A network capability opening method and system, electronic equipment and storage medium

Abstract

This application provides a method, system, electronic device, and storage medium for network capability opening, relating to the field of communication technology, and is used to reduce the cost of using network capability opening interfaces. The method includes: receiving a natural language request message sent by a third-party application; the request message requests to invoke at least one application programming interface (API) corresponding to the Network Capability Opening Function (NEF); analyzing the request message using a large-scale model and outputting an invocation scheme; the invocation scheme includes at least the invocation order of at least one API to be invoked; the large-scale model accesses all APIs supported by NEF; and invoking at least one API corresponding to NEF sequentially according to the invocation order, and sending a response message to the third-party application.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a method, system, electronic device, and storage medium for opening up network capabilities. Background Technology

[0002] Currently, network capability openness refers to the mobile communication network exposing various internal functions to external third-party applications in a secure and controllable manner through standardized application programming interfaces (APIs), enabling these applications to call them on demand.

[0003] In related technologies, the standards development and release by organizations such as the 3rd Generation Partnership Project (3GPP) take time, while market demands evolve even faster. This leads to operators and equipment manufacturers potentially releasing non-standard, proprietary APIs ahead of schedule, resulting in inconsistencies in the open network capability interfaces between different operator networks. However, third-party application developers may need to maintain different software code for different operators or devices to adapt to these interfaces, increasing the cost of using open network capability interfaces and thus creating a technical problem of high usage costs. Summary of the Invention

[0004] This application provides a method, system, electronic device, and storage medium for opening up network capabilities, which reduces the cost of using network capability opening interfaces.

[0005] In a first aspect, this application provides a method for opening up network capabilities, the method comprising: receiving a request message in natural language sent by a third-party application; the request message being used to request the invocation of at least one application programming interface (API) corresponding to the network capability opening function NEF; invoking a large model to analyze the request message and output an invocation scheme; the invocation scheme including at least the invocation order of at least one API to be invoked; the large model accessing all APIs supported by NEF; invoking at least one API corresponding to NEF in sequence according to the invocation order, and sending a response message to the third-party application.

[0006] The technical solution provided in this application brings at least the following beneficial effects: By adding a network capability open agent based on the intelligent agent model, it communicates with third-party applications externally using an interface similar to natural language, and internally interfaces with the API interfaces defined by the original NEF, thus realizing the intelligence of the network capability open interface. In other words, after receiving a natural language request message from a third-party application, the network capability open agent in this application parses the request message using its internal large model, decomposes the request message according to business logic into at least one API call corresponding to NEF, calls at least one API corresponding to NEF in the order of call, and sends a response message to the third-party application. This solves the technical problem of high usage cost of the network capability open interface and achieves the technical effect of reducing the usage cost of the network capability open interface.

[0007] One possible implementation involves calling a large model to analyze the request message and output a calling scheme, including: obtaining target interface knowledge matching the request message from the NEF interface knowledge base; the NEF interface knowledge base stores at least NEF interface documents collected according to the NEF interface version; calling an embedded model to synthesize the request message and target interface knowledge to obtain prompt words; and calling the large model to output a calling scheme based on the prompt words.

[0008] Another possible implementation scheme includes a call command for each API, which calls at least one API corresponding to NEF in the order of call. This includes: when the format type of the call command is a standardized communication interface format supported by the large model, converting the format type of the call command from the standardized communication interface format to the API format corresponding to NEF; and under the API format corresponding to NEF, calling at least one API corresponding to NEF in the order of call.

[0009] Another possible implementation scheme includes a call command for each API, which calls at least one API corresponding to NEF in the order of call, including: when the format type of the call command is the API format corresponding to NEF, calling at least one API corresponding to NEF in the order of call.

[0010] Another possible implementation method includes: in response to an event triggered by a third-party application subscription, obtaining the event message corresponding to the event; calling a large model to process the event message and generate a corresponding response message; and sending the response message to the third-party application.

[0011] Secondly, this application provides a network capability open system, which includes: a third-party message interface module for receiving natural language request messages sent by a third-party application; the request message is used to request the invocation of at least one application programming interface (API) corresponding to the network capability open function NEF; a large model for analyzing the request message and outputting an invocation scheme; the invocation scheme includes at least the invocation order of at least one API to be invoked; the large model is connected to all APIs supported by NEF; a NEF interface module for invoking at least one API corresponding to NEF in sequence according to the invocation order; and the third-party message interface module is also used to send response messages to the third-party application.

[0012] In one possible implementation, the system further includes: a NEF interface knowledge module for storing NEF interface documents collected according to the NEF interface version; and a prompt word processing module for calling the embedded model to synthesize the request message and the target interface knowledge obtained from the NEF interface knowledge module that matches the request message to obtain prompt words adapted to the input of the large model.

[0013] Another possible implementation, the invocation scheme also includes invocation commands for each API. The NEF interface module has protocol conversion capabilities and is also used to convert the format type of the invocation commands from the standardized communication interface format supported by the large model to the corresponding API format of NEF.

[0014] Thirdly, this application provides a network capability opening device, which includes: a receiving module for receiving a natural language request message sent by a third-party application; the request message is for requesting to call at least one application programming interface (API) corresponding to the Network Capability Opening Function (NEF); an analysis module for calling a large model to analyze the request message and output a calling scheme; the calling scheme includes at least the calling order of at least one API to be called; the large model is connected to all APIs supported by NEF; and a calling module for calling at least one API corresponding to NEF in sequence according to the calling order and sending a response message to the third-party application.

