Communication system and method, network function, electronic device, storage medium, and product

Abstract

The present disclosure relates to the technical field of communications, and in particular provides a communication system and method, a network function, an electronic device, a storage medium, and a product. The communication system of the present disclosure comprises first network functions and second network functions. The first network functions are used to implement communication connection functions. The second network functions are communicatively connected to at least one of the first network functions, and are used to implement at least one of the following first functions: a data management function, a data analysis function, an artificial intelligence / machine learning (AI / ML) model-related function, a computing function, and a resource collaborative control function. The technical solution provided in the present disclosure can improve system flexibility and scalability by means of a hierarchical network architecture, and provide system support for multi-dimensional service capabilities such as computing, intelligence, and perception.

Description

Communication systems and methods, network functions, electronic devices, storage media and products

[0001] Cross-reference to related applications

[0002] This application is based on and claims priority to Chinese Patent Application No. 202411182481.2, filed on August 27, 2024, the entire contents of which are incorporated herein by reference. Technical Field

[0003] This disclosure relates to the field of communication technology, and in particular to a communication system and method, network function, electronic device, storage medium and product. Background Technology

[0004] The International Telecommunication Union (ITU) has established a vision and framework for IMT-2030, with computing, intelligence, and sensing capabilities beyond communications being key features of 6G. However, the existing 5G network's centralized unit (CU) and distributed unit (DU) architectures are relatively rigid. Introducing new functions and features generally requires an upgrade of the entire base station's software capabilities. This approach neither supports flexible and dynamic loading of new capabilities on demand nor network programmability, thus failing to support multi-dimensional service capabilities such as computing, intelligence, and sensing. Summary of the Invention

[0005] This disclosure provides a communication system and method, network function, electronic device, storage medium and product to enhance system flexibility and scalability, and to provide system support for multi-dimensional service capabilities such as computing, intelligence and sensing.

[0006] According to one aspect of this disclosure, a communication system is provided, comprising:

[0007] The first network function is used to implement communication connectivity.

[0008] The second network function is communicatively connected to at least one of the first network functions and is used to implement at least one of the following first functions: data management function, data analysis function, artificial intelligence / machine learning (AI / ML) model-related functions, computing function, and resource collaborative control function.

[0009] According to another aspect of this disclosure, a communication method is provided for application to a second network function, the method comprising:

[0010] Establish a communication connection with the first network function;

[0011] Interact with the first network function and / or the core network; wherein the second network function has at least one of the following first functions: data management function, data analysis function, artificial intelligence / machine learning (AI / ML) model related functions, computing function, and resource collaborative control function.

[0012] According to another aspect of this disclosure, another communication method is provided for application to a first network function, the method comprising:

[0013] Establish a communication connection with the second network function;

[0014] The second network function interacts with the communication connection; wherein the second network function has at least one of the following first functions: data management function, data analysis function, artificial intelligence / machine learning (AI / ML) model related functions, computing function, and resource collaborative control function.

[0015] According to another aspect of this disclosure, a network function is provided, including:

[0016] Establish a unit, configured to establish a communication connection with the first network function;

[0017] The communication unit is configured to interact with the first network function and / or the core network; wherein the second network function has at least one of the following first functions: data management function, data analysis function, artificial intelligence / machine learning (AI / ML) model-related functions, computing function, and resource collaborative control function.

[0018] According to another aspect of this disclosure, another network function is provided, including:

[0019] Establish a unit, configured to establish a communication connection with the second network function;

[0020] The communication unit is configured to interact with the second network function based on the communication connection; wherein the second network function has at least one of the following first functions: data management function, data analysis function, artificial intelligence / machine learning (AI / ML) model-related functions, computing function, and resource collaborative control function.

[0021] According to another aspect of this disclosure, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the method described in any of the above embodiments.

[0022] According to another aspect of this disclosure, a computer-readable storage medium is provided that stores a computer program / instructions thereon, which, when executed by a processor, implement the methods described in any of the above embodiments.

[0023] According to another aspect of this disclosure, a computer program product is provided, including a computer program / instructions that, when executed by a processor, implement the methods described in any of the above embodiments.

[0024] As will be described in detail below, a communication system and method, network function, electronic device, storage medium, and product are disclosed according to embodiments of the present disclosure. Compared to the fixed CU / DU architecture in related technologies, the communication system provided by the present disclosure presents a two-level system architecture. A first network function in the communication system enables basic communication connectivity, while a second network function enables the first function, which may include, but is not limited to, at least one of the following: data management function, data analysis function, AI / ML model-related functions, computing function, and resource collaborative control function. Thus, on the one hand, computing and intelligent capabilities can be flexibly introduced. Compared to the traditional communication connectivity functions of the first network function, these connectivity-independent first functions can be designed based on a microservice architecture, offering better flexibility and scalability. On the other hand, the relatively centralized deployment of computing, sensing, and other first functions in the second network function provides stronger support for the distributed first network function in terms of computing power, data capabilities, and control capabilities, and achieves better resource pooling utilization efficiency through centralization, thereby improving network resource utilization efficiency. Furthermore, with future technological advancements, the primary function of the second network function can be further improved or added to without requiring extensive modifications to the entire communication system. This communication system can effectively meet the flexibility and scalability requirements of future technologies. In summary, the technical solution provided in this disclosure enhances system flexibility and scalability through a hierarchical network architecture, providing system support for multi-dimensional service capabilities such as computing, intelligence, and sensing. Attached Figure Description

[0025] Figure 1 is a schematic diagram of the architecture of a communication system provided in an embodiment of this disclosure.

[0026] Figure 2 is a schematic diagram of the architecture of another communication system provided in an embodiment of this disclosure.

[0027] Figure 3 is a flowchart illustrating a communication method provided in an embodiment of this disclosure.

[0028] Figure 4 is a flowchart illustrating another communication method provided in an embodiment of this disclosure.

[0029] Figure 5 is a schematic diagram of the interaction flow of a communication method provided in an embodiment of this disclosure.

[0030] Figure 6 is a schematic diagram of the interaction flow of another communication method provided in an embodiment of this disclosure.

[0031] Figure 7 is a schematic diagram of the interaction flow of another communication method provided in an embodiment of this disclosure.

[0032] Figure 8 is a schematic diagram of the interaction flow of another communication method provided in an embodiment of this disclosure.

[0033] Figure 9 is a schematic diagram of the interaction flow of another communication method provided in an embodiment of this disclosure.

[0034] Figure 10 is a schematic diagram of the interaction flow of another communication method provided in an embodiment of this disclosure.

[0035] Figure 11 is a schematic diagram of the interaction flow of another communication method provided in an embodiment of this disclosure.

[0036] Figure 12 is a hardware block diagram of an electronic device provided in an embodiment of this disclosure.

[0037] Figure 13 is a schematic diagram of a computer program product provided in an embodiment of this disclosure. Detailed Implementation

[0038] Currently, the 5G Radio Access Network (RAN) architecture defined in 3GPP includes Centralized Units (CUs) and Distributed Units (DUs) to support flexible RAN function deployment. The CU mainly includes protocol layers such as Service Data Adaptation Protocol (SDAP), Radio Resource Control (RRC), and Packet Data Convergence Protocol (PDCP), while the DU mainly includes protocol layers such as Radio Link Control (RLC), Medium Access Control (MAC), and Physical Layer (PHY). The CU user plane (CU-UP) and CU control plane (CU-CP) are decoupled, allowing for flexible deployment of both CU-UP and CU-CP. With the development of AI technology, RAN is also considering leveraging AI to further enhance network performance and user experience. 3GPP is currently considering supporting AI / ML model training and inference functions through NG-RAN nodes. The O-RAN Alliance introduced Non-Real Time Radio Intelligent Controllers (Non-RT RICs) and Near-Real Time Radio Intelligent Controllers (Near-RT RICs). Non-RT RICs are primarily aimed at network management and operation optimization, while Near-RT RICs are primarily aimed at network element intelligence.

[0039] This means that the existing CU / DU architecture of 5G networks is relatively rigid. The introduction of new functions and features generally requires an upgrade of the entire base station software capabilities. It does not support flexible and dynamic loading of new capabilities on demand, nor does it support network programmable features, and it cannot support multi-dimensional service capabilities such as computing, intelligence and sensing.

[0040] To address this issue, this disclosure provides a novel design concept: a two-layer deployment architecture for network functions in a communication system used to implement communication connectivity. The first layer can deploy one or more network functions to implement traditional communication connectivity, while the second layer can communicate with the first layer to implement a first function (in some scenarios, this first function can also be called an extended function) suitable for multi-dimensional service capabilities such as computing, intelligence, and sensing. Thus, through the cooperation between the two layers, both communication connectivity requirements are met, and support for current or future multi-dimensional service capabilities is provided, optimizing the system's flexibility and scalability.

[0041] In view of this, this disclosure provides a communication system and method, network function, electronic device, storage medium and product to enhance system flexibility and scalability, and to provide system support for multi-dimensional service capabilities such as computing, intelligence and sensing. The details are described below.

[0042] First, this disclosure provides a communication system. Please refer to Figure 1, which is a schematic diagram of the architecture of a communication system provided in an embodiment of this disclosure. As shown in Figure 1, the communication system 100 includes at least:

[0043] The first network function 110 is used (can be understood as configured to) implement communication connection functions;

[0044] The second network function 120 is communicatively connected to at least one first network function 110 and is used (can be understood as configured to) implement at least one of the following first functions: data management function, data analysis function, artificial intelligence / machine learning AI / ML model related functions, computing function, and resource collaborative control function.

[0045] In this disclosure, "used for" can be understood as "configured as".

[0046] In this disclosure, the first network function 110 is used to implement a communication connection function, which is equivalent to a traditional communication connection function, and can also be implemented by using existing network functions (such as CU, DU) in existing communication systems. It should be understood that the first network function 110 has at least a communication connection function, and in actual scenarios or with the development of technology, the first network function 110 may also have other capabilities, and this disclosure does not have any particular limitations in this regard.

[0047] Specifically, the communication connectivity function refers to the ability of a communication system to communicate with other systems or devices. In one exemplary embodiment, the communication system 100 provided in this disclosure can be specifically a wireless access network system. In this scenario, the first network function 110 is used to implement the communication connectivity function for the core network. In some embodiments, the first network function 110 is mainly used to implement the control plane function and user plane function of the communication connectivity. In other words, the first network function 110 can communicate with the user plane function and control plane function of the core network respectively, wherein the control plane function can connect to the core network in a service-oriented or non-service-oriented manner. In addition, the first network function 110 can also connect and communicate with terminals to assist in realizing data interaction between terminals and the core network.

[0048] In this disclosure, the second network function 120 is used to implement the first function (also referred to as the extended function or the first function). The first function referred to here is at least an open capability other than the traditional communication connection function that the first network function 110 can implement. In other words, the second network function 120 in this disclosure is mainly used to enhance the communication system's open capabilities beyond communication connections. This disclosure does not impose any particular limitation on the first function that the second network function 120 can implement. For example, it can involve at least one of the following: data-related functions, computing-related functions, AI / ML model-related functions, and wireless capability open functions; these are not exhaustive. Furthermore, with the development of technology, the second network function 120 in this disclosure can also carry more first functions to better support the communication system's multi-dimensional service capabilities in computing, intelligence, and sensing. Accordingly, in this disclosure, the second network function 120 adopts a microservice architecture, enabling the second network function 120 to have more flexible first functions and to implement more types of first functions.

