Apparatus for service-based interface (SBI) communication with at least one control function of a telecommunications network and method of operating the apparatus
By introducing the GPT model and NRF/NWDAF interface into the 6G network, intent-based service interface communication was realized, which solved the problem of high complexity of network function interaction, improved the system's flexibility and scalability, and supported multi-NF service calls and chained orchestration.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HUAWEI TECH CO LTD
- Filing Date
- 2023-10-31
- Publication Date
- 2026-05-29
AI Technical Summary
In 6G networks, the interaction complexity between network functions is high, making it difficult to flexibly respond to changes in interfaces and service points, resulting in difficulties in troubleshooting. Furthermore, existing systems cannot understand the intent of the service layer and cannot achieve generalized interaction, leading to poor system scalability.
Employing a Generative Pretrained Transformation (GPT) model interface, combined with Network Repository Function (NRF) and Network Data Analysis Function (NWDAF), it dynamically updates and modifies translation standards, rules, and strategies through a self-learning process, enabling intent-based service interface (SBI) communication, reducing the number of interactions, and providing soft error handling and flexibility.
It enables efficient and flexible interaction between different network functions, reduces the number of interactions, supports multi-NF service calls and chained orchestration, and improves the system's scalability and fault handling capabilities.
Smart Images
Figure CN122122869A_ABST
Abstract
Description
Technical Field
[0001] This disclosure generally relates to the fields of telecommunications and machine learning (ML), and more particularly to an apparatus for SBI communication with at least one control function of a telecommunications network, and a method for operating the apparatus. Background Technology
[0002] 6G systems are currently under development and are expected to widely utilize machine learning (ML) technology to achieve efficient control and interaction. A significant trend in mobile system development is the dramatic increase in the number of network functions (NFs) providing new services within the mobile network core. Furthermore, interactions between different NFs are becoming multidimensional, requiring multiple interaction steps to achieve a specific function. The application of microservices further highlights this trend, breaking down large monolithic services into much smaller, more specific services. These services can then be chained or sequentially invoked to implement larger services. Therefore, the core challenge currently is ensuring efficient interaction while minimizing the overhead of microservices.
[0003] Another trend is the urgent need in mobile networks to implement intent-based service invocation, where the requesting or consumer NF sends its service intent rather than invoking a specific API or service. This increases the flexibility of the interaction but can also increase its complexity, especially given the increasing number of NF types and the services they provide. Alternatives such as complex manual pre-configuration (the traditional way mobile systems manage their services) have limitations, failing to provide flexibility and configurability.
[0004] Furthermore, the application of machine learning (ML) in mobile systems is becoming increasingly widespread, and the Network Data Analytics Function (NWDAF) has been defined to handle data analysis and ML within mobile systems. This development benefits from the strong interest of various standardization organizations in applying machine learning, which promises to simplify network design and improve performance. However, to date, NWDAF has largely been used as a black box, assuming that sufficient network data is all that's needed to solve the problem.
[0005] Generally speaking, current networks do not support intent-based control. This leads to increased manual configuration and troubleshooting issues, and the rigid API call methods also increase the likelihood of downtime.
[0006] Furthermore, the current network faces scalability issues. This is because the large number of NF types, deployment locations, slices, and the numerous possible combinations of function calls to each NF result in many possibilities for executing and implementing services. Moreover, creating new NFs and their interfaces is both cumbersome and error-prone. If an interface fails, the entire UE request will fail, even if other methods exist to fulfill the request. The current system cannot generalize from past API call experience to new calls, thus failing to understand service-level intent. Generally, the dynamic implementation of current function calls relies on manual configuration and transformation functions, which limits its ability to implement dynamic behavior and heavily depends on the management and configuration manager.
[0007] Based on the above, 6G networks do not support intelligent SBA interaction, and therefore cannot flexibly and dynamically respond to constantly evolving interfaces and service points, nor can they properly handle errors. Furthermore, interactions between network nodes (NFs) cannot be generalized. Summary of the Invention
[0008] The aim is to overcome the above and other shortcomings.