[0015] Fourthly, this application provides an electronic device comprising: a processor and a memory; the memory storing processor-executable instructions; when the processor is configured to execute the instructions, causing the electronic device to implement the method of the first aspect described above.

[0016] Fifthly, this application provides a computer-readable storage medium comprising: computer software instructions; which, when executed in an electronic device, cause the electronic device to implement the method described in the first aspect.

[0017] Sixthly, this application provides a computer program product comprising a computer program; when the computer program is run in an electronic device, it causes the electronic device to implement the method described in the first aspect.

[0018] The beneficial effects of the second to sixth aspects mentioned above are described in the corresponding description of the first aspect and will not be repeated here. Attached Figure Description

[0019] Figure 1 This application provides an architectural diagram of a network capability open system. Figure 2 A flowchart of a network capability opening method provided in this application; Figure 3 A schematic diagram of a NEF interface provided for related technologies; Figure 4 A schematic diagram illustrating the location of a network capability open agent based on an intelligent agent model in the NEF interface, as provided in this application; Figure 5 A schematic diagram of the internal functional modules of a network capability open agent based on the intelligent agent model provided in this application; Figure 6 A flowchart illustrating the interaction between a network capability open agent, a third-party app, and NEF based on an agent model, provided for this application; Figure 7 A schematic diagram of the internal functional modules of another network capability open agent based on the intelligent agent model provided in this application; Figure 8 A schematic diagram of the composition of a network capability opening device provided in this application; Figure 9 This is a schematic diagram of the composition of an electronic device provided in this application. Detailed Implementation

[0020] The following section will describe in detail a method for opening up network capabilities provided in this application, with reference to the accompanying drawings.

[0021] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0022] The terms "first" and "second," etc., used in the specification and drawings of this application are used to distinguish different objects or to distinguish different treatments of the same object, rather than to describe a specific order of objects.

[0023] Furthermore, the terms "comprising" and "having," and any variations thereof, used in the description of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the steps or units listed, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus.

[0024] It should be noted that in the embodiments of this application, the words "exemplarily" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design scheme described as "exemplarily" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the words "exemplarily" or "for example" is intended to present the relevant concepts in a specific manner.

[0025] To facilitate a clear description of the technical solutions of the embodiments of this application, the terms "first" and "second" are used in the embodiments of this application to distinguish the same or similar items with essentially the same function and effect. Those skilled in the art can understand that the terms "first" and "second" are not intended to limit the quantity or execution order.

[0026] In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0027] Currently, the interfaces for open network capabilities are quite complex. For third-party application developers using these interfaces, the learning curve is steep due to their unfamiliarity with telecommunications network concepts, making widespread adoption difficult. With the further development of mobile communication networks, network functions are becoming increasingly complex. The six major scenarios of 6G networks include: immersive communication, ultra-large-scale connectivity, ultra-reliable low-latency communication, ubiquitous connectivity, integrated communication and AI, and integrated communication and sensing. Therefore, it is foreseeable that open network capability interfaces will become even more complex, and the logical relationships between these interfaces required to complete business processes will be more difficult to understand.

[0028] Furthermore, the development and release of standards by organizations such as 3GPP takes time, while market demands evolve much faster. This can lead to operators and equipment manufacturers releasing non-standard, proprietary APIs ahead of schedule, resulting in inconsistencies in the open network capability interfaces between different operator networks and creating a fragmented ecosystem. Third-party application developers may need to maintain different software code for different operators or devices to adapt to these interfaces, increasing the cost of using open network capability interfaces and thus creating a technical problem of high usage costs.

[0029] Based on this, this application adds a network capability open proxy based on an intelligent agent model. This proxy communicates with third-party applications externally using an interface similar to natural language, and internally interfaces with the API interfaces defined by the original NEF, thus achieving intelligent network capability open interfaces. In other words, after receiving a natural language request message from a third-party application, the network capability open proxy in this application parses the request message using its internal large model. It decomposes the request message according to business logic into at least one API call corresponding to NEF, calls at least one corresponding API in the NEF API in the order of call, and sends a response message to the third-party application. This solves the technical problem of high usage costs of network capability open interfaces and achieves the technical effect of reducing the usage costs of network capability open interfaces.

[0030] The embodiments provided in this application will now be described in detail with reference to the accompanying drawings.

[0031] The network capability opening method provided in this application can be applied to, for example... Figure 1 In the system architecture shown. Figure 1 This application provides an architectural diagram of a network capability open system, such as... Figure 1 As shown, this system can be applied to network capability open agents (hereinafter referred to as agents) based on the agent model. These agents communicate with third-party applications externally (i.e., through the northbound interface) using an interface similar to natural language; internally (i.e., through the southbound interface), they interface with the API interfaces defined by the original NEF. The system may include: a third-party message interface module 101, a large model 102, and a NEF interface module 103.

[0032] The third-party message interface module 101 is used to receive natural language request messages sent by third-party applications (APPs); the request messages are used to request the invocation of at least one application programming interface (API) corresponding to the Network Exposure Function (NEF). The third-party message interface module 101 is also used to send response messages to the third-party applications that invoke the APIs.

