Communication method and device

Abstract

A communication method and apparatus, the method comprising: an AF network element sending a first request, the first request comprising information of a first terminal and / or information of a first session; the AF network element receives a first response, the first response is a response of the first request, the first response comprises information of artificial intelligence (AI) service capability of a first network, the first network is a network accessed by the first terminal, and the first session is a session used for communication between the first terminal and a first server. Furthermore, the AF network element can determine an AI calculation auxiliary strategy according to the AI service capability of the first network, and in the AI calculation auxiliary strategy, an AI calculation function located at the first terminal and / or the first server can be migrated to the first network element of the first network, thereby reducing the AI calculation load of the first terminal and / or the first server.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a communication method and apparatus. Background Technology

[0002] By opening up their capabilities, communication networks can provide network capabilities to third-party applications, thereby enabling business innovation. Operators need to build unified and open network capabilities, providing network capability opening and security functions to third parties through a capability opening platform. On the one hand, this allows for timely response to third-party business needs, ensuring user experience and providing personalized services and security guarantees; on the other hand, it also helps third-party service providers fully utilize the operator's network and business capabilities when opening up their services. Network capability opening has gradually become a hot topic for future growth for operators and a key focus of communication network development. How communication networks can open up artificial intelligence (AI) service capabilities to third-party applications is a research direction. Summary of the Invention

[0003] In a first aspect, a communication method is provided, wherein the execution subject of the method is an AF network element, or a module, unit, or component (e.g., chip, chip system, circuit, or processor) located in the AF network element, comprising: sending a first request, the first request including information of a first terminal and / or information of a first session; receiving a first response, the first response being a response to the first request, the first response including information of the artificial intelligence (AI) service capabilities of a first network, the first network being a network accessed by the first terminal, and the first session being a communication session between the first terminal and a first server.

[0004] Through the above design, the network can open up the AI ​​service capabilities of the first network to third-party application AF network elements, thereby enabling the AF network elements to perform corresponding operations based on the AI ​​service functions of the first network. For example, the AF network elements can migrate the AI ​​computing functions located in the first terminal and / or the first server to the first network elements in the first network, reducing the AI ​​computing load on the first terminal and / or the first server.

[0005] In one possible implementation, the method further includes: determining an AI computing assistance strategy based on information about the AI ​​service capabilities of the first network, wherein the AI ​​computing assistance strategy includes migrating all or part of the AI ​​computing functions located in the first server and / or the first terminal to a first network element of the first network; and sending information about the AI ​​computing assistance strategy.

[0006] In one possible implementation, the method further includes: sending a first indication, the first indication being used to indicate that the service corresponding to the first session is provided with AI computing assistance by the network.

[0007] The second aspect is a method opposite to the first aspect, with beneficial effects as described in the first aspect. It provides a communication method, in which the executing entity is a network element X, or a module, unit, or component located in the network element X. The network element X can be an API gateway, an SMO, or a UPF network element, etc., and includes: receiving a first request, the first request including information of a first terminal and / or information of a first session, the first session being a communication session between the first terminal and a first server; responding to the first request, sending a first response, the first response including the artificial intelligence (AI) service capabilities of a first network, the first network being the network accessed by the first terminal.

[0008] In one possible implementation, the method further includes: determining the AI ​​service capabilities of the first network based on the information from the first terminal and / or the information from the first session.

[0009] In one possible implementation, determining the AI ​​service capabilities of the first network based on the information of the first terminal includes: determining the location information of the first terminal based on the information of the first terminal; and determining the AI ​​service capabilities of the first network based on the location information of the first terminal and a set of AI service capabilities of the network, wherein the set of AI service capabilities of the network includes the AI ​​service capabilities of the first network.

[0010] In one possible implementation, determining the AI ​​service capabilities of the first network based on the information of the first session includes: determining the first session based on the information of the first session; sending a second request to a first network element in the first network corresponding to the first session; and receiving a second response from the first network element, wherein the second response is a response to the second request and includes the AI ​​service capabilities of the first network element.

[0011] In one possible implementation, determining the first session based on the information of the first session includes: determining the first session based on the information of the first session and the information of the first terminal.

[0012] In one possible implementation, the method further includes: receiving a first indication, the first indication being used to indicate that the service corresponding to the first session is provided with AI computing assistance by the network; and sending a third request, the third request being used to request the first network to provide AI computing assistance to the service corresponding to the first session.

[0013] In one possible implementation, the method further includes: receiving information about an AI computing assistance strategy, the AI ​​computing assistance strategy including migrating all or part of the AI ​​computing functions located in the first server and / or the first terminal to a first network element of the first network; and sending the information about the AI ​​computing assistance strategy.

[0014] Thirdly, a communication method is provided, wherein the execution subject of the method is a second network element, or a module or unit located in the second network element. For example, the second network element may be an API gateway or an SMO, etc., and includes: receiving a first request, the first request including information about a first artificial intelligence (AI) computing resource to be subscribed; and responding to the first request by sending a first response, the first response including transport network layer information of the network element in the network that provides the first AI computing resource.

[0015] Through the above design, the network can open up idle AI computing resources in the network to third-party application AF network elements. AF network elements can subscribe to the first idle AI computing resource in the network and use the first idle AI computing resource in the network to perform corresponding operations, thereby improving the utilization rate of idle AI computing resources in the network.

[0016] In one possible implementation, the first AI computing resource is subscribed at the granularity of the entire network, or at the granularity of the first network element or the first set of network elements in the network.

[0017] In one possible implementation, the first AI computing resource is subscribed at the network-wide level, and the transport network layer information of the network element included in the first response is the transport network layer information of the second network element, wherein the second network element is the gateway of the first network element that provides the first AI computing resource in the network.

[0018] In one possible implementation, the first AI computing resource is subscribed to at the granularity of a first network element or a set of first network elements in the network, and the transmission network layer information of the network element included in the first response is the transmission network layer information of the first network element or the set of first network elements.

[0019] In one possible implementation, prior to receiving the first request, the method further includes sending a first report, which includes information about a second AI computing resource provided by the network.

[0020] In one possible implementation, the second AI computing resource is reported at the level of the entire network, or at the level of the first network element or the first set of network elements.

[0021] The fourth aspect is a method opposite to the third aspect, with beneficial effects as described in the third aspect. It provides a communication method in which the executing entity is an AF network element, or a module or unit located in the AF network element, including: sending a first request, the first request including information about a first artificial intelligence (AI) computing resource to be subscribed; and receiving a first response, the first response being a response to the first request, the first response including transmission network layer information of the network element providing the first AI computing resource in the network.

[0022] In one possible implementation, the first AI computing resource is subscribed at the granularity of the entire network, or at the granularity of the first network element or the first set of network elements in the network.

[0023] In one possible implementation, the first AI computing resource is subscribed at the network-wide level, and the transport network layer information of the network element included in the first response is the transport network layer information of the second network element, wherein the second network element is the gateway of the first network element that provides the first AI computing resource in the network.

[0024] In one possible implementation, the first AI computing resource is subscribed to at the granularity of a first network element or a set of first network elements in the network, and the transmission network layer information of the network element included in the first response is the transmission network layer information of the first network element or the set of first network elements.

[0025] In one possible implementation, prior to sending the first request, the method further includes receiving a first report, which includes information about a second AI computing resource provided by the network.

[0026] In one possible implementation, the second AI computing resource is reported at the level of the entire network, or at the level of the first network element or the first set of network elements.

[0027] Fifthly, a communication system is provided, comprising: a network element X sending a third request to a first network element, the third request being used to request a first network to provide artificial intelligence (AI) computing assistance for a service corresponding to a first session, wherein the first network element is a network element in the first network; the first network element providing AI computing assistance for the service corresponding to the first session, wherein the first session is a communication session between a first terminal and a first server. Optionally, network element X is an API gateway, or an SMO, or a UPF network element.

[0028] In one possible implementation, the method further includes: the network element X sending AI computing assistance strategy information to the first network element, and the first network element providing AI computing assistance to the service corresponding to the first session, including: the first network element providing AI computing assistance to the service corresponding to the first session according to the AI ​​computing assistance strategy.

[0029] In one possible implementation, the method further includes: network element X determining the AI ​​service capabilities of the first network based on information from the first terminal and / or information from the first session.

[0030] In one possible implementation, determining the AI ​​service capabilities of the first network based on the information of the first terminal includes: determining the location information of the first terminal based on the information of the first terminal; and determining the AI ​​service capabilities of the first network based on the location information of the first terminal and a set of AI service capabilities of the network, wherein the set of AI service capabilities of the network includes the AI ​​service capabilities of the first network.

[0031] In one possible implementation, determining the AI ​​service capabilities of the first network based on the information of the first session includes: determining the first session based on the information of the first session; sending a second request to a first network element in the first network corresponding to the first session; the first network element sending a second response to network element X, the second response being a response to the second request, and the second response including the AI ​​service capabilities of the first network element.

[0032] In one possible implementation, the method further includes: AF network element sending a first request to network element X, the first request including information about the first terminal and / or information about the first session; and network element X sending a first response to AF network element, the first response being a response to the first request, the first response including information about the artificial intelligence (AI) service capabilities of the first network.

[0033] In one possible implementation, the method further includes: the AF network element determining an AI computing assistance strategy based on information about the AI ​​service capabilities of the first network, wherein the AI ​​computing assistance strategy includes migrating all or part of the AI ​​computing functions located in the first server and / or the first terminal to the first network element of the first network; and the AF network element sending the information of the AI ​​computing assistance strategy to network element X.

[0034] In one possible implementation, the method further includes: the AF network element sends a first instruction to the network element X, the first instruction being used to indicate that the service corresponding to the first session is provided with AI computing assistance by the network.

[0035] Sixthly, an apparatus is provided capable of implementing the method of any one of the first to fourth aspects described above. For example, the apparatus includes modules, units, components, or means corresponding to any one of the first to fourth aspects. The modules, units, components, or means can be implemented in hardware, software, or a combination of hardware and software. The apparatus can be a first communication device, for example, an AF network element, a communication module within an AF network element, or a chip or chip system within an AF network element responsible for communication functions.

[0036] In one design, the device includes a unit that performs any of the first to fourth aspects described above.

[0037] In one design, the device includes a processor for executing a computer program or instructions stored in a memory, causing the device to implement the methods of any one of the first to fourth aspects described above. Optionally, the device also includes a memory, with the processor coupled to the memory.

[0038] In one design, the device includes a processor and an interface circuit. The interface circuit is used to receive signals from other devices outside the device and transmit them to the processor, or to send signals from the processor to other devices outside the device. The processor implements the methods in any one of the first to fourth aspects described above through logic circuits or executing code instructions.

[0039] In one design, the device may be a module, unit, or component (e.g., a chip, chip system, circuit, or processor, etc.) that corresponds one-to-one with the method / operation / step / action described in any one of the first to fourth aspects of the first communication device, or it may be compatible with the first communication device.

[0040] In a seventh aspect, a computer-readable storage medium is provided, storing a computer program or instructions that, when executed on a computer, cause the computer to implement the methods of any one of the first to fourth aspects described above.

[0041] Eighthly, a computer program product is provided, comprising a computer program or instructions that, when executed by a computer, cause the methods of any one of the first to fourth aspects to be performed.

