Communication method and apparatus

Abstract

The present application can be applied to the field of communications. Provided are a communication method and apparatus. Exemplarily, a session management network element can determine, on the basis of model information, a quality-of-service channel corresponding to a service. In this way, a binding manner of the quality-of-service channel is suitable for a mechanism in which a model assists an access network device in performing quality-of-service control. If the quality-of-service channel corresponding to the service is determined in the foregoing manner, then, when inference is executed on the basis of the model, data flows of different services are not mixed together during inference, which takes model information corresponding to different services into full consideration, and is beneficial to reducing the impact between the data flows of different services, thereby improving the accuracy of model inference.

Description

Communication methods and devices

[0001] This application claims priority to Chinese Patent Application No. 202510372407.5, filed on March 26, 2025, entitled "Communication Method and Apparatus", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of communications, and particularly to communication methods and apparatus in the field of communications. Background Technology

[0003] In communication networks, the management and control of Quality of Service (QoS) is one of the key factors in ensuring network performance and user experience. Currently, after policy control network elements (such as policy control function (PCF) network elements) send service-related policy information (such as policy and charging (PCC) rules) to session management network elements (such as session management function (SMF) network elements), the session management network element executes the binding of policy information and QoS channels (such as QoS flows), thereby ensuring that the aforementioned services can obtain the corresponding QoS control. The aforementioned policy information includes one or more QoS parameters used to execute QoS channel binding, such as the fifth-generation (5G) QoS identifier (5QI) and allocation and retention priority (ARP). The session management network element can execute the binding of policy information and QoS channels based on these QoS parameters.

[0004] With the development of artificial intelligence (AI), it has become possible to infer the arrival patterns of future data packets based on models (such as a data packet of size x megabytes (MB) arriving in 1 second), thus facilitating base stations to perform QoS control based on the inference results of the model. The existing QoS channel binding methods described above are not compatible with the model-assisted QoS control mechanism and require optimization. Summary of the Invention

[0005] This application provides a communication method and apparatus that are adapted to the mechanism of model-assisted QoS control and improve the accuracy of inference results.

[0006] Firstly, this application provides a communication method that can be applied to a session management network element, such as a session management network element itself, or a component configured in a session management network element (such as a processor, chip, chip system, etc.), or a logic module or software capable of implementing all or part of the functions of the session management network element. This application does not limit the application in this regard. The following example uses the execution of a session management network element.

[0007] For example, the method includes: obtaining first model information, the first model information being used for forwarding processing of a first service; determining a first quality of service channel corresponding to the first service based on the first model information, and providing the same forwarding processing for traffic belonging to the first quality of service channel.

[0008] For example, the first service can refer to the data / data packet / service flow / data stream of the first service, or the data / data packet / service flow / data stream corresponding to the first service, or the data / data packet / service flow / data stream used for the first service, or the data / data packet / service flow / data stream sent or received by the application server corresponding to the first service. For example, the first service can be identified using a service data flow / service data flow filter (e.g., Internet Protocol (IP) 5-tuple, IP triplet), or by an application identifier, or by a domain name identifier (Full Qualified Domain Name (FQDN) / Uniform Resource Locator (URL)). This application does not limit the identification method of the first service.

[0009] The aforementioned first model information, used for forwarding processing of the first service, can be replaced with: the aforementioned first model information is used for traffic forwarding of the first service; or, the aforementioned first model information is used for service quality management of the first service; or, the aforementioned first model information is used for service quality control of the first service; or, the aforementioned first model information is used for customer experience (QoE) of the first service. Customer experience can also be referred to as user experience. The word "used for" can be replaced with "used to influence," "used to control," "used to assist," or "used to execute," etc.

[0010] Optionally, the aforementioned first model information is used to indicate the forwarding process of the first service, specifically including: the aforementioned first model information is used for any one or more of the following of the first service: burst prediction, encrypted flow detection, base station resource prediction, or congestion prediction.

[0011] Similarly, traffic belonging to the first quality of service channel is provided with the same forwarding processing, which can be replaced by traffic belonging to the first quality of service channel being provided with the same traffic forwarding or traffic forwarding processing, or traffic belonging to the first quality of service channel being provided with the same quality of service control, or traffic belonging to the first quality of service channel being provided with the same quality of service management, or traffic belonging to the first quality of service channel being provided with the same customer experience.

[0012] Optionally, the aforementioned quality of service channel may be, for example, a QoS stream, a bearer, a radio bearer, or a connection.