[0033] The third-party message interface module is responsible for managing connections with third-party apps, receiving request messages from third-party apps, sending response messages for API calls or subscribed event messages to third-party apps, etc.

[0034] Large Model 102 is used to analyze request messages and output a calling scheme; the calling scheme includes at least the calling order of at least one API to be called; the large model connects to all APIs supported by NEF.

[0035] The large model, which can be an open-source model or a vendor-proprietary model, is responsible for inference based on user input. That is, it transforms user-input request messages into a series of corresponding API calls.

[0036] NEF interface module 103 is used to call at least one API corresponding to NEF in the order of invocation.

[0037] The NEF interface module is responsible for connection management with NEF, receiving response messages from NEF for API calls, and sending call messages to NEF (i.e., calling the corresponding APIs to NEF). The NEF interface module also manages the connection with the large model, connecting the large model to all APIs currently supported by NEF, receiving the call schemes output by the large model, and calling the corresponding APIs to NEF according to the call order.

[0038] In some embodiments, the system further includes: NEF interface knowledge module 104 and prompt word processing module 105.

[0039] NEF Interface Knowledge Module 104 is used to store NEF interface documents collected according to the NEF interface version.

[0040] The NEF Interface Knowledge Module can also be called the NEF Interface Knowledge Base or the Capability Open Interface Knowledge Module. The NEF Interface Knowledge Module collects specialized knowledge such as documentation, API header files, and example code for network capability open interfaces according to NEF interface versions, providing a knowledge base for embedded models.

[0041] The prompt word processing module 105 is used to call the embedded model to synthesize the request message and the target interface knowledge that matches the request message obtained from the NEF interface knowledge module to obtain prompt words that are adapted to the input of the large model.

[0042] The third-party message interface module can forward request messages received from third-party apps to the prompt word processing module. The prompt word processing module handles request messages input by third-party apps. Through an embedded model, it synthesizes the user-input request message with the target interface knowledge corresponding to the request message from the capability open interface knowledge module to obtain appropriate prompt words, which serve as input to the larger model.

[0043] In some embodiments, the invocation scheme also includes invocation commands for each API. The NEF interface module has a protocol conversion function and is also used to convert the format type of the invocation commands from the standardized communication interface format supported by the large model to the API format corresponding to NEF.

[0044] In cases where the large model uses industry-standard intelligent agent communication protocols, such as the Model Context Protocol (MCP), the NEF interface module can convert MCP interface messages into corresponding NEF API messages.

[0045] It should be noted that the system architecture described in the embodiments of this application is for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and does not constitute a limitation on the technical solutions provided in the embodiments of this application. As those skilled in the art will know, with the evolution of system architecture, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0046] Figure 2 The flowchart illustrates a network capability opening method provided in this application, where the implementing entity can be the aforementioned network capability opening agent based on the agent model. For example... Figure 2 As shown, the method specifically includes the following steps S201~S203: S201. Receive a request message in natural language sent by a third-party application.

[0047] The request message is used to request the invocation of at least one application programming interface (API) corresponding to the Network Capability Opening Function (NEF).

[0048] For example, after a connection is established between the third-party message interface module and the third-party APP, the module can receive request messages sent by the third-party APP using natural language.

[0049] S202. Analyze the request message using the large model and output the invocation plan.

[0050] The invocation scheme must include at least the invocation order of at least one API to be invoked. For example, the invocation scheme may include at least a list of invocation steps (arranged in order), step numbers, the name of at least one API to be invoked, the invocation command for each API, request parameters, execution conditions, and the response template finally returned to the third-party app. This is merely an example and does not impose specific limitations on the content of the invocation scheme.

[0051] For example, the large model is invoked to parse the request message, decompose the request message into one or more API interface calls of the original NEF according to the business logic, determine the dependencies and calling order between these API interfaces to be called, and generate a structured calling scheme.

[0052] In some embodiments, the large model can access all APIs supported by NEF. For example, you can register all APIs supported by NEF with the large model.

[0053] It should be noted that the network capability open agent based on the agent model establishes a connection with NEF upon startup. The NEF interface module inside the agent exchanges messages with NEF. At this time, the NEF interface module registers all APIs currently supported by NEF with the large model.

[0054] In some embodiments, if an adjustment to the API functionality supported by NEF is detected, the APIs registered in the large model are updated according to the adjusted API functionality.

[0055] For example, when NEF adds, modifies, or removes certain API capabilities, the NEF interface module detects the configuration or functional changes and triggers an incremental or full update mechanism to resubmit the latest API metadata to the large model. Upon receiving the update notification, the large model can synchronously refresh the API information in its local cache or registry, ensuring that subsequent calls are based on the latest interface definition, thereby maintaining service consistency and availability.

[0056] In some embodiments, the large model is invoked to analyze the request message and output an invocation scheme, including: obtaining target interface knowledge matching the request message from the NEF interface knowledge base; invoking the embedded model to synthesize the request message and target interface knowledge to obtain prompt words; and invoking the large model to output an invocation scheme based on the prompt words.

[0057] The NEF interface knowledge base can also be called the Network Capability Open Interface Professional Knowledge Base, Capability Open Interface Knowledge Module, etc. The NEF interface knowledge base stores at least the NEF interface documentation collected according to the NEF interface version. In addition, the NEF interface knowledge base also stores specialized knowledge such as API header files and example code for calls, which can provide a knowledge base for embedded models.