[0049] In one exemplary embodiment, the first function involved in the second network function 120 in this disclosure includes, but is not limited to, at least one of the following: data management function, data analysis function, artificial intelligence / machine learning AI / ML model related function, computing function, and resource collaborative control function.

[0050] Among them, the data management function, which can also be called the wireless data management function in the wireless access network scenario, refers to the ability to manage data. Its management methods may include, but are not limited to, at least one of the following: data acquisition, data preprocessing, data storage, data distribution, etc., without exhaustive list.

[0051] Data analysis functionality, also known as wireless data analysis functionality in wireless access network scenarios, refers to the ability to analyze and predict services, data, and systems based on data. For example, data analysis functionality may include, but is not limited to, at least one of the following: service identification, service characteristic perception analysis, wireless data statistical analysis, and wireless data prediction, etc., without exhaustive list. Wireless data prediction may include, but is not limited to: statistical analysis and prediction of at least one of the following at the network level / user level / slice level / service level (e.g., identifiable by 5QI or QCI): link coverage quality, load, network and user rate / latency / reliability, user KQI / QoE experience, network and terminal power consumption, etc.

[0052] AI / ML model-related functions refer to functions such as model training and model application using AI / ML technologies. For example, AI / ML model-related functions may include, but are not limited to, at least one of the following: model training, model inference, model deployment, model storage, model compression, model optimization, and model application, etc., without exhaustive list.

[0053] Computing functionality refers to the ability to perform computing tasks. For example, computing functionality may include, but is not limited to, at least one of the following: sensing of terminal and / or base station computing resources; control and / or scheduling of wireless computing resources, etc.; this is not an exhaustive list.

[0054] Resource coordination control function refers to the ability to coordinate control over communication and / or computing resources. Specifically, in addition to coordinating control over its own communication and / or computing resources, the second network function 120 can also coordinate control over the communication and / or computing resources of the first network function 110 connected to it. For example, the second network function 120 can also be used to assist the first network function 110 in intelligent optimization and enhancement of wireless resource management. In one exemplary embodiment, the second network function 120 can send at least one of the following information to the first network function 110 through the first interface Infx: wireless data analysis results, network optimization suggestions, control optimization strategies and control configurations, AI / ML models, etc., without exhaustive list.

[0055] It should be understood that the first function described above is merely an illustrative example, and in actual scenarios, the second network function 120 may have more of the first functions.

[0056] This disclosure does not impose any particular limitation on the number of first network functions 110 and second network functions 120 in the communication system 100, wherein there is at least one first network function 110 and at least one second network function 120. In some embodiments, the second network function 120 is communicatively connected to at least one first network function 110. For ease of illustration, FIG1 shows three first network functions 110 and one second network function 120, and all three first network functions 110 are communicatively connected to the second network function 120. Furthermore, it should be noted that the first network function 110 and / or the second network function 120 can be a single network function or a combination of multiple network functions. For example, the communication connection function of a first network function 110 can be implemented by a single CU, or it can be implemented by multiple CUs working together, or it can be implemented by a CU and a DU working together, etc. The following describes in detail its possible implementation methods.

[0057] In this disclosure, the first network function 110 and the second network function 120 can be connected via a first interface, Infx. Furthermore, in the wireless communication network system, the first network function 110 communicates with the user plane and control plane functions of the core network, respectively; while the second network function 120 adopts a microservice architecture and connects to the core network through a service-oriented interface. Thus, the interaction between the second network function 120 and the core network has advantages such as convenient invocation, good reusability, and good scalability. Regardless of the first function currently supported or potentially supported by the second network function 120 in the future, it can effectively communicate with the core network through the service-oriented architecture, avoiding any adverse impact on the core network from the expansion or compression of the first function, and improving the flexibility, scalability, and stability of the entire network-side system (core network and wireless access network).

[0058] It should be noted that the second network function 120 is not directly connected to the terminal, but can achieve indirect communication with the terminal through the first network function 110 and / or the core network. In some embodiments, the second network function 120 can also achieve collaborative management of (at least one) terminal. This will be explained later in conjunction with the specific implementation of the first function.

[0059] In summary, the second network function 120 in this disclosure can achieve collaborative management of at least one terminal (via the first network function 110 and / or the core network), and / or at least one first network function 110. Taking Figure 1 as an example, the communication system 100 shown in Figure 1 illustrates a second network function 120, which is communicatively connected to three first network functions 110. This allows for individual management of these three first network functions 110 and collaborative management of at least two of them. It also enables indirect management of one or more terminals through these first network functions 110 and / or the core network. Thus, the first network function 110 and the second network function 120 constitute a two-layer communication connection architecture. That is, the first layer implements traditional communication connection functions, while the second layer can communicate with the first layer to cooperate in implementing first functions applicable to multi-dimensional service capabilities such as computing, intelligence, and sensing.

[0060] In summary, compared to the fixed CU / DU architecture in related technologies, the communication system provided in this disclosure presents a two-tiered system architecture. The first network function in the communication system enables basic communication connectivity, while the second network function enables at least one of the following first functions: data management, data analysis, AI / ML model-related functions, computing, and resource collaborative control. This allows for the flexible introduction of computing and intelligent capabilities. Compared to the traditional communication connectivity provided by the first network function, these connectivity-based first functions can be designed based on a microservice architecture, offering better flexibility and scalability. Furthermore, the relatively centralized deployment of computing and sensing functions in the second network function provides stronger support for the distributed first network function in terms of computing, data, and control capabilities. Centralization also enables better resource pooling and utilization, improving network resource efficiency. Moreover, with future technological advancements, the first functions of the second network function can be further improved or added to without requiring extensive modifications to the entire communication system. This communication system effectively meets the flexibility and scalability requirements of future technologies. In summary, the technical solution provided in this disclosure can enhance system flexibility and scalability through a hierarchical network architecture, providing system support for multi-dimensional service capabilities such as computing, intelligence, and sensing.

[0061] In one exemplary embodiment, please refer to FIG2, which is a schematic diagram of the architecture of another communication system provided in this disclosure. As shown in FIG2, the communication system provided in this disclosure can specifically be a wireless access network 210. The wireless access network 210 can be connected to the core network 220 to form a complete communication system 200. This disclosure does not impose any particular limitation on the type of communication system 200, which may include, but is not limited to: Global System for Mobile communication (GSM) system, Code Division Multiple Access (CDMA) system, Wideband Code Division Multiple Access (WCDMA) system, General Packet Radio Service (GPRS), Long Term Evolution (LTE) system, Advanced Long Term Evolution (LTE-A) system, New Radio (NR) system, evolution of NR system, LTE-based access to unlicensed spectrum (LTE-U) system, NR-based access to unlicensed spectrum (NR-U) system, Non-Terrestrial Networks (NTN) system, Universal Mobile Telecommunication System (UMTS), Wireless Local Area Networks (WLAN), and Wireless Fidelity (WF). The communication system 100 provided in this disclosure can be a wireless access network portion of a communication system adapted to various standards or specifications, including Fidelity (WiFi), 5th-Generation (5G) systems, and other communication systems, without exhaustive list.

[0062] The communication system 200 can provide network services to the terminal. The terminal, also known as a terminal device, user equipment (UE), mobile station, mobile terminal, etc., is capable of communicating with network-side devices. This disclosure does not impose any particular limitations on the applicable scenarios and capabilities of the terminal. For example, the terminal 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. Based on this, the terminal can be, but is not limited to: 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 impose any particular limitations on the specific technologies and device forms used in the terminal.

[0063] For example, the terminal has at least a communication connection function, that is, it has at least the ability to communicate with network-side devices (e.g., the first network function, the core network). In some embodiments, in one possible implementation, the terminal device may also include, but is not limited to, at least one of the following functions: computing function, data function (e.g., acquisition, processing, storage, and analysis). The computing function can support online training and inference of the terminal's AI / ML models, and can also support service mesh proxy functions to achieve efficient collaboration with network computing functions (e.g., the computing functions of the wireless access network, the core network computing functions) and the computing functions of the edge cloud and the central cloud. The terminal's data function can interact with the data functions of the base station and / or the data functions of the core network to establish a logical channel for data transmission, processing, analysis, and storage, for efficient data transmission and collaborative processing among multiple devices such as the terminal, base station, and core network. As shown in Figure 2, Figure 2 specifically illustrates two terminals, both of which have communication connection functions, computing functions, and data functions. It should be understood that the number and capabilities of terminals in actual scenarios can be more or less, which will not be elaborated further.

[0064] In one exemplary embodiment, the first network function in this disclosure may also include, but is not limited to, at least one of the following: a communication connection function-related logical network function or node (xNB connection serving unit, xNB-CSU). And, the second network function may also include, but is not limited to, at least one of the following: a transcendental connection logical network function or node (xNB intelligent unit, xNB-IU).

[0065] Please refer to Figure 2. As shown in Figure 2, the xNB-IU is connected to two xNB-CSUs via the first interface Infx. Each xNB-CSU communicates with the control plane and user plane functions of the core network, respectively; while the xNB-IU connects to the core network via a service-oriented approach, as shown in Figure 2 as a bus connecting to the core network. In this way, the xNB-IU can communicate and interact with at least one of the core network's control plane, user plane, compute plane, data plane, and security plane functions.

[0066] In some embodiments, the first network function 110 is a communication connectivity function-related logical network function or node (xNB-CSU), and as mentioned above, the first network function 110, i.e., the xNB-CSU, can itself be implemented by one or more network functions. Therefore, the xNB-CSU can also have various possible implementations. These are described in detail below.

[0067] Method 1: All functions of the communication connection are constituted by a single network function. That is, a first network function (i.e., xNB-CSU) is constituted by a single network function. For example, an xNB-CSU is constituted by a single CU.

[0068] Method 2: A first network function (i.e., xNB-CSU) is implemented by a communication connection user plane network function and a communication connection control plane network function. The user plane network function is used for data transmission related to the communication user plane; the communication connection control plane network function is used for signaling transmission of the communication control plane. In other words, a first network function (i.e., xNB-CSU) consists of two network functions: a combination of a control plane network function and a user plane network function.

[0069] Method 3: A first network function (i.e., xNB-CSU) is implemented by a centralized network function (CU) and a distributed network function (DU) connected by communication. The CU can be further composed of user plane network functions and control plane network functions. In other words, a first network function (i.e., xNB-CSU) is composed of three network functions.

[0070] Method 4: A first network function (i.e., xNB-CSU) is composed of a centralized network function (CU) connected to the communication network, a distributed network function (DU) connected to the communication network, and a radio unit (RU) connected to the communication network. The capabilities supported by the DU may include, but are not limited to, at least one of the following: RLC, MAC, and partial physical layer processing; the partial physical layer processing may include, but are not limited to, at least one of the following: interleaving, modulation, layer mapping, precoding (optional), resource mapping, and I / Q data compression (optional). The capabilities supported by the RU may include, but are not limited to, at least one of the following: partial inverse fast Fourier transform (IFFT), CP addition, and analog-to-digital conversion; furthermore, in some embodiments, the optional capabilities of the RU may also include, but are not limited to, at least one of the following: I / Q data decompression, precoding, digital beamforming, and analog beamforming. For example, Figure 2 illustrates one implementation of xNB-CSU using one of the first network functions (xNB-CSU) as an example, where an xNB-CSU is composed of a combination of three network functions: CU, DU, and RU. Furthermore, it should be understood that a CU can also be implemented by one or more network functions. For example, in one possible embodiment, a CU can be implemented by combining two parts: a user plane network function and a control plane network function. It should be understood that although the names used here are all network functions, the way network functions are divided and the granularity are different. A xNB-CSU in this disclosure can reuse one or more network functions with different division methods and granularities in related technologies, which will not be elaborated further.