[0013] For example, the above solution can be applied to scenarios involving model-assisted quality of service (QoS) control / management. In this solution, the session management network element considers the model information corresponding to the service when determining the QoS channel. That is, whether different services correspond to / are bound to the same QoS channel is related to the model information corresponding to the service. This makes the QoS channel binding method compatible with the QoS control mechanism of model-assisted access network devices. Under this binding method, when performing inference based on the model, whether the data streams of different services are mixed together for inference fully considers the model information corresponding to different services, which helps reduce the impact between data streams of different services, thereby improving the accuracy of model inference.

[0014] Optionally, model-assisted service quality control / management can be replaced with AI-assisted service quality control / management, model-enhanced service quality control / management, or AI-enhanced service quality control / management, etc.

[0015] In conjunction with the first aspect, in some possible implementations of the first aspect, obtaining the first model information includes: obtaining first strategy information, which includes the first model information and the information of the first business.

[0016] The aforementioned first policy information can indicate the policy and billing control method corresponding to the first service. This first policy information may be, for example, the PCC rule corresponding to the first service, or the PCC rule of the first service, or the aforementioned first policy information may include the PCC rule of the first service. The information of the first service can also be replaced by the identifier of the first service. The information / identifier of the first service is used to identify the first service, and this information / identifier may include, but is not limited to, one or more of the following: service identifier, application identifier corresponding to the service, Internet Protocol (IP) triplet or IP quintuple. The IP triplet may, for example, be source address, source port number, and protocol type; the IP quintuple may, for example, be source address, destination address, source port number, destination port number, and protocol type.

[0017] In this implementation, determining the first quality of service channel corresponding to the first service based on the first model information includes: determining the first quality of service channel corresponding to the first policy information based on the first model information.

[0018] In other words, the session management network element can execute the mapping / binding of policy information and QoS channels based on the model information in the policy information. The aforementioned policy information corresponds to services; one policy information can correspond to one or more service information, but the QoS parameters and model information within the same policy information are identical. In this way, services and QoS channels are bound / mapped based on the model information. The above implementation involves minimal improvements to existing protocols (such as signaling improvements). In this application, it is only necessary to add model information to the policy information so that the session management network element can execute the binding of policy information and QoS channels based on the model information in the policy information.

[0019] Optionally, the above-mentioned session management network element obtains the first model information by: the session management network element receiving the first model information corresponding to the first service from the policy control network element.

[0020] Optionally, the above-mentioned session management network element obtains the first model information, including: the session management network element obtains the first model information corresponding to the first service from local storage.

[0021] In conjunction with the first aspect, in some possible implementations of the first aspect, the aforementioned first quality of service channel corresponds only to the first service; or, the aforementioned first quality of service channel does not correspond to any other service besides the first service. That is to say, the first service is mapped to / bound to a specific / separate / dedicated quality of service channel.

[0022] In other words, as long as a service corresponds to model information, or as long as the service quality control / management of the service is assisted by the model, then the service is mapped to a specific / separate / dedicated service quality channel. Alternatively, the session management network element establishes a specific / separate / dedicated service quality channel for the service. This helps ensure that the model's input parameters are separate service data streams, thereby improving model performance, increasing the accuracy of model inference, and reducing the impact on the access network equipment's mechanism for assisting service quality control based on inference results.

[0023] In conjunction with the first aspect, in some possible implementations of the first aspect, the method further includes: obtaining second model information, which is used for forwarding processing of the second service; and determining the quality of service (QoS) channel corresponding to the second service based on the second model information. Specifically, determining the QoS channel corresponding to the second service based on the second model information includes: when the first model information and the second model information are the same, determining that the second service corresponds to a first QoS channel; or, in other words, the first QoS channel corresponds to one or more services using the first model information, including the first service; or, in other words, the first QoS channel corresponds to services using the first model information (such as the first service and the second service).

[0024] For example, the second service can refer to the data / data packet / service flow / data stream of the second service, or the data / data packet / service flow / data stream corresponding to the second service, or the data / data packet / service flow / data stream used for the second service, or the data / data packet / service flow / data stream sent or received by the application server corresponding to the second service. For example, the second service can be identified using a service data stream / service data stream filter (e.g., IP 5-tuple, IP triplet), or by an application identifier, or by a domain name identifier (FQDN / URL).