[0042] A ninth aspect provides a chip system including a processor for executing computer programs or instructions stored in a memory, such that the chip system implements the methods of any one of the first to fourth aspects described above. Optionally, the chip system further includes a memory, with the processor coupled to the memory. Attached Figure Description

[0043] Figure 1 A schematic diagram of the communication system provided in the embodiments of this application;

[0044] Figure 2 , Figure 3a , Figure 3b and Figure 4 The network open architecture provided in the embodiments of this application;

[0045] Figure 3c This is a schematic diagram of the CAPIF architecture provided in an embodiment of this application;

[0046] Figure 5 , Figure 6 , Figure 8 and Figure 10 A schematic diagram of the communication process provided in the embodiments of this application;

[0047] Figure 7 , Figure 9 and Figure 11 This is a schematic diagram of the communication architecture provided in an embodiment of this application;

[0048] Figure 12 and Figure 13 This is a schematic diagram of the device provided in an embodiment of this application. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings. The specific operating methods and functional descriptions in the method embodiments can also be applied to the device embodiments or system embodiments.

[0050] It is understood that, in the embodiments of this application, the number of nouns, unless otherwise specified, refers to "singular nouns or plural nouns," that is, "one or more." "At least one" means one or more, and "more than one" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, or B exists alone, where A and B can be singular or plural. In the textual description of this application, the character " / " generally indicates that the related objects before and after are in an "or" relationship; in the formulas of this application, the character " / " indicates that the related objects before and after are in a "division" relationship. "Including at least one of A, B, or C" or similar expressions can mean: including A; including B; including C; including A and B; including A and C; including B and C; including A, B, and C, where A, B, and C can be singular or plural.

[0051] The various numerical designations used in the embodiments of this application are for descriptive convenience and are not intended to limit the scope of the embodiments of this application. The order of the process numbers does not imply the order of execution; the execution order of each process should be determined by its function and internal logic. For example, as follows... Figure 5 Step 550a can be executed before step 570a, or step 550a can be executed after step 570a, or steps 550a and 570a can be executed simultaneously. The ordinal numbers such as "first" and "second" used in the embodiments of this application are used to distinguish multiple objects and do not limit the size, order, timing, priority, or importance of the multiple objects.

[0052] Figure 1 This is a schematic diagram of one possible, non-limiting system. For example... Figure 1 As shown, the communication system 10 includes a terminal, a radio access network (RAN), and a core network (CN).

[0053] 1. Terminal

[0054] The terminal can be referred to as a terminal device, user equipment (UE), mobile station, mobile terminal, etc. Terminals can be widely used in various scenarios, such as device-to-device (D2D), vehicle-to-everything (V2X) communication, machine-type communication (MTC), Internet of Things (IoT), virtual reality, augmented reality, industrial control, autonomous driving, telemedicine, smart grids, smart furniture, smart offices, smart wearables, smart transportation, smart cities, etc. Terminals can be mobile phones, tablets, computers with wireless transceiver capabilities, wearable devices, vehicles, drones, helicopters, airplanes, ships, robots, robotic arms, smart home devices, etc. The embodiments of this application do not limit the device form of the terminal.

[0055] 2. RAN

[0056] The RAN is used to implement functions related to wireless access. The RAN includes at least one RAN node, which forms part of the communication system and helps terminals achieve wireless access. Multiple RAN nodes can be of the same type or different types.

[0057] Among them, RAN can be a cellular system related to the 3rd Generation Partnership Project (3GPP), such as 4G (4G4). th generation, 4G), fifth generation (5G) th RAN can be a generation (5G) mobile communication system, or a future-oriented evolution system (e.g., a future communication network). RAN can also be an open access network (O-RAN or ORAN), a cloud radio access network (CRAN), or a wireless fidelity (WiFi) system. RAN can also be a communication system that integrates two or more of the above systems.

[0058] In one possible scenario, a RAN node can be a base station, an evolved NodeB (eNodeB), an access point (AP), a transmission reception point (TRP), a next-generation NodeB (gNB), a next-generation base station in a future communication network, a base station in a future communication network, or an access node in a WiFi system. A RAN node can be a macro base station, a micro base station, an indoor station, a relay node, a donor node, or a radio controller in a CRAN scenario. Optionally, a RAN node can also be a server, a wearable device, a vehicle, or in-vehicle equipment. For example, the access network equipment in vehicle-to-everything (V2X) technology can be a roadside unit (RSU). All or part of the functions of the RAN node in this application can also be implemented through software functions running on hardware, or through virtualization functions instantiated on a platform (e.g., a cloud platform). The RAN node in this application can also be a logical node, logical module, or software capable of implementing all or part of the RAN node functions.

[0059] In another possible scenario, multiple RAN nodes collaborate to assist the terminal in achieving wireless access, with each RAN node performing a portion of the base station's functions. For example, RAN nodes can be central units (CUs), distributed units (DUs), CU-control plane (CPs), CU-user plane (UPs), or radio units (RUs), etc. CUs and DUs can be separate entities or included in the same network element, such as a baseband unit (BBU). RUs can be included in radio frequency equipment or radio frequency units, such as remote radio units (RRUs), active antenna units (AAUs), or remote radio heads (RRHs).

[0060] In different systems, CU (or CU-CP and CU-UP), DU, or RU may have different names, but those skilled in the art will understand their meaning. For example, in an ORAN system, CU can also be called open-CU (open-CU, O-CU), DU can also be called open-DU (open-DU, O-DU), CU-CP can also be called open-CU-CP (open-CU-CP, O-CU-CP), CU-UP can also be called open-CU-UP (open-CU-UP, O-CU-UP), and RU can also be called open-RU (open-RU, O-RU). For ease of description, this application uses CU, CU-CP, CU-UP, DU, and RU as examples. Any of the units among CU (or CU-CP, CU-UP), DU, and RU in this application can be implemented through software modules, hardware modules, or a combination of software and hardware modules.

[0061] In the embodiments of this application, the RAN node may also be referred to in other ways, such as as an access network device, RAN entity, or access node. For ease of description, the following description of specific embodiments will use an access network device as an example.

[0062] 3. Core Network

[0063] The core network is primarily used for managing terminals and providing communication functions with the external network. It includes at least one of the following network elements: User Plane Function (UPF) network element, Access and Mobility Management Function (AMF) network element, Network Exposure Function (NEF) network element, Application Function (AF) network element, or Application Programming Interface (API) gateway, etc.

[0064] Optionally, the communication system 10 also includes a data network (DN). Terminals, access network equipment, UPF network elements, and the DN are generally referred to as user plane network elements (or user plane functions and entities). User plane data flows can be transmitted through protocol data unit (PDU) sessions established between the terminal and the DN. The user plane is used to carry service data. Other network elements in the communication system 10, such as AMF, NEF, or AF, can be referred to as control plane network elements (or control plane functions and entities), primarily responsible for authentication and authorization, registration management, session management, mobility management, and policy control functions, thereby achieving reliable and stable transmission of user plane data flows. The control plane is used to carry control signaling messages.

[0065] It is understandable that the names of network elements in the core network are not restricted. For example, in a 5G communication system, the network element that implements signaling processing is called an AMF (Active Signalling Function) element. In future communication networks, the network element that implements the above functions can also be called by other names, without restriction. The core network may also include other network elements besides those described above, such as session management function (SMF) network elements and policy control function (PCF) network elements.

[0066] System 10 also includes Operation Administration and Maintenance (OAM), a collective term for a set of network management functions, including fault monitoring, fault reporting, fault location, and fault repair. The OAM corresponding to the access network and the core network can be the same or different. For example, when the OAM corresponding to the access network and the core network are different, the OAM corresponding to the access network is called the access network OAM, and the OAM corresponding to the core network is called the core network OAM.

[0067] It is understandable that access network equipment, terminals, network elements in the core network, and OAM (Operational Access Management) devices can be referred to as communication devices. For example, access network equipment can be understood as a communication device with base station functions; a terminal can be understood as a communication device with terminal functions; network elements in the core network can be understood as communication devices with core network element functions; and OAM can be understood as a communication device with OAM functions.

[0068] In communication networks, network capabilities can be exposed to third-party applications (such as AF network elements). For example... Figure 2 As shown, the current network capability opening mainly includes the following three paths:

[0069] Path #1: Open network capabilities through the north interface (Itf-N) of the OAM network element.

[0070] Path #2: Open network capabilities through the NEF_API interface of NEF network elements.

[0071] Path #3: Open network capabilities through the UPF_API interface of the UPF network element.

[0072] The network capabilities opened through the above three paths are aggregated at the API gateway (GW) and provided to the AF network element by the API gateway through the service_API interface.

[0073] Currently, the network capabilities offered are primarily communication-oriented, meaning that only communication capabilities are available to third-party applications, and this access is achieved through the core network and OAM network elements. As network architecture evolves, the access network, in addition to traditional communication capabilities, may also possess new service / business capabilities such as artificial intelligence (AI), computing, sensing, and positioning. If network capabilities continue to be offered through the current three paths, it may result in data detours, increased response latency, and a degraded user experience.

[0074] This application provides a new network capability open architecture that opens up new service capabilities on the access network side through multiple path combinations, meets the needs of third-party applications for new service capabilities, and accelerates the monetization of network service capabilities.

[0075] like Figure 3aAs shown, a RAN capability open framework is provided. New service capabilities on the access network side are carried by service units (SUs), while traditional communication capabilities are carried by CUs and DUs. The new service capabilities on the access network side include at least one of the following functions: AI, computing, sensing, and positioning. For example, AI functions include AI prediction and / or AI computing functions; sensing functions include generating sensing point cloud information and / or providing sensing object information; and positioning functions include locating the terminal's location information, accurate to the terminal's latitude, longitude, and altitude.

[0076] Specifically, the SU exposes new access network-side service capabilities to external parties through the SU_API interface. For example, the interface between the SU and the API gateway is SU_API; the SU reports new access network-side service capabilities to the API gateway through the SU_API interface, and the API gateway reports these capabilities to the AF network element through the service_API interface. Alternatively, SU_API can be the interface between the SU and the AF network element; the SU directly reports new access network-side service capabilities to the AF network element through the SU_API interface. Furthermore, the SU can also expose new access network-side service capabilities to external parties in the following ways:

[0077] Method 1: The SU (Supply Provider) exposes new access network-side service capabilities through the Itf-N interface of the OAM (Operating Access Management) network element. For example, the interface between the SU and the OAM network element is a southbound interface (Itf-S). The SU reports the new access network-side service capabilities to the OAM network element through Itf-S, and the OAM network element reports the new access network-side service capabilities to the API gateway through the Itf-N interface. The API gateway reports the new access network-side service capabilities to the AF (Action Center) network element through the service_API interface.

[0078] Method 2: The SU (Supply Provider) exposes new access network-side service capabilities through the UPF_API interface of the UPF network element. For example, the SU reports new access network-side service capabilities to the CU (Comprehensive Utility). The interface between the CU and the UPF network element is the N3 interface. The CU reports the new access network-side service capabilities to the UPF network element through the N3 interface. The UPF network element reports the new access network-side service capabilities to the API gateway through the UPF_API interface. The API gateway reports the new access network-side service capabilities to the AF (Agency Utility) network element through the service_API interface.

[0079] Method 3: The SU (Supply Entity) exposes new access network-side service capabilities through the NEF_API interface of the NEF network element. For example, the SU reports new access network-side service capabilities to the CU (Currency Entity), the CU reports these capabilities to the AMF (Aggregator Entity) network element through the N2 interface, the AMF network element reports them to the NEF network element, the NEF network element reports them to the API gateway through the NEF_API interface, and the API gateway reports them to the AF (Aggregator Entity) network element through the service_API interface.