[0071] As shown in Figure 1 or Figure 2, the number of first network functions in this disclosure can be multiple, and in this case, the multiple first network functions can be deployed in a distributed manner. In some embodiments, in a preferred embodiment of this disclosure, the first network function may also include, but is not limited to, at least one second function (also referred to as a distributed function or distributed capability): data acquisition and processing function, and AI / ML model inference function. Thus, the first network function can further support distributed intelligence of wireless network elements based on the aforementioned distributed capabilities, and in some embodiments, the aforementioned distributed capabilities can interact with data-related functions and AI / ML model-related functions in the xNB-IU to achieve multi-node distributed intelligent collaboration.

[0072] In this disclosure, the second network function can support at least one first function, and in the specific implementation of these first functions, any one of the first functions can be implemented by one or more network functions.

[0073] In one exemplary embodiment, in this disclosure, any of the first functions of the second network function may include, but is not limited to, at least one of the following: control plane function, data plane function, and computation plane function.

[0074] The control plane function is used for control and management. The objects of control and management may include, but are not limited to, at least one of the following: data, models, resources, tasks, terminals, and network functions.

[0075] The data plane function is used for at least one of data processing, data acquisition, data transmission, and data storage. For example, the data plane (processing) function for data and models.

[0076] The computing plane functionality is used for at least one of computing task processing, computing resource awareness, computing resource management, or scheduling. That is, it can be configured to implement at least one of computing task processing, computing resource awareness, and computing resource control and scheduling.

[0077] For example, the AI / ML model-related functions in the first function supported by xNB-IU can be further divided into: model control functions and model processing functions (or model data surface functions). The model processing function can include one or more of the following: model compression, model quantization, model conversion and adaptation for different hardware and / or model software operating environments, etc., without exhaustive list. As another example, the data management function in the first function can be further divided into: data control functions and data execution functions (or data processing functions). As yet another example, the computing function can be further divided into: computing control functions and computing execution functions (or computing processing functions). The computing control function can be used to control the configuration of computing resources and the lifecycle management of computing tasks. The computing execution function can receive computing service data from users, transform it into computing tasks executable by computing nodes, and act as a gateway between computing service users and computing nodes.

[0078] It should be understood that the implementation methods of any two first functions (whether implemented by a single network function or by combining multiple network functions) can be the same or different. That is, for a communication system, there are three possibilities for the implementation methods of the first functions of each second network function in the communication system: completely the same, completely different, or partially different. This disclosure does not impose any particular restrictions on this.

[0079] Furthermore, as shown in Figures 1 and 2, the first network function and the second network function in this disclosure are deployed independently. This ensures the functional independence of the first and second network functions and provides good scalability. For example, when it is necessary to update or adjust the first function of the second network function, only the second network function needs to be processed, without significantly affecting the first network function, which also helps improve the stability of the entire communication system.

[0080] Based on the aforementioned communication system, this disclosure further provides a communication method, which is executed on a first network function and a second network function respectively. Through the cooperation of the first network function and the second network function, various first functions can be realized. The details are described below.

[0081] Please refer to Figures 3 and 4. Figure 3 is a flowchart illustrating one communication method provided in an embodiment of this disclosure, and Figure 4 is a flowchart illustrating another communication method provided in an embodiment of this disclosure.

[0082] The method shown in Figure 3 is applied to the second network function. As shown in Figure 3, the method includes:

[0083] S302, the second network function establishes a communication connection with the first network function.

[0084] S304, the second network function interacts with the first network function and / or the core network.

[0085] The second network function has at least one of the following primary functions: data management function, data analysis function, artificial intelligence / machine learning (AI / ML) model-related functions, computing function, and resource collaborative control function.

[0086] In any of the communication systems described above in this disclosure, the second network function can communicate with the first network function through the first interface, and the second network function can communicate with the core network through the service interface. Thus, the second network function can interact with the first network function and / or the core network, and thereby implement the first function.

[0087] Furthermore, in some embodiments described in the communication system section above, the second network function can both implement the first function with the help of the first network function and further manage the first network function. Thus, before implementing the first function of the second network function, S302 can be executed first to establish a communication connection with the first network function. This can also provide technical support for the implementation of the first function and / or the implementation of the management capabilities of the first network function.

[0088] The method shown in Figure 4 is applied to the first network function. As shown in Figure 4, the method includes:

[0089] S402, the first network function establishes a communication connection with the second network function.

[0090] S404, the first network function interacts with the second network function based on the communication connection.

[0091] The second network function has at least one of the following first functions: data management function, data analysis function, artificial intelligence / machine learning (AI / ML) model related functions, computing function, and resource collaborative control function.

[0092] As mentioned above, the first network function in this disclosure can independently achieve traditional communication connectivity, which will not be elaborated further. In addition, in the two-tier communication system architecture scenario provided in this disclosure, the first network function can also act as a supporter of the first function, assisting the second network function in achieving its first function. Similarly, before it can be used to assist in achieving the first function, it needs to establish a communication connection with the second network function in advance for subsequent message exchange.

[0093] In the embodiments shown in Figure 3 or Figure 4, the descriptions of the first network function, the second network function and its communication system, and the first function, etc., can be found in the preceding text and will not be repeated here.

[0094] The following describes the interaction process between the second network function and the first network function and / or the core network, explaining how the first network function and the second network function establish a communication connection, and how the first function of the second network function is implemented.

[0095] This disclosure provides an implementation method for establishing a communication connection between a first network function and a second network function. This process can also be simply understood as an interface establishment process between the two. This interface establishment process includes:

[0096] A1-1, the first network function sends a first message to the second network function. The first message is used to request the establishment of a communication interface.

[0097] B1-1, the second network function receives the first message from the first network function. The first message is used to request the establishment of a communication interface.

[0098] B1-2, the second network function sends a second message to the first network function.

[0099] A1-2, the first network function receives the second message from the first network function.

[0100] In the above interface establishment process, A1-1 to A1-2 are the actions performed by the first network function, and B1-1 to B1-2 are the actions performed by the second network function.

[0101] For ease of explanation, this disclosure will use "A" to identify the action performed by the first network function side and "B" to identify the action performed by the second network function side. Different processes are distinguished by the first digit after the number, and different steps of the same process are distinguished by the number after the "-" sign. For example, in "A1-1", the first digit "1" represents the interface establishment process, and the second "1" represents step 1 performed by the first network function in that process. Overall, it means that the first network function performs step 1 in process 1 (here referring to the interface establishment process).

[0102] Based on the above interaction process, the first network function can send a request to the second network function and establish a communication connection between the two, so that the first function can be implemented subsequently.

[0103] It should be noted that the above interface establishment process can be specifically as follows: a scenario where the first network function and the second network function establish a communication connection for the first time; or, a scenario where a connection is subsequently lost and then reconnected. This disclosure does not impose any particular restrictions on when the communication connection is established between the first network function and the second network function.

[0104] In the initial communication connection establishment scenario, the network function information of the second network function (xNB-IU) can be pre-configured in the first network function (xNB-CSU). This network function information includes at least address information. This allows the first network function to send a first message to the second network function based on the address information. The pre-configured network function information of the second network function can include other content, such as the first function supported by the second network function, subscription information, name, and other customizable information. This disclosure does not exhaustively list or limit this information. Furthermore, the pre-configuration of the second network function's network function information can be implemented manually by administrators, or by a dedicated management unit, such as a network management unit or system. In some embodiments, the first network function can also store the network function information and / or address information of the second network function (in the case of storing only the address). This allows the first network function to request the establishment of a communication connection with the second network function when subsequent communication connections or other interactions are needed, using the address information.

[0105] In practical implementation scenarios, such as step B1-2 in the interface establishment process described above, after receiving the first message, the second network function can also store relevant information of the first network function. That is, in some possible embodiments, before B1-2, the second network function can also perform the following operations:

[0106] B1-x, the second network function stores relevant information of the first network function.

[0107] Specifically, this relevant information can be directly indicated by the first request information, or it can be indicated in the first request information. For example, the relevant information may include, but is not limited to, the name of the first network function, address information, etc., which can be directly indicated by the first request information. In addition, for example, the first request information may further carry any other arbitrary information of the first network function, such as the capability information, capability change information, configuration information, etc. of the first network function, and this disclosure has no particular limitations in this regard.

[0108] In contrast, after the second network function sends a second message to the first network function, it can also store relevant information of the second network function based on the instructions of the second message and / or the information carried therein.

[0109] In one exemplary embodiment, when the second message carries second information about the second network function, the interface establishment process may further include the following steps:

[0110] A1-3, the first network function stores relevant information of the second network function; the relevant information is all or part of the second information.

[0111] In this embodiment, the second message may carry second information of a custom type and content, and this disclosure does not impose any particular restrictions on this; after receiving the second message, the first network function may store all or part of the second information. For example, when the second information carried in the second message includes attribute information and capability information of the second network function, the first network function may only store the capability information (i.e., part of the information in the second message), or it may directly store all of the second information.

[0112] Furthermore, the stored procedures shown in A1-3 and B1-x can be stored in any custom location, and this disclosure does not impose any particular restrictions on this.

[0113] To fully realize this solution, the second network function needs to know the capabilities supported by the first network function in order to better manage the first network function or help it achieve intelligent optimization and enhancement of wireless resource management. Therefore, the first network function needs to report its supported capabilities to the second network function. Furthermore, when the capabilities supported by the first network function change, it also needs to proactively or passively report the capability change information upon request from the second network function. Similarly, in some scenarios, the first network function can also obtain the capability information of the second network function to better support or assist the second network function in achieving its primary function.

[0114] In practical implementation scenarios, the interaction between the first network function and the second network function is also involved. This disclosure provides two possible implementation methods for this. The first method: The interaction between the two is achieved through the communication connection process (or interface establishment process) described above. Alternatively, the second method: After establishing a communication connection, the two functions achieve interaction through an independent capability interaction process. These are explained in detail below.

[0115] In the first implementation method, the interaction between the two can be achieved by using the first message and / or the second message in the communication connection process.

[0116] Therefore, in one exemplary embodiment, the above interface establishment process can further include the following steps:

[0117] The first message carries first information about the first network function; the first information is used to indicate the capabilities supported by the first network function.

[0118] And / or,

[0119] The second message carries second information about the second network function; the second information is used to indicate the capabilities supported by the second network function.

[0120] In this embodiment, both the first information and the second information are used to describe the capability information of the network function. The capability information here can be the capabilities currently supported by the network function, and / or capability change information when the network function changes. This disclosure does not impose any particular restrictions on the content of the first and second information; based on the capabilities supported by each network function in the actual scenario, notification to the other side can be complete or partial. Specifically, at the implementation level, each network function carries its own capability information in the message sent to the other side. It should also be noted that in this embodiment, based on the different scenarios in which the first and second network functions need to notify each other of their capabilities, there can be three possibilities: the first network function reports its own capabilities to the second network function (first information), the second network function notifies the first network function of its own capabilities (second information), and both parties exchange capability information (first information and second information).

