML model training and inference entities and utilizes these services to i...
a968fed150b6156e01512b97375c462a
23.700-82
8.4.3 Corresponding APIs / information
This subclause provides a summary on the corresponding API for solution #4. - ML model training API (request / response model; API provider: AIMLE Server; known consumer: ADAES; corresponding to step 4 and 8). - ML model inference API (request / response model; API provider: AIMLE Server; known consumer: ADAES; corresp...
a968fed150b6156e01512b97375c462a
23.700-82
8.4.4 Solution evaluation
This solution addresses Key Issue #2 and introduces the capability to support ML-enabled ADAE analytics, by performing ML model training/inference on behalf of ADAES. This solution is feasible and doesn't introduce any dependency to 3GPP network systems. NOTE: Whether ML repository will be unique for both ML model info...
a968fed150b6156e01512b97375c462a
23.700-82
8.5 Solution #5: AI/ML model management
a968fed150b6156e01512b97375c462a
23.700-82
8.5.1 General
This solution addresses Key Issue #2. This solution introduces the AI/ML model information storage and discovery procedures to support AI/ML model management. In this solution, the following terminology is used: - Analytics ID is defined in 3GPP TS 23.436 [4].
a968fed150b6156e01512b97375c462a
23.700-82
8.5.2 AI/ML model information storage procedure
Figure 8.5.1-1 illustrates the procedure of AI/ML model information storage. Figure 8.5.2-1: AI/ML model information storage procedure 1. The model repository consumer (e.g. MTME-enhanced ADAES or AIML enablement server or VAL server or ADAE client) sends an AI/ML model information storage request to the model reposito...
a968fed150b6156e01512b97375c462a
23.700-82
8.5.3 AI/ML model information discovery procedure
Figure 8.5.3-1 illustrates the procedure of AI/ML model information discovery. Figure 8.5.3-1: AI/ML model information discovery procedure 1. The model repository consumer (e.g., VAL server) sends an AI/ML model information discovery request to the model repository. The request contains information elements as describe...
a968fed150b6156e01512b97375c462a
23.700-82
8.5.4 Architecture Impacts
The application enabler layer architecture impacts are the following: - An AI/ML model repository is introduced for enabling the AI/ML model information storage and discovery. Such AI/ML model repository may be part of the ML Repository as described in clause 7.2.3.
a968fed150b6156e01512b97375c462a
23.700-82
8.5.5 Corresponding APIs
a968fed150b6156e01512b97375c462a
23.700-82
8.5.5.1 AI/ML model information storage request
Table 8.5.5.1-1 describes the information flow from the repository consumer to the model repository as a request for the AI/ML model storage. Table 8.5.5.1-1: AI/ML model information storage request Information element Status Description Requester Identity M (NOTE 1) The identity of the model repository consumer perfor...
a968fed150b6156e01512b97375c462a
23.700-82
8.5.5.2 AI/ML model information storage response
Table 8.5.5.2-1 describes the information flow from the model repository to the repository consumer as a response for the AI/ML model storage request. Table 8.5.5.2-1: AI/ML model information storage response Information element Status Description Result M Indicates success or failure of the request > ML model ID O Rep...
a968fed150b6156e01512b97375c462a
23.700-82
8.5.5.3 AI/ML model information discovery request
Table 8.5.5.3-1 describes the information flow from the repository consumer to the model repository as a request for the AI/ML model discovery. Table 8.5.5.3-1: AI/ML model information discovery request Information element Status Description Requester Identity M The identity of the model repository consumer performing ...
a968fed150b6156e01512b97375c462a
23.700-82
8.5.5.4 AI/ML model information discovery response
Table 8.5.5.4-1 describes the information flow from the model repository to the repository consumer as a response for the AI/ML model discovery request. Table 8.5.5.4-1: AI/ML model information discovery response Information element Status Description Result M Indicates success or failure of the request > ML model(s) O...