[0080] It's understandable that an API gateway can also be called an API aggregator. The API gateway acts as a connection device between third-party application AF network elements and the communication network. In uplink communication, the API network element can aggregate network capabilities reported by various network elements and report them to the third-party application AF network element through the service_API interface. In one description, the service_API interface is the union of the SU_API interface, Itf-N interface, NEF_API interface, and UPF_API interface. For example, service_API interface = SU_API interface ∪ Itf-N interface ∪ NEF_API interface ∪ UPF_API interface. In downlink communication, the API gateway can receive service requests from third-party application AF network elements through the service_API interface and send these service requests to the communication network.

[0081] In one possible implementation, such as Figure 3b As shown, the access network includes SU, CU, and DU. The interface between SU ​​and CU is Si. CU and DU are responsible for traditional communication functions, while SU is responsible for beyond communication functions, which can be new service capabilities / functions on the access network side. For example... Figure 3b The diagram illustrates the RAN capability open architecture based on SU. The specific functions of SU include at least one of the following:

[0082] 1. RAN Exposure Function (REF): Based on the Common API framework (CAPIF), it is responsible for exposing new service capabilities of the access network side to the outside world.

[0083] 2. RAN service management function (RSMF): Used for service function registration, updating, discovery, status detection, authentication and authorization, and inter-service communication within the SU.

[0084] 3. Communication & Service Collaboration Function: This function is used to provide services to external parties in collaboration with communication and new business capabilities. It can be combined with the RSMF function.

[0085] Furthermore, SU also provides at least one of the following functions: RAN to AI (RAN4AI), AI to RAN (AI4RAN), or sensing / location. RAN4AI is also known as network to AI (Net4AI), and AI4RAN is also known as AI to network (AI4Net).

[0086] In one possible implementation, such as Figure 3c As shown, CAPIF can be divided into the following four logical functions: CAPIF core function, API exposing function, API publishing function, and API management function.

[0087] The core functions of CAPIF mainly include: authenticating and authorizing API users; publishing, storing, and supporting the discovery of service API information; configuring / storing policy information to control service API access; storing service API call logs and providing them to authorized entities; charging based on service API call logs; monitoring service API calls; and supporting auditing of access logs. API opening functions include: service API providers, through which API users communicate with the API. API publishing functions include: publishing service APIs to API Gateways or third-party applications. API management functions include: monitoring and querying service API call activity.

[0088] In this embodiment, the four logical functions of CAPIF are mapped to one or more network elements. For example, the core CAPIF function is mapped to the API gateway, which supports the core CAPIF function. The API open function is mapped to the SU, which supports the API open function. In this embodiment, "API invoker" refers to a third-party application that needs to invoke communication network capabilities, which can be within or outside the trusted domain of the public land mobile network (PLMN). The server corresponding to this third-party application can be deployed... Figure 1In the DN of the communication system, the descriptions of "API user," "server," and "AF network element" in the embodiments of this application are interchangeable. The interfaces between different logical functions, and the interfaces between the API user and logical functions, refer to... Figure 3c As shown.

[0089] exist Figure 3a In the RAN capability open architecture based on SU shown, SU quickly opens up new access network-side service capabilities to third-party application AF network elements through the SU_API interface. Furthermore, various network open capabilities can be aggregated at the API gateway, providing integrated network capabilities to third-party application AF network elements.

[0090] like Figure 4 As shown, an open architecture for O-RAN capabilities is provided. (And...) Figure 3a The main differences in architecture are: Figure 3a In this context, SU is replaced with near-real-time RAN intelligent controller (RIC), and API gateway is replaced with service management and orchestration (SMO).

[0091] Among them, such as Figure 4 As shown, new service functions on the access network side are carried by the near real-time RIC, while traditional communication functions are carried by the CU and DU. The interface between the near real-time RIC and the DU or CU is E2. The interface between the near real-time RIC and the SMO is AI, through which the near real-time RIC can expose new service capabilities on the access network side. For example, the A1 interface is the interface between the near real-time RIC and the SMO. The near real-time RIC reports new service capabilities on the access network side to the SMO through the A1 interface, and the SMO reports these new service capabilities to the AF network element through the service_API interface. Alternatively, the A1 interface is the interface between the near real-time RIC and the AF network element, allowing the near real-time RIC to directly report new service capabilities on the access network side to the AF network element through the A1 interface. Furthermore, the near real-time RIC can also expose new service capabilities on the access network side through the Itf-N interface of the OAM network element, the UPF_API interface of the UPF network element, or the NEF_API interface of the NEF network element. The near real-time RIC exposes new service capabilities on the access network side through these interfaces, and... Figure 3a The process by which the SU opens up new service capabilities on the access network side through the above-mentioned interface is similar, as can be found in the explanation above.

[0092] Understandably, the SMO, acting as a connection device between the third-party application AF network element and the communication network, uses the Service API interface as its interface. In uplink communication, the SMO can aggregate network capabilities reported by various network elements and report them to the third-party application AF network element via the Service API interface. In downlink communication, the SMO can receive service requests from the third-party application AF network element through the Service API interface and forward these requests to the communication network.

[0093] In the embodiments of this application, the executing entity can be an access network device (e.g., SU or near real-time RIC), or a core network device (e.g., AF network element, API gateway, or SMO, etc.), or a unit / module in the access network device (e.g., chip, chip system, circuit, or others, etc.), or a unit / module in the core network device. The following description uses access network devices and core network devices as examples of executing entities. When the executing entity is an access network device or a unit / module in the core network device, receiving / transmitting can be understood as input / output, that is, the unit / module communicates with other units / modules or components of the access network device or core network device. Furthermore, the processing performed by a single executing entity can also be divided into multiple executing entities, which can be logically and / or physically separated. For example, the processing performed by the SU can be understood as requiring the protocols of the CU and / or DU to perform the corresponding processing.

[0094] Example 1

[0095] based on Figure 3a or Figure 4 The illustrated open capability architecture, taking the AI ​​service capabilities of the open access network as an example, provides a communication method in which the AF (Analog-Based Query) network element queries the AI ​​service capabilities of the first network accessed by the first terminal. Furthermore, the AF network element can determine an AI computing assistance strategy based on the AI ​​service capabilities of the first network accessed by the terminal. For example, all or part of the functionality of the AI ​​model located on the first server and / or the first terminal can be migrated to the first network, thereby reducing the AI ​​computing load on the first terminal and / or the first server, quickly responding to the needs of AI services, and improving the service experience of AI services. Figure 5 As shown, a flowchart is provided, including:

[0096] Step 510: AF network element sends the first request, and network element X receives the first request.

[0097] The first request includes information about the first terminal and / or the first session. For example, the first terminal's information may be its identity document (ID), such as its subscription permanent identifier (SUPI) or its telephone number. Alternatively, the first terminal's information may be its transport network layer information, such as its internet protocol (IP) address or its uniform resource locator (URL) address. Or, the first terminal's information may be its generic public subscription identifier (GPSI).

[0098] The first session can be short for PDU session. The first session is used for the logical connection between the first terminal and the first server; it is a user plane connection from the first terminal to the first server. For example, the first session specifically includes: a session between the first terminal and the access network, between the access network and the UPF network element, and between the UPF network element and the first server. In the following description, the anchor point or target network corresponding to the first session specifically refers to the access network corresponding to the first session, and the first server can be deployed in the DN. For example, the first server can be a server corresponding to a third-party (3P) application, an over-the-top (OTT) consumer, or an enterprise application. In this embodiment, the first server is mainly described as a server providing third-party applications. The first server can perform AI calculations; for example, an AI model is deployed in the first server, and the first server can use the AI ​​model to perform model inference to achieve corresponding functions. Furthermore, the first server can also train the AI ​​model to update it, etc.; the first server can be called an AI server. The information for the first session can be the ID of the first session, or the transport network layer information of the first server corresponding to the first session, such as the IP address of the first server or the URL address of the first server.

[0099] Optionally, in step 520: Network element X determines the AI ​​service capabilities of the first network based on the information of the first terminal and / or the information of the first session.

[0100] In one description, the AI ​​service capability of the first network is determined based on information from the first terminal and / or information from the first session. The first network can be an access network, a core network, or a transmission network. In the description of the embodiments of this application, the "first network" is mainly used as an example of an access network. When the first network is an access network, the AI ​​service capability of the first network can be described as the AI ​​service capability of the access network. It is understood that... Figure 3a In the network open architecture shown, the AI ​​service capability of the access network refers to the AI ​​service capability of the SU network element. Figure 4 In the network open architecture shown, the AI ​​service capability of the access network refers to the AI ​​service capability of the near real-time RIC network element. Furthermore, as described above, the new service capabilities on the access network side include: AI, computing, sensing, and positioning. Using the method of this application embodiment, in addition to opening the AI ​​service capability of the access network side to third-party application AF network elements, at least one service capability of the access network side, such as computing, sensing, or positioning, can also be opened to third-party application AF network elements. The AF network element can perform corresponding operations based on the opened at least one service capability of the access network side, such as computing, sensing, or positioning, without limitation.

[0101] In one possible implementation, network element X can be... Figure 3a The API gateway in the network open architecture shown, or Figure 4 The network open architecture shown illustrates the SMO (Service Management Object) mechanism. Network element X can determine the location information of the first terminal based on its information, such as the user location information (ULI) or the global positioning system (GPS) information. Network element X then determines the AI ​​service capabilities of the first network accessed by the first terminal based on the first terminal's location information and the network's AI service capability set. For example, before step 510, each network element reports its corresponding AI service capabilities to network element X, forming a network AI service capability set. Network element X then queries the first network's AI service capabilities from the network's AI service capability set based on the first terminal's location information. The first network is the network accessed by the first terminal, or, as described, the network through which the first terminal accesses. See Example 1 below for details. The following description primarily uses the first terminal's ULI as the example to illustrate the location information of the first terminal.

[0102] In another possible implementation, network element X can refer to a UPF network element. The UPF network element can determine the first session based on the information of the first session. Furthermore, in scenarios where session identifiers are assigned at the terminal level, the information of the first session alone may not uniquely identify a session; it is also necessary to combine it with the information of the first terminal. That is, the UPF network element can determine the first session based on both the information of the first session and the information of the first terminal. The first session is the session used for communication between the first terminal and the first server. The anchor or target network in the first session can be called the first network; that is, the first network is the anchor or target network in the first session. The UPF network element can query the AI ​​service capabilities of the first network from the first network element in the first network. For example, the UPF network element sends a second request to the first network element in the first network, the second request being used to request the AI ​​service capabilities of the first network element; the first network element in the first network sends a second response to the UPF network element, the second response being a response to the second request, and the second response including the AI ​​service capabilities of the first network element. Figure 3a In the network open architecture shown, the first network element in the first network refers to the SU network element on the access network side. Figure 4 In the network open architecture shown, the first network element in the first network refers to the near real-time RIC network element on the access network side. See Example 2 below for details.

[0103] Step 530: In response to the first request, network element X sends a first response, and network element AF receives the first response. The first response includes the AI ​​service capabilities of the first network. It can be understood that the first response is a response to the first request.

[0104] Optionally, in step 540: the AF network element determines the AI ​​computing assistance strategy based on the AI ​​service capabilities of the first network.