[0009] The above and other objectives are achieved through the features of the independent claim. Other implementations will be apparent from the dependent claims, the description, and the drawings.
[0010] According to a first aspect, an apparatus is provided for communicating via a Service-Based Interface (SBI) with at least one control function of a telecommunications network. The apparatus includes a Generative Pretrained Transformer (GPT) model interface; a Network Repository Function (NRF) interface; a Network Data Analytics Function (NWDAF) interface; and a processor. The processor is configured to: obtain a GPT model for the SBI communication via the GPT model interface; obtain an indication of available services provided by the control function via the NRF interface; obtain real-time status information of the telecommunications network via the NWDAF interface; receive a first service intent or a first service request; verify the received first service intent or first service request using the GPT model; determine one or more execution services among the available services based on the verified first service intent or first service request and the real-time status information using the GPT model; and send a second service intent or a second service request for at least one of the execution services.
[0011] This has the following benefits / advantages:
[0012] - Intent-based interactions between different NFs and UEs, oriented towards the network core (i.e., the service requester can send its intent and the potential result of the request, rather than the request itself).
[0013] - A continuous self-learning process that allows translation standards, rules, and strategies to be automatically updated and modified dynamically based on current network conditions;
[0014] - Continuously learn interfaces from interactions between specifications and NFs (NWDAF, AF, NRF, etc.);
[0015] -Continuous interaction with new NF types;
[0016] -Reduce the number of interactions required;
[0017] - The GPT SBA acts as a single point of interaction between the consumer and provider NF, responding to the NF without interacting with the NRF;
[0018] -flexibility;
[0019] - Generalization of learned interaction logic (e.g., multiple calls in one go, multi-NF processes (e.g., chained orchestration), conditional request processing based on inputs from other NFs).
[0020] - Integrate new interfaces by querying the list of interfaces;
[0021] - Soft error handling.
[0022] In this article, a network function (NF) can refer to a functional network entity or building block that originates from network infrastructure virtualization and can be connected to or chained with other NFs to create and deliver services. For example, NFs can be distinguished based on their role in service invocation (consumer NF, provider NF, or cloud service NF) or based on their participating plane (control plane (network) function, user plane (network) function / UPF, or cloud-based (network) function).
[0023] In this article, Service Interface or SBI can refer to API-based communication within a service-oriented architecture.
[0024] In this article, Service-Oriented Architecture (SBA) can refer to a functional network architecture in which network nodes (NFs) communicate with each other through SBIs specific to the respective endpoints / NFs, rather than through point-to-point (P2P) interfaces specific to the respective endpoint pairs.
[0025] In this paper, Generative Pretrained Transformer (GPT) can refer to a Transformer-type artificial neural network that uses self-supervised or semi-supervised learning and / or reinforcement learning with a large amount of training data for (pre)training to obtain its patterns and structure and to generate new data with similar characteristics.
[0026] In this article, Network Repository Function or NRF can refer to the control NF of a telecommunications network, which provides registration and service discovery functions by maintaining configuration files of NF instances and the services they support, enabling NFs to discover each other and communicate through SBI.
[0027] In this article, Network Data Analysis Function or NWDAF can refer to the control NF of a telecommunications network, which collects data from the terminal (UE), network entity (NF), and operation, management and maintenance (OAM) system through a subscription or request model, calculates analysis results based on the collected data and a specific data model (e.g., AI / ML model), and shares the analysis results with other functions (data analysis consumers) in the network.
[0028] In this article, real-time status information can refer to the instantaneous utilization rate of a telecommunications network, such as in communication resource allocation and NF workload.
[0029] In this article, a service request can refer to the form of a service invocation, in which the service requester (i.e., the UE or consumer NF) provides detailed requirements and configuration parameters for the required service based on a specific SBI.
[0030] In this article, service intent can refer to the form of service invocation, in which the service requester provides instructions and information about the underlying goals and purposes of the required service.
[0031] In one possible implementation, the processor can also be used to: receive a second service response to the sent second service intent or second service request; verify the received second service response using the GPT model; and, depending on whether the verification is successful, send a first service response to the received first service intent or first service request.