[0058] For example, after receiving a request message, the prompt word processing module can first preprocess the request message (e.g., clean the text, extract key entities, convert the format, etc.), and then use an embedded model to encode the preprocessed request message into a vector, and then perform a similarity search in the network capability open interface professional knowledge base. For example, it can calculate the similarity (e.g., cosine similarity) between the vector corresponding to the request message and all interface knowledge in the knowledge base, and return the top K (Top-K) most relevant interface knowledge to constitute the target interface knowledge.

[0059] Furthermore, the request message can be combined with the retrieved target interface knowledge to obtain suitable prompt words. These generated prompt words are then used as input to a larger model, which generates output containing API calls based on the prompt words.

[0060] This application introduces an embedded model and combines it with a professional knowledge base of network capability open interfaces, enabling the large model to accurately understand the network capability context when processing request messages from third-party apps. This significantly improves the accuracy of intent recognition, thereby enhancing the rationality of API call planning and the adaptability to business scenarios.

[0061] S203. In accordance with the calling order, call at least one API corresponding to NEF in sequence and send a response message to the third-party application.

[0062] In some embodiments, the invocation scheme further includes an invocation command for each API, which invokes at least one API corresponding to NEF in the order of invocation, including: when the format type of the invocation command is a standardized communication interface format supported by the large model, converting the format type of the invocation command from the standardized communication interface format to the API format corresponding to NEF; and in the API format corresponding to NEF, invoking at least one API corresponding to NEF in the order of invocation.

[0063] The standardized communication interface format can include de facto standards for agent communication (e.g., MCP). The NEF interface module has protocol conversion capabilities. When the large model uses industry-standard interfaces such as MCP, the NEF interface module can convert MCP interface messages into corresponding NEF API messages.

[0064] For example, after the large model generates output containing API calls based on the prompt words and sequentially calls the APIs registered by the NEF interface module, the NEF interface module receives the calling scheme of the large model, performs protocol conversion on the calling command of each API in the calling scheme, calls the corresponding API interface of NEF, and receives the API call response message sent by NEF, and then sends the response message to the third-party APP.

[0065] In some embodiments, the invocation scheme further includes an invocation command for each API, which invokes at least one API corresponding to NEF in the order of invocation, including: when the format type of the invocation command is the API format corresponding to NEF, invoking at least one API corresponding to NEF in the order of invocation.

[0066] For example, when the format type of the calling command is the API format corresponding to NEF, there is no need to perform protocol conversion on the calling command; the calling command can be used to call at least one API corresponding to NEF in sequence.

[0067] When the format of the calling command is the API format corresponding to NEF, this application does not require protocol conversion of the calling command; however, when the format of the calling command is a standardized communication interface format supported by the large model (e.g., MCP), by performing protocol conversion on the format of the calling command, calling the API interface corresponding to NEF and processing the response, efficient and standardized docking between the large model and NEF can be achieved, enabling the large model to call network capabilities on demand, thereby realizing intelligent interface calling.

[0068] In some embodiments, the method further includes: in response to an event triggered by a third-party application subscription, obtaining an event message corresponding to the event; calling a large model to process the event message and generate a corresponding response message; and sending the response message to the third-party application.

[0069] Among them, the events subscribed by third-party apps can include at least user behavior events (such as login, user operation behavior, user information update, etc.), business status change events (such as status update, account change, etc.), system or device events (such as device online, fault alarm, etc.), communication and messaging events (such as receiving new messages, message sending failure, etc.), and time or periodic events (such as scheduled task triggering, expiration reminder, etc.). This is only an example and no specific restrictions are imposed on the events subscribed to.

[0070] For example, when an event subscribed to by a third-party app is triggered, it can receive an event message sent by NEF. NEF can forward the event message to the NEF interface module, which in turn forwards it to the response and subscription processing module within the proxy. The response and subscription processing module can identify and classify the message type of the event message to determine the subsequent processing logic.

[0071] For example, key fields (such as event type, timestamp, user identifier, network parameters, etc.) can be extracted from event messages and combined with historical interaction records or session context to convert unstructured or semi-structured data into a structured input format, facilitating understanding by large models. Based on message type and business scenario, prompts containing task instructions, contextual information, constraints, and output format requirements are dynamically constructed. These constructed prompts are then input into the large model, triggering its inference process. The large model generates natural language or structured response messages that conform to semantic, logical, and business rules based on built-in knowledge, training data, and context. Further post-processing of the response messages generated by the large model can be performed, such as content compliance checks, format standardization, and semantic consistency verification. Finally, the processed response messages are sent to a third-party message interface module, which then forwards them to a third-party app.

[0072] This application achieves the technical effect of providing intelligent, personalized, and real-time information services driven by the synergy of network capabilities and large models by transforming real-time events and service responses on the network side into user-oriented natural language intelligent messages and accurately pushing them to third-party apps.