[0105] Optionally, the AF network element can also send AI computing assistance strategy information to the first network element, which then executes corresponding AI computing assistance operations based on the strategy. For example, the first network element can interact with the first server and / or the first terminal with the assistance of the CU and DU, thereby migrating all or part of the AI ​​computing functions of the first server and / or the first terminal to the first network element, thus reducing the AI ​​computing load on the first server and / or the first terminal.

[0106] In one possible implementation, the AF network element can send AI computing assistance strategy information to the first network element through network element X. For example, the AF network element sends the AI ​​computing assistance strategy information to network element X, where network element X can refer to an API gateway or an SMO. Network element X then sends the AI ​​computing assistance strategy information to the first network element. Optionally, for example, Figure 5 The process also includes:

[0107] Step 550a: AF network element sends information about the AI ​​computing assistance strategy, and network element X receives the information about the AI ​​computing assistance strategy.

[0108] Step 560a: Network element X sends information about the AI ​​computing assistance strategy, and the first network element receives the information about the AI ​​computing assistance strategy.

[0109] In another possible implementation, the AF network element can send AI computing assistance strategy information to the first terminal through the application layer, and the first terminal can send the AI ​​computing assistance strategy information back to the first network element. For example, optionally, Figure 5 The process also includes:

[0110] Step 550b: The AF network element sends information about the AI ​​computing assistance strategy, and the first terminal receives the information about the AI ​​computing assistance strategy.

[0111] Step 560b: The first terminal sends information about the AI ​​computing assistance strategy, and the first network element receives the information about the AI ​​computing assistance strategy.

[0112] Optionally, the AF network element can also send a first instruction, which indicates that the service corresponding to the first session is provided with AI computing assistance by the network (or needs, desires, or requests, etc.). In one possible implementation, the AF network element can send the first instruction to network element X, where network element X refers to the API gateway or SMO. Network element X sends a third request to the first network element in the first network, which requests the first network element to provide AI computing assistance for the service corresponding to the first session. Upon receiving the third request, the first network element can perform corresponding operations according to the AI ​​computing assistance policy. That is, the third request can act as a trigger, and when the first network element receives the trigger of the third request, it performs corresponding operations according to the AI ​​computing assistance policy. Alternatively, when the first network element receives information about the AI ​​computing assistance policy, it can directly perform corresponding operations according to the AI ​​computing assistance policy, etc., without restriction. For example, optionally, Figure 5 The process also includes:

[0113] Step 570a: AF network element sends the first instruction, and network element X receives the first instruction.

[0114] For example, when network element X receives the first instruction, it obtains information about the first network element in the first network. In step 580a, network element X sends a third request to the first network element.

[0115] Step 580a: Network element X sends a third request, and the first network element receives the third request.

[0116] The third request can also be referred to as a first network element auxiliary request. For example, network element X can be an API gateway, SMO, or UPF network element, etc. Network element X can send a third request to the first network element through the corresponding interface. When the first network element receives the third request, it can establish a data channel with the CU. When the CU receives the service data corresponding to the first session, it sends the service data corresponding to the first session to the first network element through the data channel. The first network element provides AI computing assistance for the corresponding data according to the AI ​​computing assistance strategy. Optionally, the third request can carry information about the first session.

[0117] In another possible implementation, the AF network element can send a first instruction to the first terminal through the terminal's application layer, and the first terminal can send a third request to the first network element. For example, optionally, Figure 5 The process also includes:

[0118] Step 570b: The AF network element sends a first instruction, and the first terminal receives the first instruction.

[0119] Step 580b: The first terminal sends a third request, and the first network element receives the third request.

[0120] For example, the AF network element sends a first instruction to the first terminal through the application layer, and the first terminal receives the first instruction through the application layer. The first terminal obtains the service corresponding to the application layer, which is the service corresponding to the first session. The first terminal sends a third request to the first network element. In one possible implementation, the first terminal and the first network element are not directly connected; they communicate through the CU. For example, the first terminal sends a third request to the CU, and a data channel is established between the CU and the first network element. The CU can send the third request to the first network element through this data channel. Further, when the CU receives the service data corresponding to the first session, it forwards the service data of the first session to the SU through the established data channel, and the SU performs AI calculation-assisted processing on the corresponding data. Alternatively, in another possible implementation, the CU transparently transmits the information between the first terminal and the first network element, and the first terminal sends the third request to the first network element through the CU. A data channel is established between the first network element and the CU. When the CU receives the service data corresponding to the first session, it forwards the service data of the first session to the SU through the established data channel, and the SU performs corresponding AI calculation-assisted processing. Optionally, the third request may carry information about the first session.

[0121] It is understandable that the first network element can be Figure 3aThe network open architecture shown depicts a SU. A first terminal can send AI computing assistance policies and / or third requests to the SU via a DU or CU. For example, the DU receives a third request from the first terminal and sends it to the CU, which then sends the third request to the SU via the Si interface. Alternatively, the first network element could be... Figure 4 The network open architecture shown depicts a near real-time RIC. A first terminal can send AI computing assistance policies and / or third requests to the near real-time RIC via a DU and / or CU. For example, the DU receives a third request from the first terminal and sends it to the near real-time RIC via the E2 interface. Alternatively, the DU sends the third request to the CU, and the CU sends the third request to the near real-time RIC via the E2 interface.

[0122] Understandably, the AF side cannot identify the first network element in the first network. Therefore, the first indication sent by the AF is used to instruct the network (e.g., the access network) to provide AI computing assistance for the service corresponding to the first session. When network element X (API network element or SMO) or the first terminal receives the first indication, it can determine the first network accessed by the first terminal, or the anchor / target network (first network) in the first session. Network element X (API network element or SMO) or the first terminal can send a third request to the first network element in the first network. This third request is used to instruct the first network to provide AI computing assistance for the service corresponding to the first session. When the first network element in the first network receives the third request, it provides AI computing assistance for the service corresponding to the first session according to the AI ​​computing assistance strategy.

[0123] exist Figure 5 In the process, the AI ​​computing assistance strategy information in step 550a and the first instruction in step 570a can be sent separately or together in a single message. For example, AF network element sends a message to network element X, which includes the AI ​​computing assistance strategy information and the first instruction. Similarly, the AI ​​computing assistance strategy information in step 560a and the third request in step 580a, or the AI ​​computing assistance strategy information in step 550b and the first instruction in step 570b, or the AI ​​computing assistance strategy information in step 560b and the third request in step 580b, can also be sent separately or together in a single message without restriction.

[0124] Example 1

[0125] by Figure 5 The process plan is applied to Figure 3a The network open architecture shown is an example where the first network is the access network, the first network element is SU, and network element X is the API gateway. Figure 6As shown, a flowchart is provided. In this process, the SU actively reports its AI service capabilities to the API gateway. When the API gateway receives the first request from the AF, it queries the AI ​​service capabilities reported by the SU and checks the service capabilities of the SU in the access network accessed by the first terminal in the first request. This process includes:

[0126] Step 610: One or more SUs send a capability report to the API gateway, which includes the SU's AI service capabilities.

[0127] In one possible implementation, the SU in the access network can report its AI service capabilities to the API gateway via the SU_API interface. Alternatively, the SU can also... Figure 3a The other interfaces shown report SU's AI service capabilities to the API gateway. For example, SU's AI service capabilities are reported to the API gateway via OAM's ITF-N interface, UPF's UPF_API interface, or NEF's NEF_API interface. Figure 6 In this process, the AI ​​service capabilities of the SU can also be replaced by the AI ​​service capabilities of the access network. For example, the AI ​​service capabilities reported by the SU include: the AI ​​models supported by the SU, or information such as the current load of the SU.

[0128] It is understandable that the capability report submitted by the SU, in addition to AI service capabilities, also includes at least one of the following: SU information, CU information connected to the SU, access network equipment information corresponding to the SU, and cell information provided by the CU connected to the SU. For example, the SU information could be the SU's ID or its transport network layer information, such as the SU's IP address or URL address. The SU and CU can be connected via interface Si. The CU information connected to the SU could be the IP address of the CU connected to the SU, or its transport network layer information (e.g., the CU's IP address or URL address). The access network equipment corresponding to the SU could be the access network equipment to which the SU belongs. The access network equipment information could refer to the identifier of the access network equipment corresponding to the SU (e.g., the base station identifier) ​​or its transport network layer information. The cell information of the CU connected to the SU could refer to the identifier of the cell provided by the CU connected to the SU.

[0129] Understandably, the API gateway can generate a set of AI service capabilities for a given SU ​​based on capability reports submitted by one or more SUs. This set of AI service capabilities for a given SU ​​is as described above. Figure 5One possible implementation of the AI ​​service capability set of the access network in the process. The AI ​​service capability set of the SU includes one or more AI service capabilities of the SU, and each AI service capability of the SU corresponds to an identifier (or index). The identifier is used to identify an AI service capability. The identifier can be the SU information mentioned above, the CU information connected to the SU, the access network equipment corresponding to the SU, or the cell information of the CU connected to the SU, etc.

[0130] Step 620: The first terminal establishes the first session with the first server.

[0131] Understandably, the first session is used for communication between the first terminal and the first server. For example, the first server might be a server for a third-party application, such as a server for a video application or a game application. In uplink communication scenarios, the first terminal uses the first session to transmit uplink data of the application to the first server. In downlink communication scenarios, the first server uses the first session to transmit downlink data of the application to the first terminal. For example, if the first server and / or the first terminal have an AI model deployed, the first session can be used to transmit AI-related uplink and / or downlink data in the application. This AI-related uplink and / or downlink data could be AI intermediate data or AI results. For instance, the first terminal uses the AI ​​model to perform model inference and obtains an AI result. This AI result serves as AI intermediate data, which the first terminal sends to the first server through the first session. The first server uses the AI ​​model to perform model inference on the AI ​​intermediate data to obtain the final AI calculation result. Furthermore, the first server can send the final AI calculation result to the first terminal through the first session. In this case, the first session can be called a PDU session for AI services.

[0132] The first session established in step 620 may not involve SU. For example, the uplink path of the first session is: first terminal → RU → DU → CU → UPF → first server, and the downlink path of the first session is: first server → UPF → CU → DU → RU → first terminal.

[0133] Step 630: The AF network element sends a first request to the API gateway. The first request is used to request the AI ​​service capabilities of the access network.

[0134] The first request can be termed a network capability query request, and it includes information about the first terminal. Optionally, the first request may also include information about the first session. In one possible implementation, if the AF network element wishes to migrate all or part of the AI ​​computing functionality in the first server and / or the first terminal to the first network (i.e., the access network), the AF network element may send the first request to the API gateway. More specifically, for example, when the load on the first server and / or the first terminal exceeds a threshold, the AF network element may send the first request to the API gateway to query the AI ​​service capabilities of the access network. The AF network element can query the information of the first terminal of the application corresponding to it at the application layer, and further, it can query the information of the first session between the first terminal and the first server for that application, carrying the information of the first terminal and / or the first session in the first request.

[0135] In this embodiment of the application, the first server can be deployed in Figure 1 In the DN of the communication system shown, the first server can be a server for a third-party application. When the first server registers with the network (DN), the control plane generates an AF network element for that application. In the description of this application, "application" and "service" are interchangeable.

[0136] Step 640: The API gateway determines the ULI of the first terminal based on the information of the first terminal.