[0025] In the above implementation, multiple services with the same model information can be mapped to / bound to the same Quality of Service (QoS) channel. In other words, if different services have the same model information, they are mapped to the same QoS channel, or the QoS channel is bound to the services based on the model information. This allows services using the same model information to share the QoS channel. Since services using the same model information are likely to have similar data flow characteristics, the data flows of different services have less mutual influence, thus improving model performance. Furthermore, having services using the same model information share the QoS channel helps reduce resource consumption and simplifies QoS control.

[0026] In one possible implementation, after the session management network element obtains the model information corresponding to the second service, it can determine whether there is a service quality channel among the existing service quality channels that matches the model information corresponding to the second service. If there is (such as the first service quality channel), the session management network element can map the second service to the first service quality channel; if not, the session management network element can create a new service quality channel and map the second service to the new service quality channel.

[0027] In conjunction with the first aspect, in some possible implementations of the first aspect, determining the service quality channel corresponding to the second service based on the second model information includes: when the first model information and the second model information are different, determining that the second service service corresponds to a second service quality channel, which is different from the first service quality channel.

[0028] In other words, (multiple) services corresponding to different model information can be mapped to / bound to different Quality of Service (QoS) channels. This way, services using different model information are associated with different QoS channels, and the data flows of different services will not interfere with each other, thus improving model performance.

[0029] Optionally, in this application, the model information may include model identification information and / or model feature information corresponding to the model. For example, the first model information mentioned above includes the identification information of the first model and / or the model feature information corresponding to the first model, and the second model information mentioned above includes the identification information of the second model and / or the model feature information corresponding to the second model.

[0030] The model identification information is used to identify the model. This information can take various forms, such as: a model identifier (e.g., a special string representing the model), the model's address information (e.g., accessing this address allows downloading or using the model or calling its application programming interface (API)), the model's domain name information (e.g., a Uniform Resource Locator (URL) or a Full Qualified Domain Name (FQDN), where accessing this domain name allows downloading or using the model or calling its API. Alternatively, the domain name system (DNS) can be used to query the domain name to obtain the corresponding address, and then accessing that address to download or call the model), or the model's calling function (e.g., a function name, where calling this function allows access to the model's API to obtain certain functions or download the model. Alternatively, the function can be used to call the device's local interface to access locally stored model functions). This application does not limit the form of model identification information. Model identification information can be explicit or implicit, as long as the corresponding model can be identified, obtained, or the function of the model can be called through the information.

[0031] Model feature information can be used to represent the attributes, characteristics, functions, or usage of a model. For example, model feature information includes model consumers (also known as "model users", such as any one or more of UE, radio access network (RAN) equipment, and user plane function (UPF) network elements), the direction of the model's action on the service flow (e.g., uplink or downlink or both), model functions (e.g., burst prediction, encrypted flow detection), model type, and analysis identifiers (analysis identifiers are used to indicate the function of the model. The 3rd generation partnership project (3GPP) technical specification (TS) 23.288 has defined a variety of analysis identifiers, and the model functions mentioned in this application can be used as new analysis identifiers).

[0032] Correspondingly, the above-mentioned model information is the same, including: the model identification information is the same; or, the model feature information corresponding to the model is the same; or, the model identification information is the same and the model feature information corresponding to the model is the same.

[0033] Services with the same model identification information are mapped to the same Quality of Service (QoS) channel. This allows services using the same model identification information (such as those using the same model) to share the QoS channel. Since services using the same model identification information are likely to have similar data flow characteristics, the data flows of different services have less mutual influence, thus improving model performance. Furthermore, sharing the QoS channel for services using the same model identification information helps reduce resource consumption and simplifies QoS control.

[0034] Services with identical model features are assigned to the same Quality of Service (QoS) channel. Since the data flow characteristics of these services are likely to be similar, assigning them to the same QoS channel minimizes mutual interference between data flows from different services, thus improving model performance. Furthermore, assigning services with identical model features to the same QoS channel reduces resource consumption and simplifies QoS control. "Identical model features" can refer to all model features being identical, or the model features used for QoS channel binding being identical.

[0035] For example, the first model information and the second model information mentioned above are the same, including: the identification information of the first model and the identification information of the second model are the same (e.g., the model corresponding to the first business and the model corresponding to the second business are the same); or, the model feature information corresponding to the first model and the model feature information corresponding to the second model are the same; or, the identification information of the first model and the identification information of the second model are the same, and the model feature information corresponding to the first model and the model feature information corresponding to the second model are the same.