a968fed150b6156e01512b97375c462a
23.700-82
8.5.6 Solution evaluation
This solution addresses Key Issue #2 and introduces the capability to support the management of AI/ML models, by performing AI/ML model information storage and discovery procedures. The AI/ML model management functionality further enables or supports other functionalities related to AI/ML models (as described in other ...
a968fed150b6156e01512b97375c462a
23.700-82
8.6.1 Solution description
The following clauses specify procedures, information flows, and APIs for Key Issue #3 to support AIML enablement client selection. This solution uses the following definitions: - AIML client set identifier is an identifier of the selected AI/ML members. The VAL group ID may be used as AIML client set identifier and AI...
a968fed150b6156e01512b97375c462a
23.700-82
8.6.1.1 AIML Enablement Client Selection with SEAL and 5GC support
Assumptions: 1. The proposed solution is based on client-server architecture for federated learning. 2. The VAL server has already discovered and received a list of AIML enablement clients that are suitable and have available data for a particular AIML operation. 3. The discovery operation may be performed as described...
a968fed150b6156e01512b97375c462a
23.700-82
8.6.1.2 Dynamic AIML Enablement Client selection and Monitoring Subscription
Assumptions: 1. The VAL server may have performed the AIML Enablement Client Selection with SEAL and 5GC support procedure as defined in clause 8.6.1.1 and has the AIML client set identifier. 2. The VAL server may configure the AI/ML network sessions between VAL server and AI/ML Clients via SEALDD (Sdd_RegularTransmiss...
a968fed150b6156e01512b97375c462a
23.700-82
8.6.2 Architecture Impacts
This solution is based on client-server architecture as described by clause 7.2.
a968fed150b6156e01512b97375c462a
23.700-82
8.6.3 Corresponding APIs
Table 8.6.3-1 shows the request sent by a VAL server to an AIML enablement server for the AIML enablement client selection with SEAL and 5GC support procedure. Table 8.6.3-1: Request for AIML enablement client selection procedure Information element Status Description Requestor identity M The identifier of the requesto...
a968fed150b6156e01512b97375c462a
23.700-82
8.6.4 Solution evaluation
Solution #6 addresses Key Issue #3 and provides a selection mechanism for AIML enablement servers to select a group of AIML enablement clients for AIML operations. AIML enablement client discovery and selection may be performed in a static or dynamic manner in support of federated, split, and transfer learning workflow...
a968fed150b6156e01512b97375c462a
23.700-82
8.7.1 Solution description
The following clauses specify procedures, information flows, and APIs for Key Issue #3 to support AIML enablement client discovery. Assumptions: 1. The proposed solution is based on client-server architecture for federated learning. 2. The AIML enablement server has a repository of AIML enablement clients and associate...
a968fed150b6156e01512b97375c462a
23.700-82
8.7.2 Architecture Impacts
Following enhancement are required for the architecture defined in clause 8.1. Figure 8.7.2-1 illustrates the functional model for interconnection between AIML Enablement servers. Figure 8.7.2-1: Interconnection between AIML Enablement servers The AIML Enablement server interacts with another AIML Enablement server for...
a968fed150b6156e01512b97375c462a
23.700-82
8.7.3 Corresponding APIs
Table 8.7.3-1 shows the request sent by a VAL server to an AIML enablement server for the AIML enablement client discovery procedure. Table 8.7.3-1: Request for AIML enablement client discovery procedure Information element Status Description VAL server identifier M The identifier of the VAL server. Security credential...
a968fed150b6156e01512b97375c462a
23.700-82
8.7.4 Solution evaluation
This solution addresses Key Issues #3 and #7 and provides a discovery mechanism for VAL applications to find AIML enablement clients capable of and with appropriate datasets for a particular AIML operation. This solution is feasible and does not introduce any dependency to the 3GPP network system. 8.8 Solution #8: AIML...
a968fed150b6156e01512b97375c462a
23.700-82
8.8.1 Solution description
The following clauses specify procedures, information flows, and APIs for Key Issue #3 to support AIML enablement client registration. Assumptions: 1. The proposed solution is based on client-server architecture for federated learning. 2. The AIML enablement client has been pre-configured or has discovered the address ...