[0137] For example, upon receiving a first request, the API gateway obtains information about the first terminal from that request. Based on this information, the API gateway initiates a location request to the NEF network element to obtain the first terminal's ULI. For instance, the API gateway sends a location request to the NEF network element, which includes the first terminal's information. The NEF network element determines the first terminal's ULI based on this information. For example, the NEF network element can obtain the first terminal's ULI from the AMF network element. The NEF network element sends a location response to the API gateway, which includes the first terminal's ULI. For example, the first terminal's ULI can refer to: information about the cell the first terminal accesses, or information about the access network equipment the first terminal accesses, or information about the SU (Supply Unit) of the access network equipment the first terminal accesses, or information about the CU (Curricular Unit) of the access network equipment the first terminal accesses, etc., without limitation.

[0138] Step 650: The API network element determines the AI ​​service capabilities of the first SU based on the ULI of the first terminal and the AI ​​service capability set of the SU.

[0139] For example, an API network element can query the AI ​​service capabilities of the SU corresponding to the first terminal's ULI from the SU's AI service capability set, and use this as the first SU's AI service capability. It can be understood that the first SU is the SU of the access network to which the first terminal is connected, and the AI ​​service capabilities of the first SU are the AI ​​service capabilities of the SU of the access network to which the first terminal is connected. The AI ​​service capabilities of the first SU are... Figure 5 One possible implementation of the AI ​​service capabilities of the first access network in the process. For example, the ULI of the first terminal is specifically: the information of the cell accessed by the first terminal. The AI ​​service capability set of the SU includes at least one AI service capability of the SU, and the identifier corresponding to each AI service capability of the SU is the cell information provided by the CU connected to the SU. Based on the information of the cell accessed by the first terminal, the API gateway queries the AI ​​service capability corresponding to the cell accessed by the first terminal in the AI ​​service capability set of the SU. The AI ​​service capability of the cell accessed by the first terminal is the AI ​​service capability of the first SU. Alternatively, for another example, the ULI of the first terminal is specifically: the information of the access network device accessed by the first terminal. The difference is that, in this case, the identifier corresponding to each AI service capability in the AI ​​service capability set of the SU is: the information of the access network device corresponding to the SU.

[0140] Step 660: The API gateway sends a first response to the AF network element. The first response includes the AI ​​service capabilities of the first SU.

[0141] The first response can be called the network capability response.

[0142] Step 670: The AF network element determines the AI ​​computing assistance strategy based on the AI ​​service capabilities of the first SU.

[0143] Specifically, the AI ​​service capability of the first SU refers to the AI ​​service capability of the SU in the access network to which the first terminal is connected. For example, the AI ​​service capability of the first SU includes information such as the AI ​​models supported by the SU in the first access network to which the first terminal is connected, and / or the current load of the SU. The AF network element can determine an AI computing assistance strategy based on the AI ​​service capability of the SU in the first access network to which the first terminal is connected. For example, the AI ​​computing assistance strategy includes migrating all or part of the AI ​​computing functions of the first server and / or the first terminal to the SU in the first access network (i.e., the first SU), thereby reducing the AI ​​computing load of the first server and / or the first terminal. In one possible implementation, the AI ​​computing assistance strategy can be specifically an AI model segmentation strategy, which specifically involves migrating all or part of the functions of the AI ​​models in the first server and / or the first terminal to the first SU. For example, if a 100-layer model A is deployed in the first server, and the AI ​​service capability of the SU of the access network accessed by the first terminal includes: the SU supports model A, and the current load of the SU is less than the threshold, then the AI ​​computing assistance strategy determined by the AF network element can be: to migrate 50 layers of model A deployed in the first server to the SU of the access network accessed by the first terminal (i.e., the first SU).

[0144] Step 680: The AF network element initiates the first session's quality of service (QoS) modification process.

[0145] For example, the AF network element sends a QoS modification request to the NEF network element through the API gateway. The QoS modification request includes the QoS information modified for the first session. When migrating the AI ​​computing function of the first terminal and / or the first server to the SU of the access network accessed by the first terminal, the QoS of the first session between the first terminal and the first server may need to be modified accordingly. Therefore, in this embodiment of the application, it is necessary to initiate a QoS modification process for the first session.

[0146] like Figure 6As shown, the AF network element can send a QoS modification request for the first session to the gateway API through the service_API interface. This request includes not only the QoS information to be modified in the first session but also a first indication. The first indication can be called a network assistance indication (NW). Upon receiving the QoS modification request for the first session, the API gateway can obtain the first indication from it. Based on the first indication, the API gateway sends a third request to the SU in the first access network accessed by the first terminal, referring to the description of step 6010. Further, the gateway API sends the QoS modification request for the first session to the NEF network element. Upon receiving the QoS modification request for the first session, the NEF network element can initiate a QoS modification process for the first session. The QoS modification process for the first session involves the work of NEF network elements, PCF network elements, SMF network elements, AMF network elements, or UPF network elements, etc.

[0147] Step 690: QoS modification process for the first session.

[0148] Step 6010: The API network element sends a third request to the first SU.

[0149] For example, the third request can be called a SU assistance request. Upon receiving the third request, the SU executes the AI ​​computing assistance strategy. It is understood that the AF network element can notify the CU of the AI ​​computing assistance strategy determined in step 670. For example, in one possible implementation, the AF network element can send the AI ​​computing assistance strategy to the API gateway, and the API gateway can send the AI ​​computing assistance strategy to the first SU. For example, the QoS modification request for the first session sent by the AF network element to the API gateway in step 680 may also include information about the AI ​​computing assistance strategy. In step 6010, the third request sent by the API gateway to the first SU includes the AI ​​computing assistance strategy. Alternatively, in another possible implementation, the AF network element sends the AI ​​computing assistance strategy to the first terminal through the application layer, and the first terminal sends the AI ​​computing assistance strategy information to the SU, etc.

[0150] Understandably, in Figure 6 The process illustrates that the AF (Application Controller) network element sends a first instruction to the API gateway, and the API gateway sends a third request to the first SU (Application Unit). Alternatively, the AF can send a first instruction to the first terminal through the application layer, and upon receiving the first instruction, the first terminal sends a third request to the first SU to request the first SU to provide AI computing assistance for the service corresponding to the first session.

[0151] Step 6011: SU migrates the AI ​​computing functions of the first terminal and / or the first server through CU / DU.

[0152] For example, consider migrating the AI ​​computing function of the first terminal. Before step 6011, in the uplink transmission scenario: an application on the first terminal generates AI-related data, and the first terminal sends this AI-related data to the first server through the first session. The uplink path of the first session is: First Terminal → RU → DU → CU → UPF → First Server. After step 6010, since the SU undertakes the AI ​​computing function, when the CU receives the AI-related data from the first terminal, it can send the AI-related data to the SU, which then performs the corresponding AI calculation and sends it back to the CU. The CU then sends it to the first server via the UPF. In other words, the uplink path of the first session is now: First Terminal → RU → DU → CU → SU → CU → UPF → First Server. Similarly, the downlink path of the first session can be: First Server → UPF → CU → SU → CU → DU → RU → First Terminal.

[0153] against Figure 6 The flowchart shown is as follows: Figure 7 As shown in the illustration, this application also provides another form of schematic diagram, specifically: the access network includes RU, DU, CU, and SU, and the interface between SU ​​and CU is Si. SU reports its AI service capabilities to the API gateway through the SU_API interface. The AF network element sends a first request to the API gateway, and the API gateway obtains the ULI of the first terminal through the NEF network element based on the information of the first terminal included in the first request. The interface between the API gateway and the NEF network element is NEF_API. Based on the ULI of the first terminal, the API gateway determines the AI ​​service capabilities of the first SU of the access network accessed by the first terminal, and reports the AI ​​service capabilities of the first SU to the AF network element. Based on the AI ​​service capabilities of the first SU, the AF network element determines an AI computing assistance strategy, and sends the AI ​​computing assistance strategy to the first terminal through the application layer. The first terminal sends the AI ​​computing assistance strategy to the SU. The AF network element sends a first instruction to the API gateway, and the API gateway sends a third request to the SU to request the SU to provide AI computing assistance. Alternatively, the AF network element can send a first instruction to the terminal through the application layer, and the terminal can send a third request to the SU, etc., without limitation. Upon receiving a third request, the SU can perform AI-assisted computation. During this AI-assisted computation process, the data transmission path between the first terminal and the first server is: First Terminal—RU—DU—CU—SU—CU—UPF—First Server, which can be referenced. Figure 7The thick black line represents the AI ​​service of an application. Previously, AI computing was handled by the terminal and server, while the access network carried communication functions. In this embodiment, the AI ​​computing function is carried by the terminal, server, and SU (Supply Unit) in the access network, while CU (Complex Unit) and DU (Digital Unit) in the access network carry the original transmission or communication functions. Furthermore, the AF (Analog Function) network element can send a QoS modification request to the API gateway, which in turn sends a QoS modification request to the NEF (Neural Function) gateway, initiating the QoS modification process through the NEF network element.

[0154] Through the above design, the AI ​​service capabilities of the SU in the access network are opened to the AF network element. The AF network element migrates the AI ​​computing functions of the first terminal and / or the first server to the SU in the access network, thereby reducing the AI ​​computing load of the first terminal and / or the first server, improving the response speed of AI services, and enhancing the user's service experience for AI services. For example, if the AI ​​computing load of the first terminal and / or the first server is too high, the data of the application or service corresponding to the first session may need to be queued for AI computing processing. After migrating all or part of the AI ​​computing functions of the first terminal and / or the first server to the SU, the waiting latency of AI service data can be reduced, enabling AI service data to be processed in a timely manner, thereby improving the response speed and service experience of AI services.

[0155] Example 2

[0156] by Figure 5 The process plan is applied to Figure 3a The network open architecture shown is an example where the first network is the access network, the first network element is SU, and network element X is UPF. Figure 8 As shown, a flowchart is provided. In this process, when a UPF network element receives a first request from an AF network element, it determines the information of the anchor point or target access network corresponding to the first session based on the information of the first session included in the first request, and sends a request to the SU in the anchor point or target access network to request the AI ​​service capabilities of the SU, including:

[0157] Step 810: The first terminal establishes a first session with the first server.

[0158] Step 820: The AF network element sends a first request to the UPF network element. The first request is used to request the AI ​​service capability of the access network.

[0159] In one possible implementation, the AF network element sends the first request to the UPF network element through the API gateway. For example, the interface between the AF network element and the API gateway is the service_API interface, and the AF network element sends the first request to the API gateway through the service_API interface. The interface between the API gateway and the UPF network element is the UPF_API interface, and the API gateway sends the first request to the UPF network element through the UPF_API interface. Alternatively, the AF network element can directly send the first request to the UPF network element, etc., without restriction.

[0160] The first request includes information about the first session, and further, information about the first terminal. The UPF network element determines the first session based on this information. Furthermore, the UPF network element also needs to consider the first terminal's information when determining the first session. The UPF network element determines the anchor point or target access network in the first session, and further determines the SU (Supply Unit) in that anchor point or target access network, referred to as the first SU.

[0161] Step 830a: The UPF network element sends a second request to the first SU, which is used to request the AI ​​service capability of the first SU.

[0162] For example, a UPF network element can send a second request to a CU via the N3 interface, and the CU can send the second request to a first SU via the Si interface. This second request can be called a RAN capability query. When the first SU receives the second request, it can obtain the AI ​​service capabilities of the first SU.

[0163] Step 830b: The first SU sends a second response to the UPF network element, the second response including the AI ​​service capabilities of the first SU.

[0164] Similarly, SU sends a second response to CU through the Si interface, and CU sends a second response to the UPF network element through the N3 interface.