[0121] For better understanding, please refer to Figure 5, which is a schematic diagram of the interaction flow of a communication method provided in an embodiment of this disclosure. Figure 5 illustrates the interface establishment process and information / capability query / subscription / update process between the first network function (xNB-CSU) and the second network function (xNB-IU). Please refer to the interface establishment process here, as shown in Figure 5, which includes:

[0122] Step 5.1: The network management unit or system (i.e., the communication system) configures the xNB-IU information for the xNB-CSU. This includes at least the xNB-IU's address information.

[0123] Step 5.2: The xNB-CSU pre-configures the address information of the xNB-IU. This means that the xNB-CSU can store or maintain the address information of the xNB-IU, and the specific method is not limited. In some possible embodiments, steps 5.1 and 5.2 can be a single step.

[0124] Step 5.3: xNB-CSU initiates an interface establishment request to xNB-IU (i.e., the first message mentioned above).

[0125] The interface establishment request may carry xNB-CSU related information (i.e., the first information). For example, the first information may include, but is not limited to, one or more of the following: xNB-CSU's data collection and reporting service capabilities, data analysis capabilities, data openness capabilities, controllable or configurable policies, xNB-CSU's model training capabilities (such as whether it supports local federated learning model training), xNB-CSU's AI / ML model capabilities, AI / ML model performance testing capabilities, AI / ML model inference capabilities, etc.; this is not an exhaustive list.

[0126] Step 5.4: The xNB-IU stores xNB-CSU related information. In specific implementation, the xNB-IU can store all or part of the first information.

[0127] Step 5.5, the xNB-IU sends an interface establishment request response message (i.e., the second message) to the xNB-CSU. In some scenarios, the second message can be the response message of the first interface, in which case the second message can also be called the interface establishment response message.

[0128] The interface establishment request response message may carry xNB-IU related information (i.e., the second information), which is used to indicate the xNB-IU configuration and capabilities. For example, the second information carried here, namely the xNB-IU capability information, may include, but is not limited to, one or more of the following: 1) Wireless data management functions, including but not limited to, one or more of the following: data acquisition, data preprocessing, data storage, data distribution, etc.; 2) Wireless data analysis functions, including but not limited to, one or more of the following: service identification, service feature perception analysis, wireless data statistical analysis, wireless data prediction, etc.; wherein, wireless data prediction may include statistical analysis and prediction of link coverage quality, load, network and user rate / latency / reliability, user KQI / QoE experience, network and terminal energy consumption, etc. at the network level / user level / slice level / service level (e.g., can be identified by 5QI or QCI); 3) AI / ML model-related functions, including but not limited to, one or more of the following: model training, model inference, model deployment, model storage, model compression, model optimization, etc.; 4) Computation-related functions, including but not limited to, one or more of the following: perception of terminal and / or base station computing resources, wireless computing resource control and scheduling; 5) Cooperative control functions of communication and computing resources and / or tasks; further details are omitted.

[0129] Step 5.6: The xNB-CSU stores xNB-IU related information. Specifically, the xNB-CSU may store all or part of the second information. Furthermore, in some embodiments, the related information stored in this step may be specifically used in the subsequent process of the xNB-CSU initiating a service request to the xNB-IU.

[0130] In summary, the interoperability of the first and second network functions can be directly achieved through message exchange during the communication connection process. Therefore, in the subsequent implementation of the first function, there is no need to repeatedly notify the other side of its own capabilities; instead, the first function can be implemented based on the capabilities supported by both itself and the other side.

[0131] The second implementation method involves establishing a process through either of the aforementioned interfaces, and then using a separate message exchange process between the two to achieve interoperability. This process can be referred to as the information / capability query / subscription / update process, as shown in the lower half of Figure 5.

[0132] In the information / capability query / subscription / update process, any network function can request the other network function to query its capabilities, and / or any network function can respond to the request of the other network function to provide its own capabilities to the other.

[0133] For example, for the second network function, based on the aforementioned interface establishment process, some embodiments may further include the following steps: B2-1 and / or B2-2.

[0134] Specifically, B2-1 means that the second network function sends a third message to the first network function and receives a fourth message from the first network function. That is, the second network function can act as a requester, requesting the first network function on the other side to provide feedback on the capabilities of the first network function.

[0135] Specifically, B2-2 means that the second network function receives a third message from the first network function and sends a fourth message to the first network function. That is, the second network function can act as a requested party, responding to a request from the first network function on the other side and providing its own capability information to the first network function on the other side.

[0136] For example, for the first network function, based on the aforementioned interface establishment process, some embodiments may further include the following steps: A2-1 and / or A2-2.

[0137] Specifically, A2-1 means: the first network function receives a third message from the second network function and sends a fourth message to the second network function. That is, the first network function can act as a requester, requesting the second network function on the other side to provide feedback on its capabilities.

[0138] Specifically, A2-2 means that the first network function sends a third message to the second network function and receives a fourth message from the second network function. That is, the first network function can act as the requested party, responding to a request from the second network function by providing its own capability information to the second network function.

[0139] In the two embodiments described above, the third message is used to request a query or subscription to the capabilities of the message recipient. Here, "query" refers to an immediate query of the current capabilities of the other network function, while "subscription" refers to a continuous query within a certain time frame. In practical scenarios, when the third message is used to subscribe to the capabilities of the message recipient, it can be called a subscription request (or subscription message, subscription request message, the name is unrestricted). In this case, the requesting network function can request the other network function to periodically report its own information or capabilities over a period of time, or it can request the other network function to report its own capabilities and information when certain preset conditions are met. The information carried in the third message differs depending on the subscription method. For example, the third message may further carry monitoring events and / or reporting trigger events; for the message recipient (i.e., the other network function), when it detects the monitoring event in the third message, and / or when the reporting trigger event is met, it can report its own information or capabilities back to the message sender. This disclosure does not impose any particular restrictions on the subscription method or the information carried in the subscription information (e.g., at least one of reporting period, reporting trigger event, and monitoring event). Both querying and subscribing aim to obtain the current updated capabilities or changes in capabilities of the other network. Therefore, this process can also be called: information / capability query / subscription / update process. In practical scenarios, the specific use of the third message for querying or subscribing can be determined based on the actual scenario. In practice, two different types of third messages can be distinguished to implement different query methods, which will not be elaborated here.

[0140] Based on the request from the local network function, the remote network function can provide feedback on its capabilities. Specifically, the fourth message carries the third information of the message recipient; the third information includes: currently supported capabilities, and / or capability change information. This third information is similar to the first and second information mentioned above, used to describe the capability information of the network function. Its specific content is related to the specific capabilities supported by the network function, and will not be elaborated further.

[0141] For better understanding, please refer to Figure 5 for the information / capability query / subscription / update process. As shown in Figure 5, this information / capability query / subscription / update process includes at least one of the following two processes (i.e., xNB-IU query and / or xNB-CSU query):

[0142] Among them, xNB-IU query means that xNB-IU can initiate information / capability query requests or subscription requests (i.e., third-party messages) to xNB-CSU. xNB-CSU monitors the queried or subscribed information or monitors the status of information / capabilities changes, and reports xNB-CSU information / capabilities to xNB-IU.

[0143] When the third message is a subscription request, it can further carry monitoring events and / or reporting trigger events. For example, a subscription request can be used to subscribe to changes in xNB-CSU information / capabilities. Upon receiving the subscription request, the xNB-CSU can report its own information or capabilities (i.e., the third information) to the xNB IU when its capabilities change. Alternatively, a subscription request can also be used to subscribe to xNB-CSU capability information over a period of time. Upon receiving the subscription request, the xNB-CSU can periodically report its own information or capabilities (i.e., the third information) to the xNB IU during that period.

[0144] Alternatively, when the third message is a query request (or query message, query request message, name is not limited), the xNB-CSU can report its own information or capabilities (i.e., third information) to the xNB-IU upon receiving the query request.

[0145] Among these, xNB-CSU queries, meaning xNB-CSU can also initiate information / capability query requests or subscription requests (i.e., third-party messages) to xNB-IU. xNB-IU monitors the queried or subscribed information or monitors changes in information / capabilities, and reports xNB-IU information / capabilities to xNB-CSU. (Refer to the previous text; it will not be repeated here.)

[0146] Based on the above processing, a communication connection is established between the first network function and the second network function, and the ability interaction between the two can be realized by establishing a process or a separate information / capability query / subscription / update process through a reused interface.

[0147] In this disclosure, the second network function, in addition to directly communicating with the first network function, also communicates with the core network. Thus, the second network function can interact with the core network and, through this interaction, can implement the first function. In its implementation, the second network function also needs to register its capabilities with the core network.

[0148] The following section describes the capability registration process of the second network function, using the interaction flow between the second network function and the core network as an example.

[0149] In one exemplary embodiment, the second network function is capable of interacting with the first network function and / or the core network. In specific implementation, this may include the following steps:

[0150] B3-1, the second network function sends a fifth message to the core network element, the fifth message being used for capability registration.

[0151] In practice, the fifth message can carry its own service information so that the core network elements on the other side can store the service information and perform subsequent operations related to the service information as needed.

[0152] B3-2, the second network function receives the sixth message from the core network element (in some embodiments, the sixth message may also be a response message to the fifth message; in other embodiments, it may also be a separately sent message or a reused message, without any particular restrictions, and will not be elaborated further).

[0153] The core network elements mentioned here are related to the first function of the second network function. For example, when the first function includes data-related functions, the core network elements can include data-related network elements. Similarly, when the first function includes computing-related functions, the core network elements can include computing-related network elements. Please refer to the preceding text for a description of the first function; it will not be listed exhaustively here.

[0154] For example, the core network elements involved in this disclosure may include, but are not limited to, at least one of the following: service discovery capability-related network elements, computing capability-related network elements, and data capability-related network elements. Specifically, the service discovery capability-related network elements may include, but are not limited to, at least one of the following: Network repository function (NRF) and service communication proxy (SCP). The computing capability-related network elements may be any one or more network elements in the core network related to computing capabilities, such as the computing plane function shown in Figure 2. The data capability-related network elements may be any one or more network elements in the core network related to data capabilities, such as the data plane function shown in Figure 2. An exhaustive list is not provided here.

[0155] Based on the capability registration process shown in B3-1 and B3-2, the core network can learn about the capabilities of the second network function (xNB-IU). Thus, on the one hand, the core network can issue relevant computing tasks and data tasks to the xNB-IU, and the xNB-IU, as a member of the distributed network, can improve task processing efficiency and reduce the load on the core network. On the other hand, the xNB-IU can also interact with the terminal side through the core network, and further assist the terminal in computation and provide feedback on the computation results (explained in detail later).

[0156] For example, please refer to Figure 6, which is a schematic diagram of the interaction flow of another communication method provided in this embodiment of the present disclosure. Specifically, Figure 6 shows the interaction flow of the second network function (xNB-IU) requesting capability registration from the core network. As shown in Figure 6, Figure 6 exemplarily illustrates three capability registration processes: the process of xNB-IU requesting service registration from network elements related to service discovery capability (referred to as the service discovery registration process for ease of explanation), the process of xNB-IU requesting service registration from network elements related to computing capability (referred to as the computing capability registration process for ease of explanation), and the process of xNB-IU requesting service registration from network elements related to data capability (referred to as the data capability registration process for ease of explanation). It should be understood that there is no strong correlation between these three service registration processes. Figure 6 is only an illustrative example. In actual scenarios, at least one of these three processes can be executed, and the execution order is not limited. Registration can be carried out simultaneously, in a preset order, or as needed according to actual conditions. Further details are omitted.