a968fed150b6156e01512b97375c462a
23.700-82
8.8.2 Architecture Impacts
This solution is based on client-server architecture as described by clause 7.2.
a968fed150b6156e01512b97375c462a
23.700-82
8.8.3 Corresponding APIs
Table 8.8.3-1 shows the request sent by an AIML enablement client to an AIML enablement server for the AIML enablement client registration procedure. Table 8.8.3-1: Request for AIML enablement client registration procedure Information element Status Description AIML enablement client identifier M The identifier of the ...
a968fed150b6156e01512b97375c462a
23.700-82
8.8.4 Solution evaluation
This solution addresses Key Issues #3 and #7 and provides a registration mechanism for AIML enablement clients to provide their capabilities for AIML operations. AIML enablement client registration is an important part of federated, split, and transfer learning workflows. This solution is feasible and does not introduc...
a968fed150b6156e01512b97375c462a
23.700-82
8.9 Solution #9: Support for FL member registration
a968fed150b6156e01512b97375c462a
23.700-82
8.9.1 Solution description
This solution addresses Key Issue #3. This solution provides the procedures for the registration and registration update of candidate FL members in an FL Member Registry (e.g., ML repository). Candidate FL members can be application layer entities (e.g., VAL server, EAS) at the server or VAL UE side, which can potentia...
a968fed150b6156e01512b97375c462a
23.700-82
8.9.1.1 Procedure for FL member registration
Figure 8.9.1.1-1 illustrates the procedure where the registration of a candidate FL member happens via the FL member Registry. This procedure covers the registration of the candidate FL member to a global repository / registry which is introduced for keeping the FL member registrations. Such candidate member can be a s...
a968fed150b6156e01512b97375c462a
23.700-82
8.9.1.2 Procedure for FL member registration update
Figure 8.9.1.2-1 illustrates the procedure where the registration update of a candidate FL member happens via the FL member Registry. Figure 8.9.1.2-1: Procedure for registration update 1. The candidate FL member (e.g. VAL server, VAL UE) either decides to offboard or optionally change the capabilities or a trigger is ...
a968fed150b6156e01512b97375c462a
23.700-82
8.9.2 Architecture Impacts
The application enabler layer architecture impacts are the following: - An FL member Registry is introduced for enabling the registration of candidate FL members. Such FL member registry may be part of the ML member repository as described in clause 7.2.3. NOTE: Whether FL member registry will be part of the ML reposit...
a968fed150b6156e01512b97375c462a
23.700-82
8.9.3 Corresponding APIs
This subclause provides a summary on the corresponding API for solution #9. - FL member Register API (request / response model; API provider: FL member registry; known consumer: AIML enablement server; corresponding to step 2 and 4 of clause 8.9.1.1). - FL member Register Update API (request / response model; API provi...
a968fed150b6156e01512b97375c462a
23.700-82
8.9.4 Solution evaluation
This solution addresses Key Issue #3 and introduces the capability to support registration of an FL member (or a group of FL members) and registration update to a FL member registry (based on changes). Such FL members can be VAL UEs or VAL servers which are candidate to be used in ML operations. This solution is import...
a968fed150b6156e01512b97375c462a
23.700-82
8.10 Solution #10: AI/ML member participation configurations provisioning and management
a968fed150b6156e01512b97375c462a
23.700-82
8.10.1 General
The following clauses specify procedures, information flows and APIs for Key issue #7 to provision the AI/ML member participation configurations for the AI/ML Enabler services. Assumptions: - AI/ML member (e.g., AI/ML Enabler Client) is registered on the AI/ML Enabler Server to participate in the AI/ML operations (e.g....
a968fed150b6156e01512b97375c462a
23.700-82
8.10.2 Procedures
a968fed150b6156e01512b97375c462a
23.700-82
8.10.2.1 AI/ML member participation configurations provisioning and management
Figure 8.10.2.1-1: AI/ML member participation configurations provisioning and management procedure 1. The AI/ML member (e.g. AI/ML Client) sends an AI/ML member participation configurations provisioning and management request as defined in clause 8.10.3.1. 2. Upon receiving the request, the AIML Enablementserver perfor...