[0165] Step 840: The UPF network element sends a first response to the AF network element, which includes the AI ​​service capabilities of the first SU.

[0166] For example, the UPF network element sends its first response to the AF network element through the gateway API. Alternatively, the UPF network element sends its first response to the gateway API through the UPF_API interface, and the gateway API sends its first response to the AF network element through the service_API interface. Or, the UPF network element can directly send its first response to the AF network element without restriction.

[0167] Step 850: The AF network element determines the AI ​​computing assistance strategy based on the AI ​​service capabilities of the first SU.

[0168] When determining the AI ​​computing assistance strategy, the AF network element may need to adjust the QoS of the first session. Therefore, the AF network element can initiate a QoS modification request to modify the QoS of the first session.

[0169] Optionally, step 860: The AF network element initiates a QoS modification request.

[0170] For example, an AF network element sends a QoS modification request to a NEF network element, which includes the QoS information to be modified in the first session. Alternatively, the AF network element sends the QoS modification request to the NEF network element through an API gateway, and the NEF network element initiates the QoS modification process for the first session.

[0171] Optionally, step 870: QoS modification process for the first session.

[0172] exist Figure 8 In the process, the AF network element sends the first instruction and / or the AI ​​computing assistance strategy, including the following:

[0173] Case #1, in one possible implementation:

[0174] Step 880a: The AF network element sends a first indication to the UPF network element. The first indication is used to indicate that the service corresponding to the first session is provided with AI computing assistance by the access network.

[0175] For example, the AF network element can send the first indication to the UPF network element through the API gateway, or the AF network element can send the first indication directly to the UPF network element. It is understood that step 880a can also be combined with step 860. In step 860, the QoS modification request sent by the AF network element to the API gateway also includes the first indication. In this case, step 880a may not be necessary.

[0176] Step 890a: The UPF network element sends a third request to the first SU, which requests the first SU to provide AI computing assistance for the service corresponding to the first session.

[0177] Understandably, in this possible implementation: the AF network element sends a first instruction to the UPF network element, and the UPF network element sends a third request to the first SU. This possible implementation can also be used for: the AF network element sending an AI computing assistance strategy to the first SU. For example, the AF network element sends an AI computing assistance strategy to the UPF network element, and the UPF network element sends an AI computing assistance strategy to the first SU.

[0178] Case #2, in another possible implementation:

[0179] Step 880b: The AF network element sends a first instruction to the first terminal through the application layer.

[0180] Step 890b: The first terminal sends a third request to the first SU.

[0181] For example, when the CU receives a third request, it can send the third request to the first SU through the Si interface.

[0182] It is understandable that in this possible implementation: the AF network element sends a first instruction to the first terminal, and the first terminal sends a third request to the first SU. This possible implementation can also be used for: the AF network element sending an AI computing assistance strategy to the first SU. For example, the AF network element sends an AI computing assistance strategy to the first terminal, and the first terminal sends an AI computing assistance strategy to the first SU.

[0183] Step 8010: The first SU migrates the AI ​​computing functions of the first terminal and / or the first server through the CU / DU.

[0184] against Figure 8 The flowchart shown is as follows: Figure 9 As shown in the illustration, this application embodiment also provides another form of schematic diagram, including: an AF network element sending a first request to a UPF network element, the first request being used to query the AI ​​service capabilities of the first SU corresponding to the first session. The UPF network element sending a second request to the first SU through the CU of the access network, the second request being used to query the AI ​​service capabilities of the first SU, and sending the AI ​​service capabilities of the first SU to the UPF network element, the UPF network element sending the AI ​​service capabilities of the first SU to the AF network element. The AF network element determining an AI computing assistance strategy based on the AI ​​service capabilities of the first SU, and the AF network element sending the AI ​​computing assistance strategy to the first terminal through the application layer. Further, the AF network element can send a first instruction to the UPF network element through an API gateway or directly. The UPF network element sending a third request to the first SU through the CU, or the AF network element sending a first instruction to the first terminal through the application layer, and the first terminal sending a third request to the first SU. Figure 7 Similar, in Figure 9 In this context, the AI ​​computing function of an application or service is carried by the terminal, server, and SU in the access network, while the CU and DU in the access network carry the communication function.

[0185] Through the above design, the AI ​​service capabilities of the SU in the access network are opened to the AF network element. The AF network element migrates the AI ​​computing functions of the first terminal and / or the first server to the SU in the access network, thereby reducing the AI ​​computing load of the first terminal and / or the first server, improving the response speed of AI services, and enhancing the user's service experience for AI services.

[0186] Understandably, when Figure 6 or Figure 8 The process plan is applied to Figure 4When using the network open architecture shown, SU can be replaced with near real-time RIC, and the gateway API can be replaced with SMO.

[0187] Example 2

[0188] In the second embodiment, the access network can open up its idle AI computing resources to the AF network element. The AF network element can subscribe to the idle AI computing resources of the access network. The AF network element can send AI computing tasks to the access network through API gateway or SMO, etc. The first network element in the access network (e.g., SU or near real-time RIC, etc.) uses the AI ​​computing resources subscribed by the AF network element to execute the corresponding AI computing tasks. The access network provides services to the outside world with its idle AI computing resources, realizes the secondary utilization of idle AI computing resources, and improves the utilization rate of AI computing resources.

[0189] For example, an AF (Analog-First-Party) network element can send a first request to a second network element, the first request including information about a first AI computing resource that the AF network element requests to subscribe to. In response to the first request, the second network element can send a first response to the AF network element, the first response being a response to the first request, and the first response including transport network layer information of the network element providing the first AI computing resource. Optionally, before the AF network element sends the first request, the process further includes: the second network element sending a first report to the AF network element, the first report including information about a second AI computing resource that the network can provide, the second AI computing resource being all or part of the first AI computing resource. It is understood that the second network element is a gateway to the first network element providing the first AI computing resource in the network. For example, the first network element is a SU (Supply-Based Sockets) and the second network element is an API gateway. Alternatively, the second network element is a near-real-time RIC (Real-Time Integrated Circuit) and the first network element is an SMO (Single-Mobile Sockets).

[0190] Understandably, the above scheme can be applied to the access network, core network, or transmission network. That is, the second network element can report the second AI computing resources that the access network, core network, or transmission network can provide to the AF network element through a first report. The AF network element can send a first request to the second network element to subscribe to the first AI computing resources in the access network, core network, or transmission network, and provide the AF network element with the transmission network layer information of the network elements in the access network, core network, or transmission network that can provide the first AI computing resources.

[0191] In the description of Embodiment 2, the network used as the access network is mainly used as an example. It is understood that the solution in Embodiment 2 can be applied to... Figure 3a The network open architecture shown can be used as follows: the first network element can be a Subscriber Unit (SU); the "AI computing resources that the network can provide" mentioned above refers to the AI ​​computing resources that the SU in the access network can provide; the second network element can be a gateway API. Alternatively, it can be applied to... Figure 4In the network open architecture shown, the first network element can be a near real-time RIC, and the "AI computing resources that the network can provide" mentioned above refers to the AI ​​computing resources that the near real-time RIC in the access network can provide. The second network element can be an SMO.

[0192] The solution in Example 2 is applied to Figure 3a Taking the network open architecture shown as an example, such as Figure 10 As shown, a flowchart is provided:

[0193] Step 1010: One or more SUs send a capability report to the API gateway.

[0194] For example, SU sends a capability report to the API gateway through the SU_API interface, or SU can... Figure 3a Other interfaces shown, such as sending capability reports to the API gateway, are not restricted. This capability report includes information about the AI ​​computing resources provided by the SU. Furthermore, the capability report also includes information about the SU providing these AI computing resources.

[0195] Step 1020: The API gateway sends the first report to the AF network element.

[0196] For example, the API gateway can send a first report to the AF network element through the service_API interface. This first report can be called an AI network capability report. This first report includes information about the second AI computing resources provided by the SU. For example, the second AI computing resources provided by the SU could be: supporting model A, with computing specifications of 6 cores, 8GB RAM, 1Mbps bandwidth, and a duration of 1 year. It can be understood that the second AI computing resources provided by the SU can be its idle AI computing resources.

[0197] In one possible implementation, the API gateway can report information about its provided second AI computing resources to the AF network element as a whole, a process known as API gateway reporting aggregation network capability. For example, the API gateway can integrate AI computing resources reported by multiple SUs to form a network-wide AI computing resource pool, which is then reported to the AF network element. This process can be described as follows: the second AI computing resource is reported at the network-wide level.

[0198] In another possible implementation, the API gateway can report information about the second AI computing resources it provides to the AF network element at the granularity of SUs or SU sets. This process is called API gateway reporting discrete network capabilities. An SU set includes one or more SUs. It is understood that if the API gateway reports information about the second AI resources at the granularity of SUs or SU sets, then for each SU or SU set, the API gateway can send a first report separately. Further, this first report may include information about the corresponding SU or SU set. This process can be described as the second AI computing resources being reported at the granularity of SUs or SU sets in the network.

[0199] Step 1030: The AF network element sends a first request to the API gateway. The first request includes information about the first AI computing resource that the AF network element requests to subscribe to.

[0200] For example, an AF network element can send a first request to the API gateway through the service_API interface. This first request can be a service subscription request. It is understood that the first AI computing resource is all or part of the second AI computing resource. For example, in step 1020, the second AI computing resource reported by the API gateway could be: supporting model A, with computing specifications of 6 cores, 8GB RAM, 1Mbps bandwidth, and a duration of one year. Then, the AF network element can subscribe to a portion of this second AI computing resource. Specifically, the first AI computing resource subscribed by the AF network element is: model A, with computing specifications of 6 cores, 8GB RAM, 1Mbps bandwidth, and a duration of six months.

[0201] In one possible implementation, the AF network element can subscribe to the first AI computing resource at the network-wide level, meaning the first AI computing resource is subscribed to at the network-wide level. Alternatively, the AF network element can subscribe to the first AI computing resource at the unit level (SU or SU set), meaning the first AI computing resource is subscribed to at the SU or SU set level. In another possible implementation, if the AF network element subscribes to the first AI computing resource at the SU or SU set level, the AF network element can send multiple first requests to the API gateway, each first request requesting to subscribe to the first AI computing resource corresponding to the SU or SU set. Furthermore, the first request includes the identification information of the corresponding SU or SU set.

[0202] Step 1040: The API gateway sends a first response to the AF network element. The first response includes the transport network layer information of the network element in the access network that provides the first AI computing resources.

[0203] For example, the API gateway can send a first response to the AF network element through the service_API. The first response can be a service subscription response.

[0204] In one possible implementation, when an AF network element subscribes to a first AI computing resource as a whole, this first network element includes the transport network layer information of the API gateway, such as the API gateway's IP address or URL address. In this case, the API gateway is presented externally as a whole, and the AF network element can only obtain information about the API GW, but not about the access network or SU connected to that API GW.

[0205] In another possible implementation, when an AF network element subscribes to a first AI computing resource at the granularity of a SU or a set of SUs, the first response includes transport network layer information of the SU or set of SUs providing the first AI computing resource, such as the IP address or URL address of the SU or set of SUs.

[0206] Taking IP addresses as an example, the first response sent by the API gateway includes a list of IP addresses. This list includes at least one SU's IP address, which can be an IP address assigned to the SU by the API gateway. This prevents third-party application AF network elements from obtaining the SU's actual IP address, protecting the SU's privacy and security. It can be understood that there is a correspondence between the IP address assigned to the SU by the API gateway and the SU's actual IP address. The API gateway can store this correspondence. The IP address assigned to the SU by the API gateway can be called the SU's public network IP address, or the SU's IP address within the public network. The SU's actual IP address can be called the SU's internal network IP address, or the SU's IP address within the internal network. The API gateway can isolate the external public network from the internal network.