[0032] This allows the service call to complete successfully.
[0033] In this article, verification can refer to proving whether something is correct or accurate.
[0034] In one possible implementation, the second service intent or service request sent may be sent to the first provider network function (NF).
[0035] This enables multi-NF service calls.
[0036] In one possible implementation, the received second service response may be received from the first provider NF.
[0037] This enables sequential multi-NF service calls.
[0038] In one possible implementation, the received second service response may be received from a second provider NF, which is different from the first provider NF and is determined using the GPT model.
[0039] This enables chained orchestration of multi-NF service calls (Service Function Chain / SFC).
[0040] In one possible implementation, the processor can also be used to: use the GPT model to disprove the implementation result of the first service intent or first service request through the verified second service response; and, depending on whether the disprovement is successful, send a second service intent or second service request for another service in the execution service.
[0041] This enables dynamic multi-NF service calls.
[0042] In this article, falsification can refer to proving whether something is false.
[0043] In one possible implementation, the processor can also be used to: determine feedback based on an invalid or incomplete first service intent or first service request using the GPT model; and send the determined feedback.
[0044] In one possible implementation, the feedback may include a correction or query for the invalid or incomplete first service intent or first service request.
[0045] This enables soft error handling.
[0046] In this document, "correction" may refer to a suggested change to the invalid or incomplete first service intent or first service request as determined by the GPT model.
[0047] In this document, a query may refer to a suggested variation of the invalid or incomplete first service intent or first service request as determined by the GPT model.
[0048] In one possible implementation, the received first service intent or first service request may be received from the UE or the consumer NF.
[0049] This enables terminal or network entities to make service calls.
[0050] In one possible implementation, the device may be part of the service-based architecture (SBA) of the telecommunications network.
[0051] This enables direct service calls.
[0052] In one possible implementation, the device can be part of a separate NF in the telecommunications network and can be used to act as a service broker.
[0053] This enables indirect service calls to be made through the service proxy.
[0054] In this article, a service agent can refer to an NF that acts on behalf of a provider NF.
[0055] In one possible implementation, the device may also include the GPT model.
[0056] This allows for complete control over the learning / training phase of the GPT model.
[0057] In one possible implementation, the GPT model can be used to: learn potential responses to the invalid or incomplete first service intent or first service request.
[0058] This enables the implementation of learned soft error handling.
[0059] In one possible implementation, the GPT model can be used to: learn how to determine the execution service based on the verified first service intent or first service request and the real-time state information from one or more of the following: SBI standard documentation; SBI features related to available services and corresponding use cases; SBI API specifications of various provider NFs defined in machine-readable language; and recorded SBI interactions and service calls.
[0060] Log files for each provider's NF; log files for each consumer's NF; NRF information; NRF log files.
[0061] This makes it possible to learn how to break down the process into execution services.
[0062] In one possible implementation, the telecommunications network may include a mobile network.
[0063] This enables the device to be deployed in mobile networks, especially in fifth-generation (5G) networks. th generation, 5G) and sixth generation (6G) th Deployed in 3GPP networks (generation, 6G).
[0064] According to a second aspect, a telecommunications network is provided, including the device described in the first aspect or any implementation thereof. The processor may further be configured to: receive a first service intent or a first service request from outside the management domain of the telecommunications network; and send a second service intent or a second service request for at least one of the services performed outside the management domain of the telecommunications network.
[0065] This enables UEs and cloud providers to make service calls through the SBI, as the SBI bus can extend beyond the core network, that is, across the air interface and cloud systems.
[0066] In this article, for example, a management domain can refer to the scope of management authority that extends within the management boundaries of an organization.