[0157] As shown in Figure 6, during the service discovery and registration process, xNB-IU can register its services with the core network NRF and / or SCP through a service-oriented interface. In this scenario, the fifth message is the service registration request in Figure 6, which may carry xNB-IU's service information (i.e., the service capabilities that xNB-IU can provide). For example, the service information carried in the service registration request may include, but is not limited to, one or more of the following: data acquisition service, data preprocessing service, data storage service, data analysis service, AI / ML model training service, AI / ML model inference service, AI / ML model storage service, AI / ML model performance testing, AI / ML model performance verification, computing service, and RAN capability opening service; among which, RAN capability opening service may open RAN-side network capability customization capabilities, RAN analysis data opening capabilities, etc.; this is not exhaustive.

[0158] In this scenario, for the core network, the NRF and / or SCP can receive service registration requests and save the service information of the xNB-IU carried in the service registration request. Thus, the second network function xNB-IU can be discovered by other NFs (Network Functions). Furthermore, the core network can also send a service registration response to the xNB-IU. This completes the service discovery and registration process.

[0159] As shown in Figure 6, during the computing capability registration process, the xNB-IU can register its computing resource information with the core network computing-related functions through a service-oriented interface. In this scenario, the fifth message is the computing resource information registration shown in Figure 6, which can carry the xNB-IU's service information (i.e., the computing resource information in Figure 6). Based on this computing resource registration process, the xNB-IU can also accept computing tasks issued by the core network computing-related functions and request the offloading of real-time computing tasks from the terminal for the distributed network computing function.

[0160] In this scenario, for the core network, the computing-related functions can receive computing resource information registration and save the xNB-IU's computing resource information carried in the registration. Therefore, the computing-related functions can send computing task requests to the xNB-IU based on the computing resource information. Furthermore, the core network can also send a computing resource registration response to the xNB-IU. This completes the computing capability registration process.

[0161] As shown in Figure 6, during the data capability registration process, the xNB-IU can register the data capabilities it can provide with the core network data-related functions through a service-based interface. In this scenario, the fifth message is the data capability registration shown in Figure 6, which can carry the xNB-IU's service information (i.e., the data capability information in Figure 6). For example, the data capability information indicates the data capabilities that the xNB-IU can provide, and may include, but is not limited to, at least one of the following: data type (i.e., the name of the data information), data time range, spatial range, time granularity of data collection and / or reporting, data source (the cell identifier, gNB identifier, service type-related identifier (such as UE identifier, 5QI identifier, slice identifier, etc.) generated by the available data), and data characteristics (statistical data or predicted data). For predicted data, the data capability information can further indicate one or more information such as the applicable time range for prediction and prediction confidence / accuracy.

[0162] In this scenario, for the core network, the data-related functions can receive data capability registrations and store the xNB-IU's data capability information carried in the registration. Thus, the data-related functions can utilize the xNB-IU to perform data analysis. For example, the core network's data-related functions can interact with the xNB-IU to obtain real-time RAN-side data and RAN analysis data at the cell / user / 5QI / slice levels from base stations and terminals. Furthermore, the core network can also send a data capability registration response to the xNB-IU, thereby completing the data capability registration process.

[0163] As mentioned above, the first function of the second functional network includes AI / ML model-related functions, and there are multiple possible ways to implement the AI / ML model-related functions.

[0164] Implementation Method 1: xNB-CSU can request xNB-IU to perform AI / ML model-related functional processing; while xNB-IU can perform AI / ML model-related functional processing based on xNB-CSU's request and return the processing results to xNB-CSU.

[0165] At this point, it can be achieved in the following way:

[0166] A4-1, the first network function sends a seventh message to the second network function, the seventh message being used to request AI / ML model processing, the AI / ML model processing including at least one of the following: model training, model testing, and model inference.

[0167] The seventh message may carry information about the AI / ML model's processing requirements and / or data. It should be understood that the information carried in the seventh message may differ depending on the type of AI / ML model processing request. For example, the seventh message may carry one or more of the following: model type, model parameters, model accuracy, training time, feature data, data to be inferred, loss function, model termination conditions, etc., without exhaustive list.

[0168] B4-1, the second network function receives a seventh message from the first network function, the seventh message being used to request AI / ML model processing, the AI / ML model processing including at least one of the following: model training, model testing, and model inference.

[0169] B4-2, the second network function performs model processing based on the seventh message to obtain the first processing result.

[0170] Specifically, the second network function can allocate processing resources for the processing task and utilize these resources for model processing. Specifically, the processing resources mentioned here and in step d4 below can include, but are not limited to, at least one of: computing resources and storage resources.

[0171] In specific implementation scenarios, when the second network function performs model processing, it can directly process the data carried or indicated by the seventh message. Alternatively, in some possible embodiments, the second network function can also perform necessary data acquisition and / or data preprocessing to obtain feature data that meets its processing requirements.

[0172] B4-3, the second network function sends the first processing result to the first network function.

[0173] It should be understood that the first processing result (also known as the first processing data) can differ depending on the type of AI / ML model processing. For example, when the seventh message request is used for model training, the first processing result may specifically include, but is not limited to, at least one of the following: model parameters, model code data, model architecture data, and overall model data. As another example, when the seventh message request is used for model testing, the first processing result may specifically include at least one of the following: test data for the model, test conclusions, and evaluation data. Furthermore, when the seventh message request is used for model inference, the first processing result may specifically be the inference result obtained by processing the data to be inferred using the model.

[0174] It should be noted that the first processing result can be: the specific data obtained by the model processing, and / or the storage location of the specific data obtained by the model processing. For example, after training the model, xNB-IU may have a large number of data parameters and a large data volume. Directly sending the first processing result may not be efficient or secure. Therefore, the specific data obtained by the processing can be stored in a specified location (readable by both xNB-IU and xNB-CSU), and xNB-CSU can be notified of the storage location of this data (in this case, the storage location is the first processing result), and xNB-CSU can then retrieve the corresponding data itself.

[0175] In some embodiments, the first processing result may also carry more relevant data, such as test metrics like model inference accuracy and recall, training process data like model loss function, model training duration, and number of model training rounds, etc., without exhaustive enumeration, and can be customized.

[0176] A4-2, the first network function receives the first processing result from the second network function.

[0177] In this embodiment, xNB-IU can assist xNB-CSU in model processing, such as model training and / or AI / ML model inference services. It should be understood that B4-1 to B4-3 represent the actions performed by xNB-IU to assist xNB-CSU in model processing, while A4-1 to A4-2 represent the actions performed by xNB-CSU to request xNB-IU's assistance in model processing.

[0178] In practical implementations, when any network function receives data or messages from a peer network function, it can also send a response message to the peer network function so that the peer network function can be aware of the data or message reception status. For example, after step B4-1 or B4-2 above, xNB-IU can also send a model processing response message to xNB-CSU. As another example, after step A4-2 above, after xNB-CSU receives the first processing result obtained by xNB-IU from model processing, xNB-CSU can also send a model data reception response message to xNB-IU.

[0179] In some embodiments, based on the above embodiments, the method may further include:

[0180] B4-4, Second network function release processing resources.

[0181] In other words, after confirming that xNB-IU has received the first processing result, it can reclaim the processing resources allocated for model processing to xNB-CSU to avoid resource waste. Specifically, B4-4 involves the same processing resources as step B4-2 above, which may include, but are not limited to, at least one of computing resources and storage resources.

[0182] As an exemplary embodiment, please refer to Figure 7, which is a schematic diagram of the interaction flow of another communication method provided in this disclosure embodiment. Specifically, Figure 7 shows the interaction flow of xNB-IU assisting xNB-CSU in model training processing. As shown in Figure 7, the method includes:

[0183] Step 7.1: xNB-CSU sends a model training request to xNB-IU (i.e., the seventh message).

[0184] The model training request (also known as the model training request message, model training message, etc., i.e., the seventh message of this disclosure) carries model training requirement information, which may include, but is not limited to, at least one of the following: model type, model structure, model parameters, model accuracy requirements (accuracy, recall, etc.), required training duration (the required training time), feature data (or the data storage address, read by xNB-IU) or the requirement for feature data (collected and preprocessed by xNB-IU), and pruning requirements. If there are requirements for model size, the model pruning requirements can be specified in the model training message.

[0185] Step 7.2: xNB-IU allocates computing and storage resources and acquires feature data according to the model training request.

[0186] Step 7.3: xNB-IU sends a model training success response to xNB-CSU. As mentioned earlier, this step is optional.

[0187] Step 7.4: If feature data needs to be collected and processed, the xNB-IU can also collect, store, and process the data according to the feature data requirements. This step is optional, hence it is represented by a dashed box in Figure 7.

[0188] Step 7.5, xNB-IU performs model training and / or testing to obtain model parameters that meet the accuracy requirements.

[0189] Step 7.6: The xNB-IU sends the trained model parameters or the storage address of the model parameters (i.e., the first processing result) to the xNB-CSU. In other embodiments, the trained model or storage address (which the xNB-CSU can read from, i.e., another form of the first processing result) may also carry test metrics such as model inference accuracy and recall, etc., without exhaustive list.

[0190] Step 7.7: After the xNB-CSU successfully receives the model, it sends a model reception response to the xNB-IU.

[0191] Step 7.8: xNB-IU reclaims computing and storage resources.

[0192] In summary, the xNB-IU disclosed herein possesses AI / ML model-related functions. Thus, this primary function can be utilized to help the xNB-CSU complete AI / ML model training, model inference, and other related processing tasks.

[0193] It should be noted that, in the implementation of the first function described in Method 1, the xNB-CSU can autonomously send a seventh message to the xNB-IU to initiate the xNB-IU's assistance processing based on its own circumstances; alternatively, the xNB-CSU can passively trigger the above process based on a request from the terminal side. The following is a detailed explanation.

[0194] As described in the communication system section above, for example, in the communication system shown in Figure 2, the xNB-IU does not have a direct communication connection with the terminal. Therefore, communication with the terminal can be achieved through the xNB-CSU and / or the core network. This relates to the situation where the terminal can also initiate a first function implementation request to the xNB-IU through the xNB-CSU. In other words, since the xNB-IU has no communication connection function, the message interaction where the UE requests the xNB-IU to perform AI model training and / or AI / ML model inference needs to be forwarded through the xNB-CSU. Thus, in one possible embodiment of this disclosure, the seventh message involved in the above embodiments can be triggered by the first network function based on the terminal's request.

[0195] At this point, for the first network function, before step A4-1 mentioned above, the following two steps may also be included:

[0196] A4-01, before sending the seventh message to the second network function, the first network function receives a tenth message from the terminal, the tenth message being used to request the second network function to perform AI / ML model processing.

[0197] A4-02, the first network function sends the first processing result to the terminal.

[0198] At this point, please refer to Figure 8, which is a schematic diagram of the interaction flow of another communication method provided in this embodiment of the present disclosure. Specifically, Figure 8 shows the interaction flow of the terminal requesting the xNB-IU to perform model processing through the xNB-CSU.