[0207] Understandably, when an AF network element subscribes to a first AI computing resource at the granularity of a SU or a set of SUs, the AF network element can send multiple first requests to the API gateway. Each first request requests the subscription to the first AI computing resource of the corresponding SU or set of SUs. Furthermore, the first request also includes the identification information of the corresponding SU or set of SUs. For each first request, the API network element can reply with a corresponding first response, which includes the transport network layer information of the SU or set of SUs requested to be subscribed to in the first request.

[0208] Step 1050: The AF network element sends the AI ​​computing task to the API gateway.

[0209] For example, when an AF network element subscribes to AI computing resources as a whole, the API network element, upon receiving an AI computing task, needs to decompose the AI ​​computing task and determine the SU (Suite) that executes each decomposed AI computing task. Alternatively, when an AF network element subscribes to AI computing resources at the granularity of SU or SU set, in step 1050, the AI ​​computing task sent by the AF network element to the API gateway may carry the transport network layer address of its corresponding SU or SU set. Taking IP address as an example, when the API network element receives the IP address of the AI ​​computing task and its corresponding SU, the API gateway determines the actual IP address of the SU corresponding to the IP address based on the IP address mapping relationship, and sends the corresponding AI computing task to the corresponding SU based on the actual IP address of the SU.

[0210] Optionally, step 1060: The AF network element decomposes the AI ​​computing task and determines the SU corresponding to each decomposed AI computing task.

[0211] Step 1070: The API gateway sends an AI computing task to the SU.

[0212] For example, a data path is established between the API gateway and the SU (System Controller) to transmit AI computing tasks. This path can be established via the Hypertext Transfer Protocol (HTTP). Through this established data path, the SU can download and execute AI computing tasks, and then feed the corresponding AI computing results back to the AF (Analog Controller) network element.

[0213] against Figure 10 The flowchart shown is as follows: Figure 11 As shown in the illustration, this application also provides another form of schematic diagram, including: The SUs of access network devices (e.g., gNB1 and gNB2) report AI service capabilities to the API gateway through the SU_API interface. The API gateway aggregates the AI ​​service capabilities of each SU and reports them to the AF network element. The AF network element subscribes to the network's AI service capabilities through the service_API interface, and the AF network element issues AI computing tasks to the API gateway. The API gateway issues the AI ​​computing tasks to the corresponding SUs, and the SUs execute the corresponding AI computing tasks according to the first AI computing resources subscribed to by the AF. In addition, during the execution of the AI ​​computing tasks, the first server and the SUs interact with AI data, which is then processed in… Figure 11 The term "data" is used here. It can be understood that during the process of an AF network element sending an AI computing task to a SU via the API gateway: the AF network element can send the configuration of the AI ​​model corresponding to the AI ​​computing task to the SU through the API gateway. Furthermore, it can also update the configuration of this AI model, etc.

[0214] Understandable, Figure 10 or Figure 11 The solution can also be applied to Figure 4 In the network open architecture shown. When applied to Figure 4 When the network open architecture is shown, Figure 10 or Figure 11 Replace the API gateway with SMO and SU with near real-time RIC.

[0215] Through the above design, the API gateway aggregates the AI ​​computing resources of SU and provides services to the outside world. The AF gateway subscribes to the AI ​​computing resources of SU through the service_API interface, realizing the secondary utilization of the idle AI computing resources of SU and improving the utilization rate of AI computing resources.

[0216] It is understood that the various embodiments of this application may be implemented individually or in combination, and there are no limitations.

[0217] In the embodiments provided above, the methods provided by this application have been described from the perspective of interaction between terminals, access network devices, core network devices, and servers. To implement the functions of the methods provided in the embodiments of this application, terminals, access network devices, core network devices, or servers may include hardware structures and / or software modules, implementing the above functions in the form of hardware structures, software modules, or a combination of hardware structures and software modules. Whether a particular function is executed in the form of hardware structures, software modules, or a combination of hardware structures and software modules depends on the design constraints of the specific application of the technical solution.

[0218] Figure 12 and Figure 13 This is a schematic diagram of the possible communication devices provided in the embodiments of this application. These communication devices can implement one or more corresponding functions in the above-described method embodiments. For example, functions implemented by an AF network element or network element X, thus potentially achieving the beneficial effects of the above-described method embodiments. In the embodiments of this application, the communication device can be an AF network element or network element X, or a unit, module, or component (such as a chip, chip system, circuit, or processor) applied in an AF network element or network element X. In the description of the embodiments of this application, the term "unit" is used as an example. For example, in the following description, the communication device includes a processing unit and a transceiver unit as an example. It is understood that the processing unit in the following description can also be replaced by a processing module or processing component. The transceiver unit can also be replaced by a transceiver module or transceiver component. For example, a transceiver component can refer to a communication module.

[0219] like Figure 12 As shown, the communication device 1200 includes a processing unit 1210 and a transceiver unit 1220.

[0220] Optionally, the transceiver unit 1220 may also be referred to as an output unit, an interface unit, a communication unit, etc. In one possible implementation, the transceiver unit 1220 includes at least one of a sending unit or a receiving unit. The sending unit and the receiving unit may be integrated together, or be two independent units, etc.

[0221] In one possible implementation, the communication device 1200 is used to implement the functions of the first embodiment above. For example, it implements the functions of the AF network element or the network element X in the first embodiment. Figure 5 、 Figure 6 或 Figure 8 . It can be understood that in Figure 6 , the network element X refers to the API gateway, and in Figure 8 , the network element X refers to the UPF network element.

[0222] When the communication device 1200 is used to implement the functions of the AF network element in Figure 5 、 Figure 6 或 Figure 8 , specifically: The processing unit 1210 is used to generate a first request; the transceiver unit 1220 is used to send the first request, and the first request includes information of a first terminal and / or information of a first session; the transceiver unit 1220 is further used to receive a first response, and the first response is a response to the first request, and the first response includes information on the artificial intelligence AI service capability of a first network, the first network being the network accessed by the first terminal, and the first session being a session for communication between the first terminal and a first server.

[0223] In one possible implementation, the processing unit 1210 is further used to determine an AI computing assistance strategy according to the information on the AI service capability of the first network, and the AI computing assistance strategy includes migrating all or part of the functions of the AI computing functions located in the first server and / or the first terminal to a first network element of the first network. The transceiver unit 1220 is further used to send the information on the AI computing assistance strategy.

[0224] In one possible implementation, the transceiver unit 1220 is further used to send a first indication, and the first indication is used to indicate that the service corresponding to the first session is provided with AI computing assistance by the network.

[0225] When the communication device 1200 is used to implement the network element X in Figure 5 , or the API gateway in Figure 6 , or Figure 8Specifically, the functions of the UPF network element are as follows: a transceiver unit 1220 is used to receive a first request, the first request including information of a first terminal and / or information of a first session, the first session being a communication session between the first terminal and a first server; a processing unit 1210 is used to generate a first response; the transceiver unit 1220 is also used to respond to the first request by sending a first response, the first response including the artificial intelligence (AI) service capabilities of a first network, the first network being the network accessed by the first terminal.

[0226] In one possible implementation, the processing unit 1210 is further configured to determine the AI ​​service capabilities of the first network based on the information of the first terminal and / or the information of the first session.

[0227] In one possible implementation, when the processing unit 1210 determines the AI ​​service capabilities of the first network based on the information of the first terminal, it includes: determining the location information of the first terminal based on the information of the first terminal; and determining the AI ​​service capabilities of the first network based on the location information of the first terminal and the set of AI service capabilities of the network, wherein the set of AI service capabilities of the network includes the AI ​​service capabilities of the first network.

[0228] In one possible implementation, when the processing unit 1210 determines the AI ​​service capabilities of the first network based on the information of the first session, it is specifically configured to: determine the first session based on the information of the first session; and send a second request to the first network element in the first network corresponding to the first session.

[0229] Receive a second response from the first network element, the second response being a response to the second request, and the second response including the AI ​​service capabilities of the first network element.

[0230] In one possible implementation, when the processing unit 12010 determines the first session based on the information of the first session, it is specifically configured to: determine the first session based on the information of the first session and the information of the first terminal.

[0231] In one possible implementation, the transceiver unit 1220 is further configured to receive a first indication, the first indication being configured to indicate that the service corresponding to the first session is provided with AI computing assistance by the network; the transceiver unit 1220 is further configured to send a third request, the third request being configured to request the first network to provide AI computing assistance to the service corresponding to the first session.

[0232] In one possible implementation, the transceiver unit 1220 is further configured to receive information about an AI computing assistance strategy, the AI ​​computing assistance strategy including migrating all or part of the AI ​​computing functions located in the first server and / or the first terminal to a first network element of the first network; the transceiver unit 1220 is further configured to send information about the AI ​​computing assistance strategy.

[0233] In another possible implementation, the communication device 1200 is used to implement the functions of Embodiment 2 described above. For example, implementing the functions in Embodiment 2... Figure 10 The functions of AF network elements or API gateways in the system.

[0234] When the communication device 1200 is used to implement the function of the second network element in the above embodiment 2, for example, the second network element may be Figure 10 The API gateway, or SMO, specifically includes: a transceiver unit 1220 for receiving a first request, the first request including information about a first artificial intelligence (AI) computing resource to be subscribed to; a processing unit 1210 for generating a first response; and the transceiver unit 1220 for responding to the first request by sending a first response, the first response including transmission network layer information of the network element providing the first AI computing resource in the network.

[0235] In one possible implementation, the first AI computing resource is subscribed at the granularity of the entire network, or at the granularity of the first network element or the first set of network elements in the network.

[0236] In one possible implementation, the first AI computing resource is subscribed at the network-wide level, and the transport network layer information of the network element included in the first response is the transport network layer information of the second network element, wherein the second network element is the gateway of the first network element that provides the first AI computing resource in the network.

[0237] In one possible implementation, the first AI computing resource is subscribed to at the granularity of a first network element or a set of first network elements in the network, and the transmission network layer information of the network element included in the first response is the transmission network layer information of the first network element or the set of first network elements.

[0238] In one possible implementation, the transceiver unit 1220 is further configured to send a first report, the first report including information on a second AI computing resource provided by the network.

[0239] In one possible implementation, the second AI computing resource is reported at the level of the entire network, or at the level of the first network element or the first set of network elements.

[0240] When the communication device 1200 is used to implement the function of the AF network element in the above embodiment 2, for example... Figure 10 The functions of the AF network element are as follows: Processing unit 1210 is used to generate a first request; transceiver unit 1220 is used to send the first request, the first request including information about a first artificial intelligence (AI) computing resource to be subscribed; transceiver unit 1220 is also used to receive a first response, the first response being a response to the first request, the first response including transmission network layer information of the network element providing the first AI computing resource in the network.

[0241] In one possible implementation, the first AI computing resource is subscribed at the granularity of the entire network, or at the granularity of the first network element or the first set of network elements in the network.

[0242] In one possible implementation, the first AI computing resource is subscribed at the network-wide level, and the transport network layer information of the network element included in the first response is the transport network layer information of the second network element, wherein the second network element is the gateway of the first network element that provides the first AI computing resource in the network.