[0067] According to a third aspect, a method is provided for operating an apparatus for communicating via a Service-Based Interface (SBI) with at least one control function of a telecommunications network. The apparatus includes a Generative Pretrained Transformer (GPT) model interface; a Network Repository Function (NRF) interface; and a Network Data Analytics Function (NWDAF) interface. The method includes: obtaining a GPT model for API communication of the SBI through the GPT model interface; obtaining an indication of available services provided by the control function through the NRF interface; obtaining real-time status information of the telecommunications network through the NWDAF interface; receiving a first service intent or a first service request; verifying the received first service intent or first service request using the GPT model; determining one or more execution services among the available services based on the verified first service intent or first service request and the real-time status information using the GPT model; and sending a second service intent or a second service request for at least one of the execution services.
[0068] In one possible implementation, the method can be performed by a device according to the first aspect or any implementation thereof.
[0069] Fourthly, a computer program is provided, including program code, wherein when the program code is executed on a computer, the program code is used to perform the method according to the third aspect or any implementation thereof.
[0070] According to a fifth aspect, a user equipment (UE) is provided, including a processor, configured to: send a first service intent or a first service request to a device for communicating with at least one control function of a telecommunications network via a Service-Based Interface (SBI); and receive a first service response to the received first service intent or first service request.
[0071] In one possible implementation, the processor can also be used to: receive feedback on an invalid or incomplete first service intent or first service request.
[0072] In one possible implementation, the feedback may include a correction or query for the invalid or incomplete first service intent or first service request.
[0073] The technical effects and advantages of the equipment described above in relation to the first aspect also apply to the second to fifth aspects having the corresponding features. Attached Figure Description
[0074] The above aspects and implementations are explained below with reference to the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements.
[0075] The accompanying drawings should be considered as schematic illustrations, and the elements shown are not necessarily shown to scale. Rather, the various elements are shown to make their function and general purpose obvious to those skilled in the art.
[0076] Figure 1 An exemplary device according to this disclosure is shown;
[0077] Figure 2 It shows Figure 1 An exemplary 5G SBA integration implementation for a device;
[0078] Figure 3 It shows Figure 1 An exemplary 5G standalone NF implementation for a device;
[0079] Figure 4 It shows Figure 1 Exemplary 6G implementations for devices;
[0080] Figure 5 An exemplary method according to this disclosure is shown;
[0081] Figure 6 It shows the basis Figure 5 Exemplary sequence or dynamic invocation of methods;
[0082] Figure 7 It shows as based on Figure 5 A specific example of the method for creating an exemplary PDU session;
[0083] Figure 8 It shows the basis Figure 5 An example of a method is service function chaining (SFC). Detailed Implementation
[0084] In the following description, reference is made to the accompanying drawings, which form part of this disclosure, which illustratively illustrate specific aspects of implementations of this disclosure or aspects in which implementations of this disclosure may be used. It should be understood that implementations of this disclosure may be used in other ways, including structural or logical variations not depicted in the drawings. Therefore, the following detailed description should not be construed in a limiting sense, and the scope of this disclosure is defined by the appended claims.
[0085] For example, it should be understood that the disclosure relating to the described method is equally applicable to the corresponding apparatus or system for performing the method, and vice versa. For instance, if one or more specific method steps are described, the corresponding apparatus may include one or more units (e.g., functional units) to perform the described one or more method steps (e.g., one unit performs one or more steps, or multiple units perform one or more of a plurality of steps respectively), even if such one or more units are not explicitly described or illustrated in the drawings. On the other hand, for example, if a specific apparatus is described based on one or more units (e.g., functional units), the corresponding method may include a step to perform the function of one or more units (e.g., one step performs the function of one or more units, or multiple steps perform the function of one or more of a plurality of units respectively), even if such one or more steps are not explicitly described or illustrated in the drawings. Furthermore, it should be understood that, unless otherwise explicitly stated, features of the various exemplary implementations and / or aspects described herein can be combined with each other.
[0086] Figure 1 Device 1 according to this disclosure is shown.
[0087] Device 1 is adapted to communicate via SBI with at least one control function 24 of telecommunications network 2.
[0088] For example, telecommunications network 2 may include a mobile (i.e., 3GPP) network.
[0089] Device 1 includes a processor 16, which may include a virtualized portion of an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a network processing unit (NPU), a digital signal processor (DSP), a microprocessor (μP), etc., or generally includes any processing logic for performing the method 3 described in the third aspect or any implementation thereof (see Figure 5 ).