[0073] The technical solutions provided by the above embodiments bring at least the following beneficial effects. The network capability opening method provided in this application, by adding a network capability opening agent based on the intelligent agent model, communicates with third-party applications externally using an interface similar to natural language, and internally interfaces with the API interfaces defined by the original NEF, thereby realizing the intelligence of the network capability opening interface. That is to say, after receiving a natural language request message sent by a third-party application, the network capability opening agent in this application parses the request message through an internal large model, decomposes the request message into at least one API call corresponding to NEF according to business logic, calls at least one API corresponding to NEF in the order of call, and sends a response message to the third-party application. This solves the technical problem of high usage cost of the network capability opening interface and achieves the technical effect of reducing the usage cost of the network capability opening interface.

[0074] The following describes a specific embodiment of the network capability opening method of this application.

[0075] Currently, network capability openness refers to the mobile communication network exposing various internal functions (such as connection management, traffic policies, location, billing, etc.) to external third-party applications (including application developers, vertical industry platforms, other operators, or end users) in a secure and controllable manner through standardized API interfaces. This enables external third-party applications to call upon these functions as needed, thereby rapidly incubating new businesses and new business models, and empowering various industries to carry out business innovation and digital transformation.

[0076] Current methods for opening up capabilities in mobile communication networks mainly include capability opening as defined by 3GPP, capability opening through Open Radio Access Network (O-RAN), and Open Gateway as defined by the Global System for Mobile Communications Association (GSMA).

[0077] Among them, 3GPP defined a network capability open framework in the 5G standard. This network capability open framework can be divided into two categories: one is to expose core network capabilities to external applications through the 5G NEF via the network northbound interface; the other is to expose them to various vertical applications through enabling abstraction layer APIs (such as SEAL, EDGEAPP, V2XAPP, etc.), which is a parallel abstraction layer provided on top of the abstract functions already provided by the network northbound interface, and can also be exposed through the Common API Framework (CAPIF). The O-RAN Alliance proposed near real-time Radio Access Network Intelligent Controller (RIC) and non-real-time RIC solutions, supporting third-party applications to call radio-related capabilities through RIC applications (such as xApps and rApps). The O-RAN Alliance has completed research on providing computing services and integrated computing and communication services to the Radio Access Network (O-RAN) and is developing relevant standards for integrated computing and communication services. The GSMA proposed Open Gateway, which aims to facilitate faster access to operator networks for developers and cloud service providers, transforming telecommunications network capabilities into standardized services, with open interfaces implemented through the core network NEF function.

[0078] Figure 3 A schematic diagram of a NEF interface provided for related technologies, such as... Figure 3 As shown, the entire structure is divided into three layers: the third-party application layer, the NEF layer, and the network function layer, which are connected through northbound and southbound interfaces. The third-party application layer includes multiple third-party applications (Application Functions, AFs), for example, AF1 represents the first application function, and AFn represents the nth application function. Figure 3 It contains two independent NEF instances, each handling different API requests. For example, the left NEF contains API1 and API2, while the right NEF contains API3 through APIn. The Network Function Layer (NF) contains multiple Network Functions (NFs), such as NF1, NF2, and NFn. Each NEF connects to the upper-layer third-party application (AF) via a northbound interface and to the lower-layer network function (NF) via a southbound interface.

[0079] Currently, the interfaces for open network capabilities are quite complex. For third-party application developers using these interfaces, the learning curve is steep due to their unfamiliarity with telecommunications network concepts, making adoption difficult. As mobile communication networks continue to develop, their functions become increasingly complex. The six key scenarios of 6G networks include: immersive communication, massive MIMO, ultra-reliable low-latency communication, ubiquitous connectivity, integrated communication and AI, and integrated communication and sensing. Therefore, it is foreseeable that open network capability interfaces will become even more complex, and the logical relationships between interfaces required to complete business processes will be more difficult to understand.

[0080] Furthermore, the development and release of standards by organizations such as 3GPP takes time, while market demands evolve much faster. This can lead to operators and equipment manufacturers releasing non-standard, proprietary APIs ahead of schedule, resulting in inconsistent open capability interfaces across different operator networks and creating a fragmented ecosystem. Third-party application developers may need to maintain different software code for different operators or devices to adapt to these interfaces, increasing the cost of using open capability interfaces.

[0081] Based on this, considering that current network capability open interfaces lack intelligent capabilities, and with the rapid development of AI technology, network capability open interfaces can also evolve towards intelligence. This application proposes a mobile communication network capability open method based on the Agent model. By using an Agent as a capability open proxy, it uses an interface that approximates natural language and provides business-level calling capabilities. That is, the intelligent agent parses user input and, according to business needs, sequentially calls multiple interfaces according to the internal logic flow of the mobile network, realizing intelligent calling of multiple APIs, thereby achieving intelligent network capability open interfaces. Simultaneously, this method also shields the complexity and internal changes of the interfaces; upgrades and changes to internal interfaces are learned and adapted by the intelligent agent within the proxy, eliminating the need for third-party applications to be aware of or perform adaptation development work, thus reducing the difficulty of calling by third-party applications.

[0082] Capability sharing based on the agent model can adopt the agent model. Figure 4 A schematic diagram illustrating the location of a network capability open agent based on an intelligent agent model in the NEF interface, as provided in this application, is shown below. Figure 4 As shown, this application is in Figure 3Building upon the existing framework, a new capability open proxy based on the intelligent agent model has been added. This agent-based proxy connects to upper-layer third-party applications (AFs) via a new northbound interface. Each NEF connects to the agent-based proxy via its original northbound interface and to the lower-layer network functions (NFs) via a southbound interface. In other words, the proxy communicates with third-party applications (AFs) externally (northbound) using an interface resembling natural language; internally (southbound), it interfaces with the API interfaces defined by the original NEFs. Internally, the proxy analyzes the request messages sent by the AFs, decomposes them according to business logic into one or more API calls to the original NEFs, executes them sequentially, and ultimately returns the results to the user.