[0036] In conjunction with the first aspect, in some possible implementations of the first aspect, the above method further includes: obtaining a first service quality parameter corresponding to the first service and a second service quality parameter corresponding to the second service; determining the service quality channel corresponding to the second service based on the second model information includes: when the first model information and the second model information are the same, and the first service quality parameter and the second service quality parameter are the same, determining the first service quality channel corresponding to the second service.

[0037] In other words, services with the same model information and the same service quality parameters are assigned to the same service quality channel.

[0038] In the above scheme, when determining the service quality channel corresponding to a service, the session management network element considers not only model information but also service quality parameters. This not only helps to reduce the impact between different services when inferring data flow on the service quality channel based on the model, but also enables services with the same service quality parameters to receive the same service quality control.

[0039] Optionally, in this application, the above-mentioned model feature information includes one or more of the following: model consumer, the direction of the model's effect on the business flow, model function, model type, analysis identifier, or whether the model feature information is used for the correspondence of the quality of service channel.

[0040] In this context, a model consumer can be understood as a device / equipment that performs inference based on a model, or a device / equipment applicable to this model (meaning that the model is only effective when running on this device / equipment, or that the model can only run on this device / equipment). A model consumer can be, for example, one or more of the following: a terminal, a user plane function network element (such as a UPF network element), or an access network device. A model consumer can also be referred to as a model user; this application does not limit the terminology.

[0041] The direction in which a model affects a service flow includes, but is not limited to, uplink, downlink, or both. This direction can also be referred to as the model direction, the used direction, or whether it's used for uplink / downlink. An uplink model means the model is used to perform inference on uplink service flows (e.g., a terminal can use the model to perform inference on uplink service flows); a downlink model means the model is used to perform inference on downlink service flows (e.g., user plane function elements can use the model to perform inference on downlink service flows); and a model that affects both uplink and downlink means the model is used to perform inference on both uplink and downlink service flows.

[0042] Model functionality can be understood as the functions a model possesses; it can also be referred to as model role, model capability, or model usage / use case. Model functionality includes, but is not limited to, one or more of the following: burst prediction, encrypted flow detection, base station resource prediction, or congestion prediction.

[0043] Model Type: Indicates whether model information is used for service quality channel mapping. This information can be a "yes" or "no" indication. When the model type indicates that the model information is used for service quality channel mapping, services with the same model information correspond to the same service quality channel; when the model type indicates that the model information is not used for service quality channel mapping, multiple services with different model information can still be bound to the same service quality channel. Regardless of whether the model information is used for service quality channel mapping, the session management network element can determine the service quality channel corresponding to the service based on other parameters, such as QoS parameters (when model information is used for service quality channel mapping, both model information and QoS parameters can be considered simultaneously, or only model information can be considered; this application does not limit this).

[0044] The analysis identifies the role of the identifier in the model.

[0045] Whether model feature information is used for the correspondence of quality of service channels can be understood as whether the session management network element considers model feature information when performing the binding of quality of service channels.

[0046] In conjunction with the first aspect, in some possible implementations of the first aspect, the method further includes: sending a first rule corresponding to the first service to the terminal and / or user plane function network element, the first rule being used to control the traffic forwarding of the quality of service channel, the first rule including first model information.

[0047] By indicating the model information corresponding to the first service to the terminal and / or user plane functional network elements, the terminal and / or user plane functional network elements can determine the correspondence between the service and the model information, thereby facilitating the terminal and / or user plane functional network elements to perform inference on the service data flow based on the corresponding model.

[0048] In conjunction with the first aspect, in some possible implementations of the first aspect, the above method further includes: sending a first indication message to the access network device, the first indication message being used to instruct the first quality of service channel to enable model-assisted quality of service control / management.

[0049] Model-assisted service quality control / management can be understood as service quality control / management based on the inference results of a model. Model-assisted service quality control / management can be replaced with AI-assisted service quality control / management, model-enhanced service quality control / management, or AI-enhanced service quality control / management.

[0050] In conjunction with the first aspect, in some possible implementations of the first aspect, the aforementioned first instruction information is also used to indicate the first model information.

[0051] For example, the aforementioned first indication information may also indicate the function of the model corresponding to the first service, such as one or more of burst prediction, encrypted flow detection, base station resource prediction, or congestion prediction.