a968fed150b6156e01512b97375c462a
23.700-82
8.10.3 Information flows
a968fed150b6156e01512b97375c462a
23.700-82
8.10.3.1 AI/ML member participation configurations provisioning and management request
Table 8.10.3.1-1 describes the information flow from the AI/ML member to the AI/ML Enabler server as a request for the AI/ML member participation configurations provisioning and management. Table 8.10.3.1-1: AI/ML member participation configurations provisioning and management request Information element Status Descrip...
a968fed150b6156e01512b97375c462a
23.700-82
8.10.3.2 AI/ML member participation configurations provisioning and management response
Table 8.10.3.2-1 describes the information flow from the AI/ML Enabler server to the AI/ML member as a response for the AI/ML member participation configurations provisioning and management. Table 8.10.3.2-1: AI/ML member participation configurations provisioning and management response Information element Status Descr...
a968fed150b6156e01512b97375c462a
23.700-82
8.10.4 Architecture Impacts
This solution is based on the architecture in clause 7 with added functionality for the AIML Enablement Client Configuration Provisioning and Management.
a968fed150b6156e01512b97375c462a
23.700-82
8.10.5 Corresponding APIs
This clause provides a summary on the corresponding API for solution #10. - AIML Enablement Client Configuration Provisioning and Management service operations are a part of AIMLE Client Registration API (request/response model; API provider: AIMLE server; known consumer: AIMLE Client).
a968fed150b6156e01512b97375c462a
23.700-82
8.10.6 Solution evaluation
This solution addresses Key Issue #7 and introduces the capability for the AIMLE Client to provision the participation configurations that include time and spatial conditions, supported ML modes and roles, etc. This solution is ML mode agnostic (i.e., applicable for registration and configuration provisioning for FL, D...
a968fed150b6156e01512b97375c462a
23.700-82
8.11 Solution #11: AIML service lifecycle management procedure
a968fed150b6156e01512b97375c462a
23.700-82
8.11.1 General
The following clauses specify procedures, information flows and APIs for Key issue #1 to enable the AIML service lifecycle management procedure. The AIML service opeation is a interaction between VAL server (e.g., AIML server) and AIML client for the AIML service. The VAL server uses the AIML enabler layer to support t...
a968fed150b6156e01512b97375c462a
23.700-82
8.11.2 AIML service lifecycle management procedure
Figure 8.11.2-1: AIML service lifecycle management procedure 1. The VAL server sends AIML service lifecycle management request to the AIML Enablement server. The request contains AIML service operation ID, AIML client ID, AIML clients group ID, AIML service operation mode like start, pause, continue, and finish. AIML s...
a968fed150b6156e01512b97375c462a
23.700-82
8.11.3 Architecture Impacts
The solution is based on the functional architecture described in clause 7 and it has no architecture impacts.
a968fed150b6156e01512b97375c462a
23.700-82
8.11.4 Corresponding APIs
This subclause provides a summary on the corresponding API for solution #11. - AIML service lifecycle management request API (request/response model; API provider: AIMLE server; known consumer: AIML consumer (like VAL server); corresponding to step 1 and step 4 of clause 8.11.2). - AIML Enablement client service operat...
a968fed150b6156e01512b97375c462a
23.700-82
8.11.5 Solution evaluation
This solution addresses Key Issue #1(bullet #2 - Whether and how the above architecture enhancement and related functions supporting the management/execution of AI/ML lifecycle operations, bullet #3 - Whether and how the architecture enhancement and related functions leveraging the existing 5GC capabilities and assista...
a968fed150b6156e01512b97375c462a
23.700-82
8.12 Solution #12: AI/ML model lifecycle management
a968fed150b6156e01512b97375c462a
23.700-82
8.12.1 Solution Description on AI/ML Model Lifecycle Management
This solution addresses Key Issues #1 and #6. The solution provides a procedure to support AI/ML model lifecycle management for ML model re-training and update when model performance degradation is observed by AI/ML Enablement. The solution also supports using an existing model to re-train the model using Transfer Lear...