[0083] Figure 5 A schematic diagram of the internal functional modules of a network capability open agent based on the intelligent agent model provided in this application is shown below. Figure 5 As shown, the internal functional modules of the agent may include at least: a third-party message interface module 501, a prompt word processing module 502, a capability open interface knowledge module 503, a NEF interface module 504, a response and subscription processing module 505, a historical data module 506, a large model 507, and an embedded model 508.

[0084] The third-party message interface module 501 is responsible for connection management with the third-party app, as well as receiving and sending messages. It can forward the content of requests sent by the third-party app to the prompt word processing module and send the final response message or event message to the third-party app.

[0085] The prompt word processing module 502 processes the input from third-party apps and synthesizes prompt words from user input and open capability interface knowledge through an embedded model, which are then used as input to the large model.

[0086] The Capability Open Interface Knowledge Module 503 collects specialized knowledge such as documentation (i.e., NEF interface documentation), API header files, and example code for calling the capability open interface according to the NEF interface version, providing a knowledge base for embedded models.

[0087] The NEF interface module 504 is responsible for connection management, message receiving and sending, and connection management with the NEF and the large model, as well as protocol conversion. It can register the API interfaces currently supported by NEF with the large model, receive the calling schemes output by the large model, process them sequentially to call the corresponding APIs to NEF, and finally send the returned response messages and / or subscribed event messages to the response and subscription processing modules. The NEF interface module 504 also has protocol conversion capabilities. When the large model uses industry-standard intelligent agent communication protocols (such as MCP), the NEF interface module 504 can convert MCP interface messages into corresponding NEF API messages.

[0088] The response and subscription processing module 505 is used to collect response messages and event messages returned by NEF. If necessary, it can send a request to the large model, which will then generate the format and content of a message to be sent back to the third-party app and reply to the app.

[0089] The historical data module 506 is used to collect historical data processed by the large model in order to further train or fine-tune the large model.

[0090] The large model 507 and the embedded model 508 can adopt open-source large models or vendor-owned models, and are responsible for reasoning from user input and converting user input into a series of corresponding API call behaviors.

[0091] The processing flow of the internal functional modules of the agent-based network capability open proxy can include: Upon startup, the agent-based network capability open proxy establishes a connection with NEF; the NEF interface module and NEF exchange messages; the NEF interface module registers the API interfaces currently supported by NEF with the large model; the third-party message interface module establishes a connection with the third-party application and receives request messages from the third-party application; the third-party message interface module sends the content of the request message to the prompt word processing module; the prompt word processing module uses an embedded model to query capability open interface professional knowledge based on the content of the request message, that is, it queries the capability open interface professional knowledge to find the interface professional knowledge that matches the request message, and synthesizes the request message with the corresponding interface professional knowledge to obtain a prompt word suitable for processing by the large model; the prompt word processing module sends the prompt word to the large model.

[0092] The large model generates output containing API calls based on the prompt words and sequentially calls the APIs registered by the NEF interface module. The NEF interface module receives the large model's call request (i.e., the call scheme), performs protocol conversion on the request, calls the corresponding NEF API interface, and receives the response message and subscribed event messages from NEF. The NEF interface module forwards the received response message and subscribed event messages to the response and subscription processing module. The response and subscription processing module calls the large model to generate a reply message. The response and subscription processing module sends the generated reply message to the third-party message interface module, which then sends the reply message to the third-party app.

[0093] Figure 6 This application provides a flowchart illustrating the interaction between a network capability open agent, a third-party app, and NEF based on an intelligent agent model. The executing entity can be the agent. Figure 6 As shown, the message flow between the agent, third-party app, and NEF may include the following steps: Step S601: The agent receives the request message sent by the third-party APP.

[0094] Once the third-party message interface module within the agent establishes a connection with the third-party APP, it can receive request messages sent by the third-party APP.

[0095] In step S602, the agent sequentially calls the corresponding APIs of NEF based on the request message.

[0096] Upon receiving a request message, the third-party message interface module can send the request message to the prompt word processing module. The prompt word processing module uses an embedded module to query the corresponding interface knowledge from the capability open interface knowledge module, and then combines the request message with the corresponding interface knowledge to obtain prompt words suitable for the large model input. The large model outputs output containing API calls based on the prompt words, and sequentially calls the APIs registered by the NEF interface module. After receiving the call request from the large model, the NEF interface module can perform protocol conversion on the call request and then call the corresponding NEF API interface.

[0097] In other words, the proxy internally parses the request message and decomposes it into one or more API calls from the original NEF according to business logic. Then, it sequentially calls the corresponding API1, API2 to APIn of the NEF for processing according to the process.

[0098] In step S603, the agent sends the response message to the third-party app.

[0099] The NEF interface module receives response messages from NEF and forwards them to the response and subscription processing module. The response and subscription processing module then forwards the response messages to the third-party message interface module, which in turn sends them to the third-party application.