[0199] As shown in Figure 8, the UE can send an AI model training / inference request (i.e., the tenth message) to the xNB-CSU via an air interface message. The xNB-CSU can then convert the air interface message into an Infx interface message and send the AI ​​model training / inference request (the seventh message) to the xNB-IU via the Infx interface message. The processing flow on the xNB-IU side can be referred to the previous text and will not be repeated here. Similarly, messages sent by the UE to the xNB-IU via the xNB-CSU can be processed in a similar manner, such as the AI ​​model training / inference result response in Figure 8.

[0200] Furthermore, messages sent from the xNB-IU to the UE via the xNB-CSU can also be forwarded by the xNB-CSU after message conversion. As shown in Figure 8, after receiving the AI ​​model training / inference response or AI model training / inference result from the xNB-IU, the xNB-CSU converts the received Infx interface message into an air interface message and sends it to the UE.

[0201] Implementation Method Two: xNB-IU can proactively collaborate with multiple xNB-CSUs to achieve federated AI / ML model processing, which can be referred to as a joint AI / ML model processing flow. For example, xNB-IU can collaborate with multiple xNB-CSUs to perform federated learning.

[0202] Taking the AI / ML model-related functions based on the second network function as an example to illustrate the joint processing of AI / ML models, the joint processing flow of AI / ML models can be implemented as follows:

[0203] B5-1, the second network function identifies at least one target network function among the first network functions.

[0204] In practical implementation, based on the aforementioned interface, a process or a separate information / capability query, subscription, or update process is established. The second network function can obtain the capabilities and / or relevant information of each first network function. Therefore, when executing the joint processing process, the second network function can select one or more suitable first network functions to achieve joint AI / ML model processing. In this disclosure, the model processing method here is not limited; referring to the preceding text, it will not be repeated here.

[0205] B5-2, the second network function iteratively processes each target network function until the preset requirements are met.

[0206] The preset requirements can be customized. For example, they may include, but are not limited to, at least one of the following: reaching a preset number of iterations, reaching a preset iteration duration, reaching a preset model processing accuracy, etc., without exhaustive list.

[0207] Specifically, in implementing step B5-2, the model data processing in each iteration is achieved jointly by one or more first network functions through message exchange with the first network functions. Specifically, any iteration process includes:

[0208] B5-2-1, the second network function sends an eighth message to each target network function, the eighth message being used to request joint model processing.

[0209] Specifically, the eighth message is used to request the first network function to perform joint model processing. It may specifically carry or indicate model information (e.g., model parameters, model architecture, initial model, etc., not exhaustively listed). In some embodiments, the eighth message may also carry or indicate data information. Specifically, the indication of model information and data information may be achieved by carrying the storage location of the model information / data information in the eighth message.

[0210] Furthermore, in practical implementation scenarios, multiple first network functions are often involved. In such cases, the information carried in the eighth message sent to different first network functions can be the same, different, or partially different. For example, a second network function can uniformly send global model data to multiple first network functions; or, for another example, a second network function can split the global model into multiple sub-models (or sub-modules, which are essentially parts of the model) and send the multiple sub-models (or parameters of the sub-models) to multiple first network functions respectively. In this way, each first network function can be used to process a part of the model.

[0211] B5-2-2, the second network function receives the second processing results from each target network function.

[0212] There can be multiple second processing results, and any one of the second processing results is obtained by joint model processing of the target network functions.

[0213] B5-2-3, the second network function aggregates the models based on the results of each of the second processing steps to obtain a global model.

[0214] For the second network function, the B5-2-2 step can receive multiple second processing results. In order to determine the specific effect of the model, the second network function also needs to perform model aggregation processing on multiple second processing results to obtain relevant data of the global model, and then perform global model evaluation to obtain a more accurate model training result.

[0215] In practical implementation, if B5-2-1 is distributed to multiple first network functions in the form of splitting into sub-models for joint training, then in this step, it is only necessary to concatenate and restore according to the splitting logic. Alternatively, for example, if B5-2-1 is the second processing result obtained by training multiple global models separately from multiple first network functions, then in this step, the multiple second processing results can be processed, such as averaging, taking the maximum value, or other preset fusion processing methods, to obtain the global model. Further details are omitted.

[0216] B5-2-4, the second network function determines whether the global model meets the preset requirements.

[0217] As mentioned earlier, in each round of iterative training, it is necessary to determine whether the preset requirements are met. If the preset requirements are not met, the process is repeated starting from B5-2-1. If the preset requirements are met, the result of meeting the preset requirements can be directly used as the model training result.

[0218] Accordingly, for the first network function (here only considering xNB-CSU selected as the target network function by the second network function), in any of the above iterations, i.e., between B5-2-1 and B5-2-2, the first network function only needs to process and return the processing result based on the joint processing request of the second network function. At this time, between B5-2-1 and B5-2-2, the following steps may also be included:

[0219] A5-1, the first network function receives an eighth message from the second network function, the eighth message being used to request joint model processing.

[0220] A5-2, the first network function performs local model processing to obtain the second processing result.

[0221] Specifically, the first network function only needs to train the model locally based on the model and data indicated by the eighth message. For example, the first network function can train the model using local data or using data from a specified storage location indicated by the eighth message.

[0222] A5-3, the first network function sends the second processing result to the second network function.

[0223] For example, please refer to Figure 9, which is a schematic diagram of the interaction flow of another communication method provided in this embodiment of the disclosure. Specifically, Figure 9 illustrates the process of joint federated learning between xNB-IU and multiple xNB-CSUs. As shown in Figure 9, the joint federated learning process includes:

[0224] Step 9.1: The xNB-IU acts as the aggregation node and selects a suitable xNB-CSU as a participating node (i.e., the target network function mentioned above) for federated learning.

[0225] Step 9.2: xNB-IU sends the global model to xNB-CSU participating in federated learning (i.e., the model information is carried in the eighth message).

[0226] It should be understood that Figure 9 is merely exemplary, and the information carried in the eighth message shown here is only one possible implementation. Specifically, the global model here may further include: initial model parameters, wherein the initial model parameters may include, but are not limited to, at least one of the following: initial weights, biases, learning rate, batch size, loss function, number of iterations, model architecture parameters, accuracy requirements, etc., without exhaustive list. In addition, as mentioned above, the eighth message may also carry more (e.g., further indicating data, indicating data type, etc.) or less (e.g., pre-preset or specified) information, and this disclosure has no particular limitation in this regard.

[0227] Step 9.3: xNB-CSU uses local data for model training. That is, xNB-CSU performs joint model processing.

[0228] Step 9.4: xNB-CSU sends local model parameters (i.e., the second processing result) to xNB-IU.

[0229] If the required accuracy is achieved or the number of iterations reaches the set number (i.e., when the preset requirements are met), in addition to sending the local model parameters to xNB-IU, the second processing result can also carry more information, such as gradients and accuracy.

[0230] In step 9.5, xNB-IU aggregates the models sent by xNB-CSU participating in federated learning to obtain a global model.

[0231] Step 9.6: xNB-IU performs a performance evaluation on the global model.

[0232] This involves evaluating whether the preset requirements have been met. If they are met, the training ends. If not, the process repeats to the next iteration.

[0233] For example, when the preset requirement is to achieve a preset model accuracy, an accuracy test is required. If the accuracy does not meet the requirements, the global model is sent to xNB-CSU for federated learning until the preset requirements are met.

[0234] The above uses AI / ML model-related functions as an example to illustrate the interaction process between xNB-IU, xNB-CSU, and UE. In this interaction process, xNB-CSU can proactively initiate an application request for the first function to xNB-IU; xNB-CSU can also act as a message relay between the UE and xNB-IU based on UE-side requests, enabling the UE to request and apply the first function of xNB-IU; and xNB-IU can also act as an active initiator, assigning its own tasks to one or more xNB-CSUs, with at least one xNB-CSU working together to implement its own first function. It should be understood that other first functions of xNB-IU can also be implemented in the manner described above, and will not be listed exhaustively. For example, xNB-CSU can proactively request the computing function of xNB-IU; xNB-IU can proactively implement its own resource coordination control function for multiple xNB-CSUs; and the UE can initiate a request for data analysis function from xNB-IU through xNB-CSU.

[0235] As mentioned above, in this disclosure, the xNB-IU can communicate with the UE through the xNB-CSU and / or the core network. Therefore, for the UE, when the UE needs to invoke or request the first function of the xNB-IU, it can do so through the xNB-CSU and / or the core network.

[0236] To further illustrate this process, the following example uses the scenario of a UE requesting computing services from an xNB-IU to explain the interaction between the two.

[0237] At this point, the implementation of the computational function of the second network can proceed as follows:

[0238] B6-1, the second network function receives a ninth message from the first network function or the core network, the ninth message being used to request computing services to the terminal.

[0239] In other words, the UE can initiate a computing service request (or computing request) to the xNB-IU through the xNB-CSU or the core network. It should be understood that the computing services that the second network function can provide can be of various kinds, such as computing offloading tasks.

[0240] B6-2, the second network function allocates computing resources to the computing tasks of the terminal, and uses the computing resources to process the computing tasks to obtain a third processing result.

[0241] Computational tasks primarily involve computing resources; in addition, storage resources may also be involved in practical implementation scenarios. That is, before a computing task is executed, the second network function can allocate computing and storage resources for the terminal's computing task; after the task is completed, the second network function can also reclaim the computing and storage resources.

[0242] The computational task performed here can be provided or specified by the UE. This can be done by implicitly sending it in the ninth message or by sending it separately. Taking xNB-CSU as the target object as an example, after allocating resources to xNB-IU, xNB-CSU can initiate a user plane bearer request to the UE to establish a user plane bearer between the UE, xNB-CSU, and xNB-IU, thereby enabling the transmission of computational user plane data (i.e., computational task-related data).

[0243] B6-3, the second network function sends the third processing result to the first network function or the core network.

[0244] The results of the computational tasks performed by the second network function can also be fed back to the UE through the xNB-CSU or the core network. For example, task data can be sent to the UE through the user plane bearer. In addition, in embodiments involving user plane bearer establishment, when B6-3 ends, the xNB-IU can also initiate a computational user plane bearer deletion request to the UE through the xNB-CSU to delete the computational user plane bearer and reclaim computational and storage resources.

[0245] Based on the above process, for the second network function, the ninth message it receives may originate from the first network function or from the core network. Therefore, different interaction methods can be used for different target objects.

[0246] In one possible implementation, the UE requests computing services from the xNB-IU via the xNB-CSU. In this way, the xNB-IU can help the UE perform computing tasks and return the computing results to the UE via the xNB-CSU.

[0247] At this point, for the first network function, the following processing is possible:

[0248] A6-1, the first network function receives an eleventh message from the terminal, the eleventh message being used to request the second network function to provide computing services to the terminal.

[0249] Since the xNB-IU and UE cannot communicate directly, the UE can send requests to the xNB-CSU to the xNB-IU, which will then forward them to the xNB-IU.

[0250] A6-2, the first network function sends a ninth message to the second network function, the ninth message being used to request computing services to the terminal.

[0251] In practice, the eleventh message can be an air interface message, and xNB-CSU will convert the air interface message into an Infx interface message, i.e., the ninth message, and forward it.

[0252] After A6-2, xNB-IU executes the aforementioned B6-1 to B6-3, performs the computation task, and provides feedback.

[0253] A6-3, the first network function receives the third processing result from the second network function.

[0254] A6-4, the first network function sends the third processing result to the terminal.

[0255] In contrast to A6-2, in this step, the xNB-CSU converts the received Infx interface message into an air interface message and sends it to the UE.