[0052] In conjunction with the first aspect, in some possible implementations of the first aspect, the above method further includes: sending a second rule corresponding to the first quality of service channel to the access network device, the second rule including the first model information.

[0053] By indicating the model information corresponding to the first quality of service channel to the access network device, the access network device can determine the correspondence between the quality of service channel and the model information, thereby facilitating the access network device to perform inference on the data flow on the quality of service channel based on the corresponding model.

[0054] In conjunction with the first aspect, in some possible implementations of the first aspect, the above method further includes: receiving a request message from a terminal, the request message being used for request model-assisted quality of service control.

[0055] This request message can be a service quality control mechanism that assists in the request model of a specific session / service / terminal. For example, the aforementioned request message could be a session establishment request.

[0056] For example, before obtaining the first model information, the terminal can send a request message to the session management network element to request model-assisted quality of service control. Correspondingly, the session management network element can receive the request message from the terminal.

[0057] In conjunction with the first aspect, in some possible implementations of the first aspect, the above method further includes: receiving second instruction information from a unified data management network element, the second instruction information being used to instruct authorized model-assisted quality of service control.

[0058] The second instruction information may be a service quality control aided by a specific session / service / terminal authorization model.

[0059] For example, after receiving the above request message from the terminal, the session management network element can query the unified data management network element to see if the service quality control is assisted by an authorized model before obtaining the first model information. In one case, the unified data management network element instructs the service quality control to be assisted by an authorized model, and in another case, the unified data management network element instructs the service quality control to be assisted by an unauthorized model.

[0060] Secondly, this application provides a communication method that can be applied to the policy control network element side, such as the policy control network element itself, or components configured in the policy control network element (such as processors, chips, chip systems, etc.), or logic modules or software capable of implementing all or part of the policy control network element functions. This application does not limit the application in this regard. The following example uses the execution of a policy control network element.

[0061] For example, the method includes: obtaining first model information, which is used for forwarding processing of a first service; sending first policy information, which includes information about the first service and the first model information, and the first model information is further used to determine the quality of service channel corresponding to the first policy information.

[0062] The aforementioned first model information used for forwarding processing of the first service can be replaced with: the aforementioned first model information used for traffic forwarding of the first service; or, the aforementioned first model information used for service quality management of the first service; or, the aforementioned first model information used for service quality control of the first service; or, the aforementioned first model information used for customer experience of the first service. The term "used for" can be replaced with "used to influence," "used to control," "used to assist," or "used to execute."

[0063] Optionally, the aforementioned first model information is used to indicate the forwarding process of the first service, specifically including: the aforementioned first model information is used for any one or more of the following of the first service: burst prediction, encrypted flow detection, base station resource prediction, or congestion prediction.

[0064] For example, the above scheme can be applied to scenarios involving model-assisted quality of service (QoS) control. In this scheme, model information is used to bind policy information and QoS channels. That is, whether different policy information corresponds to / binds to the same QoS channel is related to the model information corresponding to the service. This makes the QoS channel binding method adaptable to the mechanism of model-assisted access network devices for QoS control. Under the above binding method, when performing inference based on the model, whether the data streams of different services are mixed together for inference fully considers the model information corresponding to the different services, which helps reduce the impact between the data streams of different services, thereby improving the accuracy of model inference.

[0065] In conjunction with the second aspect, in some possible implementations of the second aspect, obtaining the first model information corresponding to the first service includes: obtaining the first model information from local storage; or receiving the first model information corresponding to the first service from a network data analysis function / application function network element.

[0066] The first model information corresponding to the aforementioned first service can be carried in a first message. This first message may be forwarded through one or more intermediate network elements. The content / information elements (such as the first model information) of the first message remain unchanged, but the name and format (such as the order of information elements / the number of information elements) of the first message may change. For example, the first message may include the identifier of the first service and the first model information.

[0067] In one implementation, the policy control network element obtains the model information corresponding to the service from other network elements. For example, application function network elements / network data analysis function network elements can provide the policy control network element with the model information corresponding to the service, or in other words, indicate the correspondence between the service and the model information. Another implementation is that the policy control network element can locally store the model information corresponding to each service, or in other words, the correspondence between the service and the model information.

[0068] Optionally, the aforementioned first model information includes the identification information of the first model corresponding to the first service and / or the model feature information corresponding to the first model; the model feature information includes one or more of the following: model consumer, the direction of the model's effect on the business flow, model function, model type, and whether the analysis identifier or model feature information is used for the corresponding service quality channel. For an explanation of the above parameters, please refer to the first aspect, which will not be repeated here.