[0243] In one possible implementation, the first AI computing resource is subscribed to at the granularity of a first network element or a set of first network elements in the network, and the transmission network layer information of the network element included in the first response is the transmission network layer information of the first network element or the set of first network elements.

[0244] In one possible implementation, the transceiver unit 1220 is further configured to receive a first report, the first report including information about a second AI computing resource provided by the network.

[0245] In one possible implementation, the second AI computing resource is reported at the level of the entire network, or at the level of the first network element or the first set of network elements.

[0246] For a more detailed description of the processing unit 1210 and the transceiver unit 1220, please refer to the descriptions of the first and second method embodiments above, which will not be repeated here.

[0247] It is understood that the division of units in this application embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods. In addition, the functional units in this application embodiment can be integrated into a physical device (e.g., in a processor), or each functional unit can be a separate physical device, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional module, etc.

[0248] like Figure 13 As shown, the communication device 1300 includes a processor 1310 and an interface circuit 1320. The processor 1310 and the interface circuit 1320 are coupled to each other. It is understood that the interface circuit 1320 can be a transceiver or an input / output interface. Optionally, the communication device 1300 may also include a memory 1330 for storing instructions executed by the processor 1310, or storing input data required by the processor 1310 to execute instructions, or storing data generated after the processor 1310 executes instructions.

[0249] When the communication device 1300 is used to implement the method shown in Embodiment 1 or Embodiment 2, the processor 1310 is used to implement the function of the processing unit 1210, and the interface circuit 1320 is used to implement the function of the transceiver unit 1220.

[0250] When the aforementioned communication device is a chip applied to an AF network element, the chip implements the functions of the AF network element in the above method embodiments. The chip receives information sent to the AF network element by the network element X through other modules (such as radio frequency modules or antennas) in the AF network element; or, the chip sends information to other modules (such as radio frequency modules or antennas) in the AF network element, which is information sent by the AF network element to the network element X.

[0251] When the aforementioned communication device is a module applied to network element X, the module implements the functions of network element X in the above method embodiments. This module receives information from other modules (such as radio frequency modules or antennas) within network element X, and this information is sent to network element X; alternatively, the module sends information to other modules (such as radio frequency modules or antennas) within network element X, and this information is sent by network element X to the AF network element. Here, the module of network element X can be an API gateway, SMO, or UPF network element, etc.

[0252] This application also provides a communication device, which includes a processor. The processor is used to implement the functions of the AF network element or network element X in Embodiment 1, or to implement the functions of the AF network element or second network element in Embodiment 2. Optionally, the communication device further includes a memory, and the processor is coupled to the memory.

[0253] This application also provides a communication device, including a processor and an interface circuit. The interface circuit is used to receive signals from other devices outside the device and transmit them to the processor, or to send signals from the processor to other devices outside the device. The processor is used to implement the functions of the AF network element or network element X in Embodiment 1, or to implement the functions of the AF network element or second network element in Embodiment 2, through logic circuits or execution code instructions.

[0254] This application also provides a computer-readable storage medium storing instructions, which may also be referred to as computer programs, computer program code, etc. When these instructions are executed on a computer, they cause the computer to perform the functions of the AF network element or network element X in the first embodiment of the above method, or to perform the functions of the AF network element or the second network element in the second embodiment of the above method.

[0255] This application also provides a computer program product, including a computer program or instructions, which, when run on a computer, implement the function of the AF network element or network element X in the first method embodiment above, or implement the function of the AF network element or second network element in the second method embodiment above.

[0256] This application also provides a chip or chip system, which includes a processor coupled to a memory. The processor executes computer programs or instructions stored in the memory to implement the functions of the AF network element or network element X in the first method embodiment, or to implement the functions of the AF network element or second network element in the second method embodiment. Optionally, the chip or chip system further includes a memory, and the processor is coupled to the memory.

[0257] It is understood that the processor in the embodiments of this application can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. A general-purpose processor can be a microprocessor or any conventional processor.

[0258] The memory in the embodiments of this application may be random access memory (RAM), flash memory, read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), register, hard disk, portable hard disk, CD-ROM, or any other form of storage medium known in the art.

[0259] The method steps in the embodiments of this application can be implemented in hardware or in software instructions executable by a processor. The software instructions can consist of corresponding software modules, which can be stored in random access memory, flash memory, read-only memory, programmable read-only memory, erasable programmable read-only memory, electrically erasable programmable read-only memory, registers, hard disks, portable hard disks, CD-ROMs, or any other form of storage medium well known in the art. An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. The storage medium can also be a component of the processor. The processor and the storage medium can reside in an ASIC.

[0260] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. This computer program product includes one or more computer programs or instructions. When the computer program or instructions are loaded and executed on a computer, the processes or functions described in the embodiments of this application are performed entirely or partially. The computer can be a general-purpose computer, a special-purpose computer, a computer network, a network device, a user equipment, or other programmable device. The computer program or instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another. For example, the computer program or instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center integrating one or more available media. The available medium can be a magnetic medium, such as a floppy disk, hard disk, or magnetic tape; it can also be an optical medium, such as a digital video optical disc; or it can be a semiconductor medium, such as a solid-state drive. The computer-readable storage medium may be a volatile or non-volatile storage medium, or may include both types of storage media.

[0261] In the various embodiments of this application, unless otherwise specified or in case of logical conflict, the terminology and / or descriptions of different embodiments are consistent and can be referenced by each other. The technical features of different embodiments can be combined to form new embodiments according to their inherent logical relationship.

Claims

1. A communication method, characterized in that, include: Send a first request, the first request including information about the first terminal and / or information about the first session; Receive a first response, which is a response to the first request. The first response includes information about the artificial intelligence (AI) service capabilities of the first network. The first network is the network accessed by the first terminal. The first session is a communication session between the first terminal and the first server.

2. The method as described in claim 1, characterized in that, Also includes: Based on the information on the AI ​​service capabilities of the first network, an AI computing assistance strategy is determined. The AI ​​computing assistance strategy includes migrating all or part of the AI ​​computing functions located in the first server and / or the first terminal to the first network element of the first network. Send information about the AI ​​computing assistance strategy.

3. The method as described in claim 1 or 2, characterized in that, Also includes: Send a first instruction, which is used to indicate that the service corresponding to the first session is provided with AI computing assistance by the network.

4. A communication method, characterized in that, include: Receive a first request, the first request including information about a first terminal and / or information about a first session, the first session being a session used for communication between the first terminal and a first server; In response to the first request, a first response is sent, the first response including the artificial intelligence (AI) service capabilities of a first network, the first network being the network accessed by the first terminal.

5. The method as described in claim 4, characterized in that, Also includes: The AI ​​service capabilities of the first network are determined based on the information from the first terminal and / or the information from the first session.

6. The method as described in claim 5, characterized in that, Determining the AI ​​service capabilities of the first network based on the information from the first terminal includes: Based on the information from the first terminal, determine the location information of the first terminal; Based on the location information of the first terminal and the set of AI service capabilities of the network, the AI ​​service capabilities of the first network are determined, wherein the set of AI service capabilities of the network includes the AI ​​service capabilities of the first network.

7. The method as described in claim 5, characterized in that, Determining the AI ​​service capabilities of the first network based on the information from the first session includes: The first session is determined based on the information from the first session; Send a second request to the first network element in the first network corresponding to the first session; Receive a second response from the first network element, the second response being a response to the second request, and the second response including the AI ​​service capabilities of the first network element.

8. The method as described in claim 7, characterized in that, Determining the first session based on the information from the first session includes: The first session is determined based on the information from the first session and the information from the first terminal.

9. The method according to any one of claims 4 to 8, characterized in that, Also includes: Receive a first instruction, the first instruction being used to indicate that the service corresponding to the first session is provided with AI computing assistance by the network; Send a third request, which is used to request the first network to provide AI computing assistance for the service corresponding to the first session.

10. The method according to any one of claims 4 to 9, characterized in that, Also includes: Receive information about an AI computing assistance strategy, wherein the AI ​​computing assistance strategy includes migrating all or part of the AI ​​computing functions located in the first server and / or the first terminal to the first network element of the first network; Send information about the AI ​​computing assistance strategy.

11. A communication method, characterized in that, include: Receive a first request, the first request including information about a first artificial intelligence (AI) computing resource to be subscribed to; In response to the first request, a first response is sent, the first response including transport network layer information of the network element in the network that provides the first AI computing resources.

12. The method as described in claim 11, characterized in that, The first AI computing resource is subscribed at the level of the entire network, or at the level of the first network element or the first set of network elements.

13. The method as described in claim 12, characterized in that, The first AI computing resource is subscribed at the network-wide level, and the transport network layer information of the network element included in the first response is the transport network layer information of the second network element. The second network element is the gateway of the first network element that provides the first AI computing resource in the network.

14. The method as described in claim 12, characterized in that, The first AI computing resource is subscribed to at the granularity of the first network element or the first set of network elements in the network, and the transmission network layer information of the network element included in the first response is the transmission network layer information of the first network element or the first set of network elements.

15. The method according to any one of claims 11 to 14, characterized in that, Before receiving the first request, the method further includes: Send a first report, which includes information about the second AI computing resources provided by the network.

16. The method as described in claim 15, characterized in that, The second AI computing resource is reported at the level of the entire network, or at the level of the first network element or the first set of network elements.

17. A communication method, characterized in that, include: Send a first request, the first request including information about a first artificial intelligence (AI) computing resource to be subscribed to; Receive a first response, which is a response to the first request, and the first response includes transport network layer information of the network element that provides the first AI computing resources in the network.

18. The method as described in claim 17, characterized in that, The first AI computing resource is subscribed at the level of the entire network, or at the level of the first network element or the first set of network elements.

19. The method as described in claim 18, characterized in that, The first AI computing resource is subscribed at the network-wide level, and the transport network layer information of the network element included in the first response is the transport network layer information of the second network element. The second network element is the gateway of the first network element that provides the first AI computing resource in the network.

20. The method as described in claim 18, characterized in that, The first AI computing resource is subscribed to at the granularity of the first network element or the first set of network elements in the network, and the transmission network layer information of the network element included in the first response is the transmission network layer information of the first network element or the first set of network elements.

21. The method according to any one of claims 17 to 20, characterized in that, Before sending the first request, the method also includes: Receive a first report, which includes information about a second AI computing resource provided by the network.

22. The method as described in claim 21, characterized in that, The second AI computing resource is reported at the level of the entire network, or at the level of the first network element or the first set of network elements.

23. A communication device, characterized in that, It includes units for implementing the method as described in any one of claims 1 to 3, or units for implementing the method as described in any one of claims 4 to 10, or units for implementing the method as described in any one of claims 11 to 16, or units for implementing the method as described in any one of claims 17 to 22.

24. A communication device, characterized in that, The device includes a processor configured to cause the communication device to perform the method as described in any one of claims 1 to 3, or the method as described in any one of claims 4 to 10, or the method as described in any one of claims 11 to 16, or the method as described in any one of claims 17 to 22.

25. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed, cause the communication device to perform the method as described in any one of claims 1 to 3, or the method as described in any one of claims 4 to 10, or the method as described in any one of claims 11 to 16, or the method as described in any one of claims 17 to 22.

26. A computer program product, characterized in that, The computer program product includes instructions that, when executed, cause the communication device to perform the method as described in any one of claims 1 to 3, or the method as described in any one of claims 4 to 10, or the method as described in any one of claims 11 to 16, or the method as described in any one of claims 17 to 22.

Images