[0256] Based on the above process, with the assistance of the xNB-CSU, the UE can initiate a request to the xNB-IU for its primary function.

[0257] For easier explanation, please refer to Figure 10, which is a schematic diagram of the interaction flow of another communication method provided in this embodiment of the present disclosure. Specifically, Figure 10 shows the interaction flow of a terminal requesting xNB-IU to provide compute offloading services through xNB-CSU. The following explanation uses compute offloading services as an example to illustrate how xNB-IU provides compute services. As shown in Figure 10, the process may include:

[0258] Step 10.1: The UE packages the "computation request" into an air interface message (i.e., the eleventh message) and sends it to the xNB-CSU.

[0259] As mentioned earlier, computation requests can have various possible scenarios in real-world situations. For example, Figure 10 illustrates the "computation unloading request" as an example.

[0260] The computation request may carry resource requirement information for the computation task, which may include, but is not limited to, at least one of the following: the type of computation resource required by the computation task and the size of the computation resource; in addition, in some embodiments, it may also include, but is not limited to, at least one of the following: the size of the data required for computation, the transmission rate requirement, the computation time requirement, etc.

[0261] Step 10.2: The xNB-CSU forwards the message to the xNB-IU. This is the ninth message sent to the xNB-IU.

[0262] In one exemplary embodiment, the xNB-CSU can parse the message as a message for the xNB-IU, and then send it to the xNB-IU via the Infx interface.

[0263] Step 10.3: The xNB-IU allocates appropriate computing and storage resources based on the computing request, such as GPU resources and container resources.

[0264] Step 10.4: The xNB-IU's computation control initiates a computation user plane bearer establishment message to the xNB-CSU and the UE.

[0265] The computing user plane bearer is used to transmit computing-related data, which may include, but is not limited to, at least one of the following: bearer ID, data transmission rate requirements (maximum rate, minimum rate), time requirements, data endpoint (IP, port) information, etc.

[0266] Step 10.5: After receiving the bearer establishment message, the xNB-CSU first allocates transmission-related resources to establish a user plane data transmission channel with the xNB-IU. After success, it then initiates a dedicated bearer establishment to the UE.

[0267] Step 10.6: The xNB-CSU initiates a dedicated bearer establishment message for calculating data transmission to the UE via the air interface. This message may include, but is not limited to, at least one of the following information: bearer ID, configuration information of each layer of the user plane (such as the PDCP layer's drop timer, the RLC layer's transmission mode, the MAC layer's HARQ retransmission count, and other protocol layer configurations).

[0268] Step 10.7: The UE sends a dedicated bearer establishment response message to the xNB-CSU.

[0269] Step 10.8: The xNB-CSU sends a computation user plane bearer establishment response message to the xNB-IU. This completes the transmission channel (computation user plane data bearer) for computation-related data from the UE to the xNB-IU. The UE and xNB-IU can then transmit computation-related data (i.e., the third processing result) over this bearer, including computation files, data required for computation, data interactions during the computation process, and computation results.

[0270] Step 10.9: After completing this computation task, the xNB-IU initiates the computation user plane data deletion process and reclaims the allocated computation and storage resources.

[0271] In another possible implementation, the UE requests computing services from the core network, the core network offloads the task to the xNB-IU, the xNB-IU assists the UE in performing the calculation, and returns the calculation result to the UE through the core network.

[0272] At this point, please refer to Figure 11. Specifically, Figure 11 illustrates the interaction process of a terminal requesting computing services from the xNB-IU through the core network. Figure 11 further explains the process using computing offloading as an example. As shown in Figure 11, this process may include:

[0273] Step 11.1: The xNB-IU has completed the registration process of its computing resource information with the core network. Refer to the capability registration process described earlier for details.

[0274] Step 11.2: The UE's computing function initiates a computing request to the core network computing function. The resource requirement information for the computing task carried in the computing request may specifically include, but is not limited to, at least one of the following: the type and size of computing resources required by the computing task; furthermore, in some embodiments, it may also include, but is not limited to, at least one of the following: the data size required for the computing, transmission rate requirements, and computing time requirements.

[0275] Step 11.3: The core network computing function searches for suitable computing resources to offload the computing task based on the information in the computing service request. In the embodiment shown in Figure 11, the core network can select a suitable xNB-IU to execute the computing task.

[0276] In other words, the UE can initiate a compute offload request to the core network, and the core network can either handle it itself or distribute the compute offload task to other execution entities, such as the xNB-IU, which will then provide compute offload services to the UE.

[0277] This disclosure does not impose any particular restrictions on how the core network determines the processing node for computing tasks. Since the aforementioned xNB-IU has registered its capabilities and computing resources with the core network, the core network can determine the processing node for computing tasks from within the scope including the xNB-IU. Figure 11 illustrates the specific scenario where the core network selects the xNB-IU as the processing node for the UE's computing tasks.

[0278] Step 11.4: The core network sends a computing request (i.e., the ninth message) to the xNB-IU. In the computing offload service scenario shown in Figure 11, the computing request can be specifically a task offload indication message, which carries the computing requirements sent by the UE.

[0279] Step 11.5: The xNB-IU allocates appropriate computing and storage resources according to the computing request.

[0280] Step 11.6: The xNB-IU sends a compute offload response message to the core network and requests the establishment of a compute data transmission channel. This dedicated data transmission channel can either be directly offloaded to the xNB-IU via xNB-CSI or forwarded via UPF. When the former is selected, the core network needs to configure xNB-CSI to instruct it to send the corresponding compute task data directly to the corresponding xNB-IU.

[0281] Step 11.7: Establish a dedicated bearer for computing data in the core network to create a dedicated transmission channel for computing task data transmission.

[0282] Step 11.8: On the dedicated data transmission channel for computation, the UE and xNB-IU interact to process computation-related data, including computation files, data required for computation, data interaction during the computation process, and computation results (i.e., through the dedicated data transmission channel for computation, xNB-IU sends the third processing result to the UE via the core network).

[0283] Step 11.9: After completing this computation task, the xNB-IU reclaims the allocated computing and storage resources and sends a computation task completion response to the core network.

[0284] Step 11.10: The core network releases the dedicated bearer for computing data with the UE.

[0285] The above, taking the computing function as an example and referring to Figures 10 and 11, specifically illustrates that under the two-layer communication connection architecture provided in this disclosure, the UE can either directly request its first function from the xNB-IU through the xNB-CSU; or, the UE can directly request it from the core network. The core network, based on the xNB-IU's capability registration process, can include the xNB-IU within the scope of processing nodes providing services to the UE. Thus, through task allocation by the core network, the xNB-IU can also indirectly provide its first function to the UE. It should be understood that other first functions besides the computing function can also be provided to the UE in this manner.

[0286] In summary, the technical solution provided in this disclosure, compared to existing solutions, further deploys an intelligent unit (i.e., a second network function, xNB-IU) that transcends connectivity, building upon the traditional communication connection network function (i.e., the first network function xNB-CSU). This forms a two-tiered system architecture, and both xNB-CSU and xNB-IU can be independently deployed and upgraded as needed. This allows for the flexible introduction of computing and intelligent capabilities. These capabilities, compared to traditional connection functions, can be designed based on a microservice architecture, offering better flexibility and scalability. Simultaneously, the relatively centralized computing and intelligent network functions provide more powerful computing capabilities for the distributed connection network functions, and through centralization, achieve better resource pooling and utilization efficiency, thereby improving network resource utilization efficiency.

[0287] In addition, this disclosure also provides a network function (a second network function), including:

[0288] Establish a unit, configured to establish a communication connection with the first network function;

[0289] The communication unit is configured to interact with the first network function and / or the core network; wherein the second network function has at least one of the following first functions: data management function, data analysis function, artificial intelligence / machine learning (AI / ML) model-related functions, computing function, and resource collaborative control function.

[0290] In one exemplary embodiment, the communication unit is specifically configured as follows:

[0291] Receive a first message from the first network function, the first message being used to request the establishment of a communication interface;

[0292] Store relevant information related to the first network function;

[0293] Send a second message to the first network function.

[0294] In one exemplary embodiment, the first message carries first information about the first network function; the first information is used to indicate the capabilities supported by the first network function; and / or, the second message carries second information about the second network function; the second information is used to indicate the capabilities supported by the second network function.

[0295] In one exemplary embodiment, the communication unit is specifically configured as follows:

[0296] Send a third message to the first network function and receive a fourth message from the first network function;

[0297] And / or,

[0298] Receive a third message from the first network function and send a fourth message to the first network function;

[0299] The third message is used to query or subscribe to the capabilities of the message recipient; the fourth message carries third information about the message recipient; the third information includes: currently supported capabilities, and / or capability change information.

[0300] In one exemplary embodiment, the communication unit is specifically configured as follows:

[0301] Send a fifth message to the core network element, the fifth message being used for capability registration;

[0302] Receive the sixth message from the core network element;

[0303] The core network elements include at least one of the following: network elements related to service discovery capability, network elements related to computing capability, and network elements related to data capability.

[0304] In one exemplary embodiment, the communication unit is specifically configured as follows:

[0305] Receive a seventh message from the first network function, the seventh message being used to request AI / ML model processing, the AI / ML model processing including at least one of the following: model training, model testing, and model inference;

[0306] Based on the seventh message, model processing is performed to obtain the first processing result;

[0307] Send the first processing result to the first network function.

[0308] In one exemplary embodiment, the seventh message is triggered by the first network function based on a request from the terminal.

[0309] In one exemplary embodiment, the communication unit is specifically configured as follows:

[0310] Identify at least one target network function from the first network functions of each institution;

[0311] Iterative processing is performed based on the functions of each target network until the preset requirements are met; wherein, any round of the iterative processing includes:

[0312] Send an eighth message to each target network function, the eighth message being used to request joint model processing;

[0313] Receive the second processing results from each target network function;

[0314] Based on the results of each of the second processing steps, the models are aggregated to obtain a global model;

[0315] Determine whether the global model meets the preset requirements.

[0316] In one exemplary embodiment, the communication unit is specifically configured as follows:

[0317] The system receives a ninth message from the first network function or the core network, the ninth message being used to request computing services to be provided to the terminal; wherein the target object is: the first network function or the core network;

[0318] Allocate computing resources to the computing tasks of the terminal, and use the computing resources to process the computing tasks to obtain a third processing result;

[0319] The third processing result is sent to the first network function or the core network.

[0320] In addition, this disclosure also provides another network function (first network function), including:

[0321] Establish a unit, configured to establish a communication connection with the second network function;

[0322] The communication unit is configured to interact with the second network function based on the communication connection; wherein the second network function has at least one of the following first functions: data management function, data analysis function, artificial intelligence / machine learning (AI / ML) model-related functions, computing function, and resource collaborative control function.

[0323] In one exemplary embodiment, the communication unit is specifically configured as follows:

[0324] Send a first message to the second network function, the first message being used to request the establishment of a communication interface;

[0325] Receive a second message from the first network function.

[0326] In one exemplary embodiment, the first message carries first information about the first network function; the first information is used to indicate the capabilities supported by the first network function; and / or, the second message carries second information about the second network function; the second information is used to indicate the capabilities supported by the second network function.

[0327] In one exemplary embodiment, when the second message carries second information about the second network function, the communication unit is specifically configured as follows:

[0328] Store relevant information for the second network function; the relevant information may be all or part of the second information.