[0069] Thirdly, this application provides a communication method that can be applied to the access network device side, such as the access network device itself, or components configured in the access network device (such as processors, chips, chip systems, etc.), or logic modules or software capable of implementing all or part of the functions of the access network device. This application does not limit the application in this regard. The following example demonstrates the implementation of the method by the access network device.

[0070] For example, the method includes: receiving a rule corresponding to a third quality of service channel, the rule including one or more model information, the one or more model information being used for forwarding processing of the third quality of service channel; and performing forwarding processing of the third quality of service channel according to the one or more model information.

[0071] The function of performing forwarding processing on the third quality of service channel can be replaced by controlling the forwarding processing on the third quality of service channel. Forwarding processing can be replaced by traffic forwarding, customer experience, quality of service management, or quality of service control.

[0072] In the above scheme, the access network device can determine the corresponding model based on one or more model information corresponding to the QoS channel, and then use the corresponding model to perform inference on the data flow on the QoS channel, and perform QoS control on the QoS channel based on the inference results. In this way, the access network device does not need to obtain the model inference results from other network elements, which helps to reduce signaling overhead and reduce latency.

[0073] In conjunction with the third aspect, in some possible implementations of the third aspect, the above-mentioned forwarding processing of the third quality of service channel based on one or more model information includes: performing forwarding processing of the third quality of service channel based on the inference results corresponding to one or more model information.

[0074] The inference results corresponding to the above one or more model information can be understood as the inference results of one or more models corresponding to the above one or more model information.

[0075] Fourthly, this application provides a communication method that can be applied to the terminal side, such as a terminal, or components configured in the terminal (such as processors, chips, chip systems, etc.), or logical modules or software capable of implementing all or part of the terminal's functions. This application does not limit the application in this regard. The following example demonstrates terminal execution. Alternatively, this method can be applied to the user plane function network element side, such as a user plane function network element, or components configured in the user plane function network element (such as processors, chips, chip systems, etc.), or logical modules or software capable of implementing all or part of the user plane function network element's functions. This application does not limit the application in this regard.

[0076] For example, the method includes: receiving first information, the first information indicating a quality of service channel corresponding to a first service and model information corresponding to the quality of service channel; and performing inference on data packets in the quality of service channel using a first model, the first model being determined based on the first information and being a model corresponding to the quality of service channel.

[0077] In the above scheme, the terminal / user plane function network element can determine the correspondence between the service, quality of service channel and model information based on the received first information, thereby facilitating the inference of data packets in the quality of service channel based on the model corresponding to the quality of service channel.

[0078] It should be noted that in the above scheme, the Quality of Service (QoS) channel corresponds to the model information; for example, one QoS channel may correspond to one model information (or model), but this should not constitute any limitation on this application. The model information (or model) may also correspond one-to-one with a business; for example, businesses with the same model feature information correspond to the same QoS channel. In this case, one QoS channel may correspond to multiple model information (or models), but one business corresponds to one model information. Correspondingly, in the above case, the first model may refer to the model corresponding to the first business.

[0079] In conjunction with the fourth aspect, in some possible implementations of the fourth aspect, the above method further includes: sending a request message for requesting model-assisted quality of service control.

[0080] This request message can be a service quality control mechanism that assists in the request model of a specific session / service / terminal. For example, the aforementioned request message could be a session establishment request.

[0081] In conjunction with the fourth aspect, in some possible implementations of the fourth aspect, the above method further includes: indicating the inference result corresponding to the first model to the access network device.

[0082] By indicating the inference results corresponding to the first model to the access network devices, the access network devices can perform quality of service control of the quality of service channel based on the inference results.

[0083] Fifthly, this application provides a communication apparatus capable of implementing the methods described in the first to fourth aspects and any possible implementation thereof. The apparatus includes corresponding modules for performing the described methods. These modules can be implemented in software and / or hardware.

[0084] In a sixth aspect, this application provides a communication device including a processor that can execute a computer program in a memory to implement the methods described in the first to fourth aspects and any possible implementation of the first to fourth aspects.

[0085] Optionally, the communication device further includes a memory. The memory can be used to store instructions (or computer programs, or code) and data. The memory is coupled to the processor, which, when executing the instructions stored in the memory, can implement the methods described in the foregoing aspects.