[0100] Step S604: The agent receives the event message sent by NEF.

[0101] When an event subscribed to by a third-party app occurs, the NEF interface module can receive the event message sent by NEF and forward it to the response and subscription processing module. The response and subscription processing module can then send the event message to the large model, which will infer the event message and generate the corresponding response information.

[0102] In step S605, the agent forwards the reply message corresponding to the event message to the third-party APP.

[0103] The large model sends the generated response information to the response and subscription processing module. The response and subscription processing module then forwards the event message to the third-party message interface module, which in turn sends the event message to the third-party app.

[0104] It should be noted that the way the agent interacts with NEF can be either by converting the NEF protocol defined by 3GPP internally into the de facto standard for intelligent agent communication in the industry, such as MCP; or the NEF can directly use protocols such as MCP to communicate with the agent, avoiding the protocol conversion work inside the agent.

[0105] Figure 7 A schematic diagram of the internal functional modules of another network capability open agent based on the intelligent agent model provided in this application is shown below. Figure 7 As shown, the internal functional modules of this agent-based capability openness agent can include at least: a third-party message interface module 701, a prompt word processing module 702, a capability openness interface knowledge module 703, NEF 704, a response and subscription processing module 705, a historical data module 706, a large model 707, and an embedded model 708. Besides adopting an agent-based approach, network capability openness based on the agent model can also be designed in a hybrid manner. That is, the NEF can be supplemented with the aforementioned prompt word processing module, capability openness interface knowledge module, historical data module, third-party interface module, large model, and embedded model functional modules, while the NEF interface module can be directly replaced by the NEF, thus realizing an intelligent NEF based on the agent model. It should be noted that... Figure 7 The functions of each module and the processing flow between modules are the same. Figure 5 This will not be elaborated upon here.

[0106] In summary, this application achieves intelligent network capability opening functionality by using an intelligent agent model in mobile communication networks. The network capability opening agent based on the intelligent agent model uses an interface that approximates natural language. This interface is analyzed by large and embedded models and parsed into API call logic that meets business requirements, enabling intelligent API calls. Simultaneously, it shields the complexity and internal changes of the interface; upgrades and changes to the internal interface are learned and adapted by the agent within the agent, eliminating the need for third-party applications to be aware of or perform adaptation development. This solves the technical problem of high usage costs for network capability opening interfaces and achieves the technical effect of reducing the usage costs of network capability opening interfaces.

[0107] As can be seen, the above mainly describes the solutions provided by the embodiments of this application from a methodological perspective. To achieve the above functions, the embodiments of this application provide corresponding hardware structures and / or software modules for executing each function. Those skilled in the art should readily recognize that, in conjunction with the modules and algorithm steps of the various examples described in the embodiments disclosed herein, the embodiments of this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed by hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this invention.

[0108] In some embodiments, this application also provides a network capability opening device. This network capability opening device may include one or more functional modules for implementing the network capability opening method of the above method embodiments.

[0109] Figure 8 This application provides a schematic diagram of the composition of a network capability opening device, as shown below. Figure 8 As shown, the network capability opening device 800 includes: The receiving module 801 is used to receive natural language request messages sent by third-party applications; the request messages are used to request the invocation of at least one application programming interface (API) corresponding to the Network Capability Open Function (NEF).

[0110] Analysis module 802 is used to call the large model to analyze the request message and output the calling scheme; the calling scheme includes at least the calling order of at least one API to be called; the large model accesses all APIs supported by NEF.

[0111] Module 803 is used to call at least one API corresponding to NEF in the order of invocation and send response messages to third-party applications.

[0112] In some embodiments, the analysis module 802 includes: an acquisition unit, configured to acquire target interface knowledge matching the request message from the NEF interface knowledge base; the NEF interface knowledge base stores at least NEF interface documents collected according to the NEF interface version; a synthesis unit, configured to call an embedded model to synthesize the request message and the target interface knowledge to obtain prompt words; and an output unit, configured to call a large model to output a calling scheme based on the prompt words.

[0113] In other embodiments, the invocation scheme further includes an invocation command for each API. The invocation module 803 includes: a conversion unit, used to convert the format type of the invocation command from the standardized communication interface format supported by the large model to the API format corresponding to NEF when the format type of the invocation command is the standardized communication interface format supported by the large model; and a first invocation unit, used to invoke at least one API corresponding to NEF in the API format corresponding to NEF according to the invocation order.

[0114] In some other embodiments, the invocation scheme further includes an invocation command for each API, and the invocation module 803 includes: a second invocation unit, used to invoke at least one API corresponding to NEF in sequence according to the invocation order when the format type of the invocation command is the API format corresponding to NEF.

[0115] In some other embodiments, the device further includes: an acquisition module, configured to acquire an event message corresponding to an event in response to an event triggered by a third-party application subscription; a processing module, configured to call a large model to process the event message and generate a corresponding reply message; and a sending module, configured to send the reply message to the third-party application.

[0116] In the case of implementing the functions of the integrated modules described above in hardware, this embodiment of the invention provides a possible structural schematic diagram of the electronic device involved in the above embodiments. Figure 9 A schematic diagram of the composition of an electronic device provided in this application, such as... Figure 9 As shown, the electronic device 900 includes: a processor 902, a communication interface 903, and a bus 904. Optionally, the electronic device 900 may also include a memory 901.