[0329] In one exemplary embodiment, the communication unit is specifically configured as follows:

[0330] Receive a third message from the second network function and send a fourth message to the second network function;

[0331] And / or,

[0332] Send a third message to the second network function and receive a fourth message from the second network function;

[0333] The third message is used to query or subscribe to the capabilities of the message recipient; the fourth message carries third information about the message recipient; the third information includes: currently supported capabilities, and / or capability change information.

[0334] In one exemplary embodiment, the communication unit is specifically configured as follows:

[0335] Send a seventh message to the second network function, the seventh message being used to request AI / ML model processing, the AI / ML model processing including at least one of the following: model training, model testing, and model inference;

[0336] Receive the first processing result from the second network function.

[0337] In one exemplary embodiment, the communication unit is specifically configured as follows:

[0338] Before sending the seventh message to the second network function, a tenth message is received from the terminal, the tenth message being used to request the second network function to perform model processing;

[0339] The first processing result is sent to the terminal.

[0340] In one exemplary embodiment, the communication unit is specifically configured as follows:

[0341] Receive an eighth message from the second network function, the eighth message being used to request joint model processing;

[0342] Perform local model processing to obtain the second processing result;

[0343] Send the second processing result to the second network function.

[0344] In one exemplary embodiment, the communication unit is specifically configured as follows:

[0345] Receive an eleventh message from the terminal, the eleventh message being used to request the second network function to provide computing services to the terminal;

[0346] Send a ninth message to the second network function, the ninth message being used to request computing services to the terminal;

[0347] Receive the third processing result from the second network function;

[0348] The third processing result is sent to the terminal.

[0349] Figure 12 is a hardware block diagram of an electronic device provided in an embodiment of this disclosure. The electronic device 1200 according to an embodiment of this disclosure includes at least a memory, a processor, and a computer program stored in the memory. The processor executes the computer program to implement the communication method described in any of the above embodiments.

[0350] The electronic device 1200 shown in Figure 12 specifically includes a central processing unit (CPU) 1201, a graphics processing unit (GPU) 1202, and a memory 1203. These units are interconnected via a bus 1204. The CPU 1201 and / or GPU 1202 can function as the aforementioned processor, and the memory 1203 can function as the aforementioned memory for storing computer-readable instructions. Furthermore, the electronic device 1200 may also include a communication unit 1205, a storage unit 1206, an output unit 1207, an input unit 1208, and an external device 1209, all of which are also connected to the bus 1204.

[0351] Figure 13 is a schematic diagram of a computer-readable storage medium provided in an embodiment of this disclosure. As shown in Figure 13, the computer-readable storage medium 1300 according to an embodiment of this disclosure stores a computer program / instructions 1301. When executed by a processor, the computer program / instructions 1301 implements the communication method described in any of the preceding embodiments of this disclosure. The computer-readable storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, optical disk, magnetic disk, etc.

[0352] This disclosure further provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the communication method described in any of the preceding embodiments of this disclosure.

[0353] The basic principles of this disclosure have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this disclosure are merely examples and not limitations, and should not be considered as essential features of each embodiment of this disclosure. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the scope of this disclosure to the necessity of employing the aforementioned specific details for implementation.

[0354] The block diagrams of devices, apparatuses, devices, and systems disclosed herein are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0355] Additionally, as used herein, the “or” used in a list of items beginning with “at least one” indicates a separate list, such that a list of, for example, “at least one of A, B, or C” means A or B or C, or AB or AC or BC, or ABC (i.e., A and B and C). Furthermore, the word “exemplary” does not imply that the described example is preferred or better than other examples.

[0356] It should also be noted that in the systems and methods of this disclosure, the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered as equivalent solutions to this disclosure.

[0357] Various changes, substitutions, and modifications can be made to the technology described herein without departing from the teachings defined by the appended claims. Furthermore, the scope of the claims of this disclosure is not limited to the specific aspects of the processes, machines, manufactures, events, means, methods, and actions described above. Currently existing or later-developed processes, machines, manufactures, events, means, methods, or actions that perform substantially the same function or achieve substantially the same result as the corresponding aspects described herein can be utilized. Therefore, the appended claims include such processes, machines, manufactures, events, means, methods, or actions within their scope.

[0358] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of this disclosure. Therefore, this disclosure is not intended to be limited to the aspects shown herein, but rather to be carried out within the widest scope consistent with the principles and novel features disclosed herein.

[0359] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this disclosure to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations therein.

Claims

1. A communication system, wherein, The communication system includes: The first network function is used to implement communication connectivity. The second network function is communicatively connected to at least one of the first network functions and is used to implement at least one of the following first functions: data management function, data analysis function, artificial intelligence / machine learning (AI / ML) model-related functions, computing function, and resource collaborative control function.

2. The system according to claim 1, wherein, The communication system is a wireless access network system; The first network function is communicatively connected to the user plane function and control plane function of the core network, respectively. The second network function connects to the core network through a service-oriented interface.

3. The system according to claim 1 or 2, wherein, Any of the first functions includes at least one of the following: control plane function, data plane function, and computation plane function; The control plane function is used for control and management; the data plane function is used for at least one of data processing, data acquisition, data transmission, and data storage; and the computing plane function is used for at least one of computing task processing, computing resource awareness, computing resource management or scheduling.

4. The system according to any one of claims 1-3, wherein, The first network function also includes at least one second function: data acquisition and processing function, AI / ML model inference function; The first network function includes at least one of the following: communication connection function related logical network functions or nodes.

5. The system according to any one of claims 1-4, wherein, The first network function and the second network function are deployed independently.

6. A communication method, wherein, Applied to a second network function, the method includes: Establish a communication connection with the first network function; Interact with the first network function and / or the core network; The second network function has at least one of the following first functions: data management function, data analysis function, artificial intelligence / machine learning (AI / ML) model related functions, computing function, and resource collaborative control function.

7. The method according to claim 6, wherein, Establishing a communication connection with the first network function includes: Receive a first message from the first network function, the first message being used to request the establishment of a communication interface; Send a second message to the first network function.

8. The method according to claim 7, wherein, The first message carries first information about the first network function; the first information is used to indicate the capabilities supported by the first network function. And / or, The second message carries second information about the second network function; the second information is used to indicate the capabilities supported by the second network function.

9. The method according to claim 6, wherein, The method further includes: Send a third message to the first network function and receive a fourth message from the first network function; And / or, Receive a third message from the first network function and send a fourth message to the first network function; The third message is used to query or subscribe to the capabilities of the message recipient; the fourth message carries third information about the message recipient; the third information includes: currently supported capabilities, and / or capability change information.

10. The method according to any one of claims 6-9, wherein, The interaction with the first network function and / or core network includes: Send a fifth message to the core network element, the fifth message being used for capability registration; Receive the sixth message from the core network element; The core network elements include at least one of the following: network elements related to service discovery capability, network elements related to computing capability, and network elements related to data capability.

11. The method according to any one of claims 6-10, wherein, The interaction with the first network function and / or core network includes: Receive a seventh message from the first network function, the seventh message being used to request AI / ML model processing, the AI / ML model processing including at least one of the following: model training, model testing, and model inference; Based on the seventh message, model processing is performed to obtain the first processing result; Send the first processing result to the first network function.

12. The method according to claim 11, wherein, The seventh message is triggered by the first network function based on a request from the terminal.

13. The method according to any one of claims 6-12, wherein, The interaction with the first network function and / or core network includes: Identify at least one target network function among the first network functions; Iterative processing is performed based on the functions of each target network until the preset requirements are met; wherein any round of iterative processing includes: Send an eighth message to each target network function, the eighth message being used to request joint model processing; Receive the second processing results from each target network function; Based on the results of each of the second processing steps, the models are aggregated to obtain a global model; Determine whether the global model meets the preset requirements.

14. The method according to any one of claims 6-13, wherein, The interaction with the first network function and / or core network includes: Receive a ninth message from the first network function or the core network, the ninth message being used to request computing services to be provided to the terminal; Allocate computing resources to the computing tasks of the terminal, and use the computing resources to process the computing tasks to obtain a third processing result; The third processing result is sent to the first network function or the core network.

15. A communication method applied to a first network function, the method comprising: Establish a communication connection with the second network function; Interact with the second network function based on the communication connection; The second network function has at least one of the following first functions: data management function, data analysis function, artificial intelligence / machine learning (AI / ML) model related functions, computing function, and resource collaborative control function.

16. The method according to claim 15, wherein, The establishment of a communication connection with the second network function includes: Send a first message to the second network function, the first message being used to request the establishment of a communication interface; Receive a second message from the first network function.

17. The method according to claim 16, wherein, The first message carries first information about the first network function; the first information is used to indicate the capabilities supported by the first network function. And / or, The second message carries second information about the second network function; the second information is used to indicate the capabilities supported by the second network function.

18. The method according to claim 16, wherein, When the second message carries second information about the second network function, the method further includes: Store relevant information for the second network function; the relevant information may be all or part of the second information.

19. The method according to claim 15, wherein, The method further includes: Receive a third message from the second network function and send a fourth message to the second network function; And / or, Send a third message to the second network function and receive a fourth message from the second network function; The third message is used to query or subscribe to the capabilities of the message recipient; the fourth message carries third information about the message recipient; the third information includes: currently supported capabilities, and / or capability change information.

20. The method according to any one of claims 15-19, wherein, The interaction with the second network function based on the communication connection includes: Send a seventh message to the second network function, the seventh message being used to request AI / ML model processing, the AI / ML model processing including at least one of the following: model training, model testing, and model inference; Receive the first processing result from the second network function.

21. The method according to claim 20, wherein, The method further includes: Before sending the seventh message to the second network function, a tenth message is received from the terminal, the tenth message being used to request the second network function to perform AI / ML model processing; The first processing result is sent to the terminal.

22. The method according to any one of claims 15-21, wherein, The interaction with the second network function based on the communication connection includes: Receive an eighth message from the second network function, the eighth message being used to request joint model processing; Perform local model processing to obtain the second processing result; Send the second processing result to the second network function.

23. The method according to any one of claims 15-22, wherein, The method further includes: Receive an eleventh message from the terminal, the eleventh message being used to request the second network function to provide computing services to the terminal; Send a ninth message to the second network function, the ninth message being used to request computing services to the terminal; Receive the third processing result from the second network function; The third processing result is sent to the terminal.

24. A network function, wherein, include: Establish a unit, configured to establish a communication connection with the first network function; The communication unit is configured to interact with the first network function and / or the core network; wherein the second network function has at least one of the following first functions: data management function, data analysis function, artificial intelligence / machine learning AI / ML model related functions, computing function, and resource collaborative control function.

25. A network function, wherein, include: Establish a unit, configured to establish a communication connection with the second network function; The communication unit is configured to interact with the second network function based on the communication connection; wherein the second network function has at least one of the following first functions: data management function, data analysis function, artificial intelligence / machine learning (AI / ML) model-related functions, computing function, and resource collaborative control function.

26. An electronic device comprising a memory, a processor, and a computer program stored in the memory, wherein, The processor executes the computer program to implement the method according to any one of claims 6-23.

27. A computer-readable storage medium having a computer program / instructions stored thereon, wherein, When the computer program / instructions are executed by the processor, they implement the method described in any one of claims 6-23.

28. A computer program product comprising a computer program / instructions, wherein, When the computer program / instructions are executed by the processor, they implement the method described in any one of claims 6-23.

Images