[0090] The processor 16 is used to obtain the GPT model 4 (Ngpt_obtain_model) of the 301SBI communication through the GPT model interface 11 of the device 1.
[0091] Processor 16 is also used to obtain, via the NRF interface 12 of device 1, indication 121 (Nnrf_get_services) of available services provided by control function 24.
[0092] The processor 16 is also used to obtain real-time status information 131 (Nnwdaf_subscribe) of the 303 telecommunications network 2 through the NWDAF interface 13 of the device 1.
[0093] exist Figure 1 It should be noted that the NRF interface 12 and the NWDAF interface 13 are respectively converged into the SBI bus 23, which provides a connection to the NRF and NWDAF control NF 24.
[0094] Processor 16 is also used to receive 304 first service intent 141a or first service request 141b (Nue_gpt_sba). This can be achieved through the service consumer interface 14 of device 1.
[0095] The first service intent 141a or the first service request 141b received may be received from the user equipment (UE) 21 or the consumer NF 22.
[0096] The UE 21 includes a processor 211, which is used to send a first service intent 141a or a first service request 141b to the device 1.
[0097] Processor 16 is also used to verify the first service intent 141a or the first service request 141b received by 305 using GPT model 4.
[0098] In the case of the first service intent 141a, a service request 141b matching it was detected based on the GPT model.
[0099] The service request body will be validated based on the GPT model. This validation primarily checks whether the input meets expectations and whether it is consistent with the training data.
[0100] Processor 16 can also be used to determine 306 feedback 142 based on an invalid or incomplete first service intent 141a or first service request 141b using GPT model 4; and send 307 the determined feedback 142.
[0101] If the service request is invalid, a response 142 will be returned instead of a simple error message, as the latter would not be helpful to the service requester.
[0102] At this point, prompt word engineering can be applied to generate the best possible answer for the initial request without requiring multiple back-and-forth interactions. Prompt word engineering refers to an engineering and design-based approach used to normalize the intent or request input into the GPT module.
[0103] On the UE 21 side, the processor 211 can also be used to receive feedback 142 for invalid or incomplete first service intent 141a or first service request 141b.
[0104] Feedback 142 may include corrections or queries for invalid or incomplete first service intent 141a or first service request 141b. For example, feedback 142 may include suggestions for corrected service intent 141a, or queries for suggested choices of corrected service intent 141a, and all feedback always conforms to GPT Model 4.
[0105] Processor 16 is also used to determine one or more execution services 41 among the 308 available services using GPT model 4 based on the verified first service intent 141a or first service request 141b and real-time status information 131.
[0106] In other words, service requests are broken down into one or more underlying NF requests. For example, GPT Model 4 transforms a service request (e.g., establishing a PDU session for a UE with ID 1111 and enabling it to access a low-latency slice) into multiple requests sent to the SMF, NSSMF, and AMF, which trigger the underlying requests required based on training data.
[0107] The processor 16 is also configured to send 309 a second service intent 151a or a second service request 151b for at least one of the services 41 (Nnf_service_call). This can be achieved through the service provider interface 15 of the device 1.
[0108] Device 1 is used to process multiple (potentially in parallel) requests and ensures that each provider NF 25 returns a valid response.
[0109] Processor 16 can also be used to receive a second service response 152 to the second service intent 151a or the second service request 151b sent by 310; verify the second service response 152 received by 311 using GPT model 4; and send a first service response 143 to the first service intent 141a or the first service request 141b received by 313, depending on whether the verification is successful.
[0110] On the UE 21 side, the processor 211 is also configured to receive a first service response 143 to the received first service intent 141a or first service request 141b.
[0111] According to one implementation, device 1 may also include GPT model 4.
[0112] GPT Model 4 can be used to learn potential responses 142 to invalid or incomplete first service intent 141a or first service request 141b.