[0117] Processor 902 may implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 902 may be a central processing unit, a general-purpose processor, a digital signal processor, an application-specific integrated circuit, a field-programmable gate array, or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 902 may also be a combination that implements computing functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0118] The communication interface 903 is used to connect to other devices via a communication network. This communication network can be Ethernet, wireless access network, wireless local area network (WLAN), etc.

[0119] The memory 901 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto.

[0120] In one possible implementation, the memory 901 can exist independently of the processor 902. The memory 901 can be connected to the processor 902 via a bus 904 and is used to store instructions or program code. When the processor 902 calls and executes the instructions or program code stored in the memory 901, it can implement the network capability opening method provided in this embodiment of the invention.

[0121] In another possible implementation, the memory 901 can also be integrated with the processor 902.

[0122] The 904 bus can be an extended industry standard architecture (EISA) bus, etc. The 904 bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 9 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0123] Through the above description of the implementation methods, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the service calling device can be divided into different functional modules to complete all or part of the functions described above.

[0124] This application also provides a computer-readable storage medium. All or part of the processes in the above method embodiments can be executed by computer instructions instructing related hardware. The program can be stored in the aforementioned computer-readable storage medium, and when executed, it can include the processes of the above method embodiments. The computer-readable storage medium can be any of the foregoing embodiments or memory. The aforementioned computer-readable storage medium can also be an external storage device of the aforementioned service invocation device, such as a plug-in hard drive, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the aforementioned service invocation device. Further, the aforementioned computer-readable storage medium can include both internal storage units of the aforementioned service invocation device and external storage devices. The aforementioned computer-readable storage medium is used to store the aforementioned computer program and other programs and data required by the aforementioned service invocation device. The aforementioned computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0125] This application also provides a computer program product comprising a computer program that, when run on a computer, causes the computer to execute any of the network capability opening methods provided in the above embodiments.

[0126] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A network capability exposure method, characterized by, The method includes: Receive a request message in natural language sent by a third-party application; the request message is used to request the invocation of at least one application programming interface (API) corresponding to the Network Capability Opening Function (NEF); The large model is invoked to analyze the request message and output an invocation plan; the invocation plan includes at least the invocation order of at least one API to be invoked; the large model is connected to all APIs supported by the NEF; In accordance with the calling order, at least one API corresponding to the NEF is called sequentially, and a response message is sent to the third-party application.

2. The method of claim 1, wherein, The large-scale invocation model analyzes the request message and outputs an invocation scheme, including: Obtain the target interface knowledge that matches the request message from the NEF interface knowledge base; the NEF interface knowledge base stores at least NEF interface documents collected according to the NEF interface version; The embedded model is invoked to synthesize the request message and the target interface knowledge to obtain prompt words; The large model is invoked to output the invocation scheme based on the prompt words.

3. The method of claim 1, wherein, The invocation scheme also includes invocation commands for each API, wherein in the invocation order, at least one API corresponding to the NEF is invoked sequentially, including: If the format type of the calling command is a standardized communication interface format supported by the large model, the format type of the calling command will be converted from the standardized communication interface format to the API format corresponding to the NEF. Under the API format corresponding to the NEF, at least one API corresponding to the NEF is called sequentially according to the calling order.

4. The method of claim 1, wherein, The invocation scheme also includes invocation commands for each API, wherein in the invocation order, at least one API corresponding to the NEF is invoked sequentially, including: When the format type of the calling command is the API format corresponding to the NEF, at least one API corresponding to the NEF is called sequentially according to the calling order.

5. The method according to any one of claims 1 to 4, characterized in that, The method further includes: In response to the triggering of an event subscribed by the third-party application, obtain the event message corresponding to the event; The large model is invoked to process the event message and generate a corresponding response message; The reply message is sent to the third-party application.

6. A network capability open system, characterized in that, The system includes: The third-party message interface module is used to receive natural language request messages sent by third-party applications; the request messages are used to request the invocation of at least one application programming interface (API) corresponding to the Network Capability Opening Function (NEF). A large model is used to analyze the request message and output a calling scheme; the calling scheme includes at least the calling order of at least one API to be called; the large model is connected to all APIs supported by the NEF; The NEF interface module calls at least one API corresponding to the NEF in the order described above. The third-party message interface module is also used to send response messages to the third-party application.

7. The system according to claim 6, characterized in that, The system also includes: The NEF Interface Knowledge Module is used to store NEF interface documents collected according to the NEF interface version. The prompt word processing module is used to call the embedded model to synthesize the request message and the target interface knowledge obtained from the NEF interface knowledge module that matches the request message, so as to obtain prompt words that are adapted to the input of the large model.

8. The system according to claim 6, characterized in that, The invocation scheme also includes invocation commands for each API. The NEF interface module has a protocol conversion function and is also used to convert the format type of the invocation commands from the standardized communication interface format supported by the large model to the API format corresponding to the NEF.

9. An electronic device, characterized in that, It includes a processor and a memory, the processor being coupled to the memory; the memory is used to store computer instructions, which are loaded and executed by the processor to enable the computer device to perform the method as described in any one of claims 1 to 5.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes computer-executable instructions that, when executed on a computer, cause the computer to perform the method of any one of claims 1 to 5.

Images