a968fed150b6156e01512b97375c462a
23.700-82
8.12.2 Consumer-based ML Model Performance Degradation Detection
Figure 8.12.2-1: Consumer-based ML Model Performance Degradation Detection This procedure corresponds to the step 0b in clause 8.12.1. 0. An ADAE Server, as an AI/ML Enablement Consumer, receives trained ML model (or ML model information) from the AI/ML Enablement Server. 1. The ADAE Server receives request from consum...
a968fed150b6156e01512b97375c462a
23.700-82
8.12.2 Architecture Impacts
The application enabler layer architecture impacts are the following: - AI/ML Enablement Server is introduced to support AI/ML model lifecycle management. - AI/ML Enablement Consumer is introduced to consume AI/ML Enablement Server services. - AI/ML Enablement Consumer is introduced to detect ML model performance degra...
a968fed150b6156e01512b97375c462a
23.700-82
8.12.3 Corresponding APIs
Table 8.12.3-1 details the ML model performance degradation notification IEs. Table 8.12.3-1: AI/ML model update request Information element Status Description AI/ML Enablement Consumer Identity M The identity of the AI/ML Enablement Consumer sending the notification. AI/ML model ID M Provides the ID of AI/ML model for...
a968fed150b6156e01512b97375c462a
23.700-82
8.12.4 Solution evaluation
This solution addresses Key Issues #1 and #6, and details the procedures for AI/ML model lifecycle management to detect performance degradation and trigger AI/ML model update. The solution also supports scenarios of Transfer Learning, by triggering proactive updates to related models if needed. This solution is feasibl...
a968fed150b6156e01512b97375c462a
23.700-82
8.13 Solution #13: Analytics and Assistance Information Collection for Supporting FL Member (Re)Selection with the ADAES capabilities
a968fed150b6156e01512b97375c462a
23.700-82
8.13.1 General
The following clauses specify procedures, information flows and APIs for Key Issues #1 and #3 to support AI/ML Enablement Server interact with ADAES and NEF for analytics and assistance information to support FL member (re)selection in the VAL server driven and the AI/ML Enablement server driven FL members (re)selectio...
a968fed150b6156e01512b97375c462a
23.700-82
8.13.2 Procedure on Analytics/Assistance Information Collection for FL member (re)selection with the ADAES capabilities
Figure 8.13.2-1: Analytics/Assistance information collection procedure for FL member (re)selection 0. The consumer (e.g. VAL server) may request AI/ML Enablement Server to start assisting the VAL server for the FL member (re)selection or request the AI/ML Enablement Server to coordinate the AI/ML operations for member ...
a968fed150b6156e01512b97375c462a
23.700-82
8.13.3 Architecture Impacts
This solution is based on the architecture described in clause 7. The application enabler layer architecture impacts are the following: - Interactions between AI/ML Enablement Server and ADAES are introduced to support AI/ML Enablement Server requests/subscribes analytics from ADAES. - Interactions between AI/ML Enable...
a968fed150b6156e01512b97375c462a
23.700-82
8.13.4 Corresponding APIs
None.
a968fed150b6156e01512b97375c462a
23.700-82
8.13.5 Solution evaluation
This solution addresses Key Issues #1 (open issue 3), #2 (open issues 2 and 4), and #3 (open issue 1), and introduces the interactions between AI/ML Enablement Server with ADAES for analytics and with NEF for assistance information. This solution provides architecture enhancement for leveraging the existing ADAE servic...
a968fed150b6156e01512b97375c462a
23.700-82
8.14 Solution #14: AI/ML policies provisioning and management
a968fed150b6156e01512b97375c462a
23.700-82
8.14.1 General
The following clauses specify procedures, information flows and APIs for Key issue#7 to provision and manage the AI/ML policies for the AI/ML Enablement services. The AI/ML data transfer policies allow the VAL server to provide requirements for the AI/ML data (e.g., the current status of the trained ML model) from the ...