[0113] GPT Model 4 can be used to learn how to determine the execution service 41 based on one or more of the following, based on a verified first service intent 141a or first service request 141b and real-time status information 131: SBI standard documentation; SBI characteristics related to available services and corresponding use cases; SBI API specifications for each provider NF defined in machine-readable language; recorded SBI interactions and service calls; log files for each provider NF; log files for each consumer NF; NRF information; and NRF log files.
[0114] Figure 2 It shows Figure 1 The 5G SBA integration implementation method of device 1.
[0115] The SBA of Telecommunication Network 2 includes a typical control NF 24 connected via SBI bus 23 (solid line). Telecommunication Network 2 also includes a typical user plane NF connected via a P2P interface (dashed line).
[0116] Device 1 can be a component of the SBA, such as Figure 2 As shown, SBI bus 23 "integrates" device 1. In other words, device 1 is directly integrated into SBI bus 23, and the service requester is unaware of its existence because it is completely transparent. Device 1 itself is aware of SBI communication on telecommunications network 2.
[0117] Figure 3 It shows Figure 1The implementation method of 5G standalone NF for device 1.
[0118] Similarly, the SBA of telecommunications network 2 includes a typical control NF 24 connected via SBI bus 23 (solid line). Telecommunications network 2 also includes a typical user plane NF connected via a P2P interface (dashed line).
[0119] Device 1 can be part of a separate NF in telecommunications network 2, such as Figure 3 As shown, another control NF 1 on SBI bus 23 acts as a service broker. In other words, the service requester needs to discover the other control NF 1 via NRF 24 and then send a request to the other control NF 1 that acts as its service broker.
[0120] Figure 4 It shows Figure 1 The 6G implementation method of device 1;
[0121] In the sixth generation (6 th In the 6G architecture, the SBI bus 23 can be extended beyond the core network, that is, across the air interface and to the cloud provider. This enables the UE 21, user plane nodes, and cloud providers to make service calls through the SBI.
[0122] In this scenario, device 1 can function as part of a separate NF within telecommunications network 2, such as... Figure 4 As shown, another control NF 1 on SBI bus 23 is used as a service agent.
[0123] Therefore, its processor 16 can be used to receive 304' first service intent 141a or first service request 141b from outside the management domain 26 of the telecommunications network 2; and send 309' second service intent 151a' or second service request 151b' for at least one of the services 41 to outside the management domain 26 of the telecommunications network 2.
[0124] Figure 5 Method 3 according to this disclosure is shown.
[0125] Method 30 defines the operation of device 1 according to the first aspect or any implementation thereof. Conversely, method 30 can be executed by device 1 according to the first aspect or any implementation thereof.
[0126] Method 30 includes the following steps: obtaining the GPT model 4 of the API communication of 301 SBI through the GPT model interface 11 of device 1.
[0127] Method 30 further includes the step of obtaining, via the NRF interface 12 of device 1, an indication 121 of available services provided by control function 24.
[0128] Method 30 further includes the following step: obtaining real-time status information 131 of telecommunications network 2 303 through the NWDAF interface 13 of device 1.
[0129] Method 30 further includes the following steps: receiving 304 first service intent 141a or first service request 141b.
[0130] Method 30 further includes the following steps: verifying the first service intent 141a or the first service request 141b received by 305 using GPT model 4.
[0131] Method 30 may further include the following steps: using GPT model 4 to determine 306 feedback 142 based on an invalid or incomplete first service intent 141a or first service request 141b; and sending 307 the determined feedback 142.
[0132] Method 30 further includes the following steps: based on the verified first service intent 141a or first service request 141b and real-time status information 131, using GPT model 4 to determine one or more execution services 41 among the available services 308.
[0133] The method 30 further includes the step of: sending 309 a second service intent 151a or a second service request 151b for at least one of the services 41.
[0134] Method 30 may further include the following steps: receiving a second service response 152 to the sent second service intent 151a or second service request 151b; verifying the second service response 152 received by 311 using GPT model 4; and, depending on whether the verification is successful, sending a first service response 143 to the received first service intent 141a or first service request 141b.
[0135] Method 30 may also include the following intermediate steps: using GPT model 4, falsifying the implementation result of the first service intent 141a or the first service request 141b by the verified second service response 152; and sending 309''' or 'Second service intent 151a'' or 'Second service request 151b'' for another service in the execution service 41, depending on whether the falsification is successful.
[0136] Figure 6 It shows the basis Figure 5 Method 3 can be called sequentially or dynamically.
[0137] In this example, the first service intent 141a or service request 141b may be expressed as or include a UE's request for registration, session, etc.
[0138] For example, the second service intent 151a or service request 151b sent may be expressed as or include an SBI call to the first provider NF 25 (NF1), and the second service response 152 received may be received from the first provider NF 25 (NF1).
[0139] The above message exchange will be repeated for intermediate provider NF 25 (NF2) and the second (i.e. the last) provider NF 25 (NF3).
[0140] In other words, provider NF 25 (NF1, NF2, NF3) can sequentially receive the corresponding second service request 151b under the intermittent control of device 1.
[0141] The term “sequential” as used in this article can refer to one after another (see loop 309-310-311-312-309-…), while “dynamic” can refer to whether the previous request 151b was successful (see the branch at 312).
[0142] For this purpose, processor 16 can also be used to use GPT model 4 to falsify the implementation result of first service intent 141a or first service request 141b by verified second service response 152; and send 309'' second service intent 151a'' or second service request 151b'' for another service in execution service 41, depending on whether the falsification is successful.
[0143] The final generated first service response 143 may include the UE's response to registration, session, etc.
[0144] Figure 7 It shows as based on Figure 5 Method 3 provides a specific example of PDU session creation.
[0145] In this example, the first service intent 141a or service request 141b may be expressed as or include a session establishment request.
[0146] The second service intent 151a or service request 151b may be expressed as or include a session establishment request directed to the Access and Mobility Management Function (AMF) 25 of the telecommunications network 2.
[0147] Then, AMF 25 can send a request to Network Repository Function (NRF) 25 of Telecom Network 2 to perform service discovery. NRF 25 can return a list of Session Management Functions (SMF) 25 as a response.
[0148] Next, AMF 25 can send a request to one of the SMF 25s in Telecommunication Network 2 to perform session establishment. SMF 25 can return session management context data as a response.
[0149] Subsequently, AMF 25 can return a session establishment response as the first service response 152.
[0150] The final generated first service response 143 may also include a session establishment response.
[0151] Figure 8 It shows the basis Figure 5 Method 3 is service function chaining (SFC).
[0152] In this example, the first service intent 141a or service request 141b may include the UE's request for registration, session, etc.
[0153] The second service intent 151a or service request 151b can be sent to the first provider NF 25 (NF1), and the received second service response 152 can be received from the second provider NF 25 (NF3), which is different from the first provider NF (NF1) and has been determined using GPT model 4 308.
[0154] In other words, provider NF 25 (NF1, NF2, NF3) can receive the corresponding second service request 151b in sequence according to the SFC determined by GPT model 4 308, i.e., without the need for intermittent control of device 1.
[0155] The final generated first service response 143 may include the UE's response to registration, session, etc.
[0156] This disclosure has described various implementations as examples. However, based on a study of the drawings, this disclosure, and the independent claims, those skilled in the art will be able to understand and implement other variations when implementing the claimed subject matter. In the claims and the description, the word "comprising" does not exclude other elements or steps, and the quantifier "a" does not exclude a plurality. A single element or other unit may fulfill the function of several entities or items described in the claims. The enumeration of certain measures in dissimilar dependent claims does not imply that combinations of these measures cannot be used in advantageous implementations. A computer program may be stored / distributed on a suitable medium, such as an optical storage medium or a solid-state medium provided together with or as part of other hardware, and may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunications systems.