a968fed150b6156e01512b97375c462a
23.700-82
8.14.2 Procedures
a968fed150b6156e01512b97375c462a
23.700-82
8.14.2.1 AI/ML policies provisioning and management
Figure 8.14.2-1: AI/ML policies provisioning and management procedure 1. The VAL server sends an AI/ML service request with policies provisioning and management information as defined in clause 8.A.3.1. 2. Upon receiving the request, the AI/ML Enablement Server performs an authorization check of the VAL server. 3. If t...
a968fed150b6156e01512b97375c462a
23.700-82
8.14.3 Information flows
a968fed150b6156e01512b97375c462a
23.700-82
8.14.3.1 AI/ML service request with policies provisioning and management information
Table 8.14.3.1-1 describes the information flow from the VAL server to the AI/ML Enablement Server as a request that contains the AI/ML policies provisioning and management information. Table 8.14.3.1-1: AI/ML policies provisioning and management request Information element Status Description Requester Identity M The i...
a968fed150b6156e01512b97375c462a
23.700-82
8.14.3.2 AI/ML service response with policies provisioning and management information
Table 8.14.3.2-1 describes the information flow from the AI/ML Enablement Server to the VAL server as a response that contains the AI/ML policies provisioning and management result. Table 8.14.3.2-1: AI/ML member selection policies provisioning and management response Information element Status Description Result M Ind...
a968fed150b6156e01512b97375c462a
23.700-82
8.14.4 Architecture Impacts
This solution is based on the architecture in clause 7 with added functionality for the AIML Policy Provisioning and Management.
a968fed150b6156e01512b97375c462a
23.700-82
8.14.5 Corresponding APIs
This clause provides a summary on the corresponding API for solution #14. - AIML Policies Provisioning and Management service operations can be a part of AIMLE Client Discovery and Selection API (request/response model; API provider: AIMLE server; known consumer: VAL server and/or AIMLE Client).
a968fed150b6156e01512b97375c462a
23.700-82
8.14.6 Solution evaluation
This solution addresses Key Issue #7 and introduces the capability for the VAL server and/or AIMLE Client to provision and manage the AI/ML policies for AIMLE Client selection and re-selection, data transfer. This solution is ML mode agnostic (i.e., applicable for FL, DML, ML, Transfer and/or Split Learnings) and impor...
a968fed150b6156e01512b97375c462a
23.700-82
8.15 Solution #15: ADAES support for AI-enabled DN Energy Analytics
a968fed150b6156e01512b97375c462a
23.700-82
8.15.1 Solution description
In the edge scenarios, the energy consumption for EDN can be due to the EES/EAS vCPU usage, the API invocations (for edges services produced or consumed by the EDGE platform) and other energy consumptions (e.g HW/NFVI layer). Some of this part can be fixed; however, lots of the processing is analogous to the applicatio...
a968fed150b6156e01512b97375c462a
23.700-82
8.15.2 Architecture Impacts
The architecture impacts are mainly the enhancement of ADAES to support an additional analytics service for DN energy analytics.
a968fed150b6156e01512b97375c462a
23.700-82
8.15.3 Corresponding APIs
This subclause provides a summary on the corresponding API for solution #15. - DN Energy Analytics API (request-response or subscribe-notify model; API provider: ADAES; known consumers: VAL server, EES; corresponding to step 1 and 7) - Energy-related Data Collection APIs: - Load / Usage data API (request-response or su...
a968fed150b6156e01512b97375c462a
23.700-82
8.15.4 Solution evaluation
This solution addresses Key Issue #2 and introduces a new ML-enabled analytics capability to support ML-enabled DN energy analytics. This solution is re-using Solution #2 for the interaction with AIML enablement server, and mainly enhances ADAES. This solution is feasible and doesn't introduce any dependency to 3GPP ne...
a968fed150b6156e01512b97375c462a
23.700-82
8.16 Solution #16: Support for FL event notifications