Claims
1. A device (1), characterized in that, For communication via a Service-Oriented Interface (SBI) with at least one control function (24) of the telecommunications network (2), including: Generative pre-trained transform GPT model interface (11); Network repository functionality NRF interface (12); Network data analysis function NWDAF interface (13); Processor (16), used for: - Obtain the GPT model (4) of the SBI communication through the GPT model interface (11) (301); - Obtain (302) an indication (121) of the available services provided by the control function (24) through the NRF interface (12); - Obtain (303) the real-time status information (131) of the telecommunications network (2) through the NWDAF interface (13). - Receive (304) first service intent (141a) or first service request (141b). - Use the GPT model (4) to verify (305) the first service intent (141a) or the first service request (141b) received. -Based on the verified first service intent (141a) or first service request (141b) and the real-time status information (131), use the GPT model (4) to determine (308) one or more execution services (41) among the available services. - Send (309) a second service intent (151a) or a second service request (151b) for at least one of the execution services (41).
2. The device (1) according to claim 1, characterized in that, The processor (16) is also used for: - Receive (310) a second service response (152) to the second service intent (151a) or the second service request (151b) sent. - Use the GPT model (4) to verify (311) the received second service response (152). - Depending on whether the verification is successful, send (313) a first service response (143) to the received first service intent (141a) or first service request (141b).
3. The device (1) according to claim 2, characterized in that, The second service intent (151a) or service request (151b) is sent to the first provider network function NF (25).
4. The device (1) according to claim 2 or 3, characterized in that, The second service response (152) received is received from the first provider NF (25).
5. The device (1) according to claim 2 or 3, characterized in that, The second service response (152) received is received from the second provider NF (25), which is different from the first provider NF and is determined using the GPT model (4).
6. The device (1) according to any one of claims 2 to 5, characterized in that, The processor (16) is also used for: - Using the GPT model (4), the implementation result of the first service intent (141a) or the first service request (141b) is falsified (312) by the verified second service response (152); - Depending on whether the disproving is successful, send (309'') a second service intent (151a'') or a second service request (151b'') for another service in the execution service (41).
7. The device (1) according to any one of claims 1 to 6, characterized in that, The processor (16) is also used for: - Use the GPT model (4) to determine (306) feedback (142) based on invalid or incomplete first service intent (141a) or first service request (141b). - Send (307) the feedback (142) that is determined.
8. The device (1) according to claim 7, characterized in that, The feedback (142) includes corrections or queries for the invalid or incomplete first service intent (141a) or first service request (141b).
9. A method (30), characterized in that, Device (1) for operating at least one control function (24) of the telecommunications network (2) to communicate via a service interface (SBI). The method (30) includes: - Obtain (301) the GPT model (4) of the SBI API communication through the generative pre-trained transform GPT model interface (11) of the device (1). - Obtain (302) an indication (121) of the available services provided by the control function (24) through the Network Repository Function (NRF) interface (12) of the device (1); - Obtain (303) the real-time status information (131) of the telecommunications network (2) through the network data analysis function NWDAF interface (13) of the device (1). - Receive (304) first service intent (141a) or first service request (141b). - Use the GPT model (4) to verify (305) the first service intent (141a) or the first service request (141b) received. -Based on the verified first service intent (141a) or first service request (141b) and the real-time status information (131), use the GPT model (4) to determine (308) one or more execution services (41) among the available services. - Send (309) a second service intent (151a) or a second service request (151b) for at least one of the execution services (41).
10. The method (30) according to claim 9, characterized in that, The method is performed by the device (1) according to any one of claims 2 to 8.
11. A computer program, characterized in that, Includes program code, wherein when the program code is executed on a computer, the program code is used to perform the method (30) according to claim 9 or 10.
12. A user equipment (UE) (21), characterized in that, include: Processor (211), for: - Send a first service intent (141a) or a first service request (141b) to a device (1) for communicating with at least one control function (24) of the telecommunications network (2) via a service interface (SBI). - Receive a first service response (143) to the received first service intent (141a) or first service request (141b).
13. The UE (21) according to claim 12, characterized in that, The processor (211) is also used for: - Receive feedback (142) for invalid or incomplete first service intent (141a) or first service request (141b).
14. The UE (21) according to claim 13, characterized in that, The feedback (142) includes corrections or queries for the invalid or incomplete first service intent (141a) or first service request (141b).