[0086] Optionally, the communication device further includes a communication interface. The communication interface is used for communication between the communication device and other communication devices. For example, the communication interface may be a transceiver, circuit, bus, module, pin, or other type of communication interface.

[0087] In a seventh aspect, this application provides a computer-readable storage medium storing a computer program (or instructions or code) that, when executed, performs the method described in the first to fourth aspects and any possible implementation thereof.

[0088] Eighthly, this application provides a computer program product including instructions (or computer program, or code) that, when executed, implement the methods described in the first to fourth aspects and any possible implementation of the first to fourth aspects.

[0089] Ninthly, this application provides a chip system including a processor and potentially a memory for implementing the methods described in the first to fourth aspects and any possible implementation thereof. The chip system may be composed of chips or may include chips and other discrete devices.

[0090] It should be understood that the fifth to ninth aspects of this application correspond to the technical solutions of the first to fourth aspects of this application, and the beneficial effects obtained by each aspect and the corresponding feasible implementation are similar, and will not be repeated here. Attached Figure Description

[0091] Figure 1 is a schematic diagram of the network architecture applicable to the communication method provided in the embodiments of this application;

[0092] Figure 2 is a schematic diagram of a QoS architecture;

[0093] Figure 3 is a schematic diagram of the model's reasoning process;

[0094] Figure 4 is a schematic flowchart of the communication method provided in an embodiment of this application;

[0095] Figure 5 is a detailed flowchart of the communication method provided in the embodiments of this application;

[0096] Figure 6 is another detailed flowchart of the communication method provided in an embodiment of this application;

[0097] Figure 7 is another detailed flowchart of the communication method provided in the embodiments of this application;

[0098] Figure 8 is a further detailed flowchart of the communication method provided in the embodiments of this application;

[0099] Figure 9 is a schematic block diagram of a communication device provided in an embodiment of this application;

[0100] Figure 10 is another schematic block diagram of the communication device provided in the embodiments of this application. Detailed Implementation

[0101] To facilitate understanding of the embodiments of this application, the following points are explained first:

[0102] First, in the embodiments of this application, the terms "first" and "second" are used to distinguish identical or similar items with essentially the same function and purpose. For example, the first strategy information and the second strategy information are only used to distinguish strategy information corresponding to different businesses, and do not limit their order. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and the terms "first" and "second" are not necessarily different; for example, the first strategy information and the second strategy information may be the same or different.

[0103] 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 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.

[0104] In the embodiments of this application, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following," "one or more of the following," or similar expressions refer to any combination of these items, including any combination of singular or plural items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0105] Second, in the embodiments of this application, "send" and "receive" indicate the direction of signal transmission. For example, "send information to the second device" can be understood as the destination of the information being the second device, which may include sending directly via the air interface or sending indirectly via the air interface from other units or modules. "Receive information from the second device" can be understood as the source of the information being the second device, which may include receiving directly from the second device via the air interface or receiving indirectly from the second device via the air interface from other units or modules. "Send" can also be understood as the "output" of the chip interface, and "receive" can also be understood as the "input" of the chip interface.

[0106] In other words, sending and receiving can be done between devices, such as between a second device and a first device; or it can be done within a device, such as between components, modules, chips, software modules, or hardware modules within a device via a bus, wiring, or interface.

[0107] It is understandable that information may undergo necessary processing, such as encoding and modulation, before being sent from the source to the destination. Similarly, the destination, upon receiving information from the source, can also perform corresponding processing, such as decoding and demodulation, to interpret the valid information from the source. Similar expressions in this application can be understood in a similar way and will not be elaborated further.

[0108] Third, in the embodiments of this application, "instruction" can include direct instruction and indirect instruction, as well as explicit instruction and implicit instruction. The information indicated by a certain piece of information is called the information to be instructed. In the specific implementation process, there are many ways to instruct the information to be instructed, such as, but not limited to, directly instructing the information to be instructed, such as the information to be instructed itself or its index. It can also indirectly instruct the information to be instructed by instructing other information, where there is a relationship between the other information and the information to be instructed; or it can only instruct a part of the information to be instructed, while the other parts of the information to be instructed are known or pre-agreed upon. For example, the instruction of specific information can be achieved by using a pre-agreed (e.g., protocol predefined) arrangement of various pieces of information, thereby reducing instruction overhead to a certain extent. This application does not limit the specific method of instruction.

[0109] It is understandable that, for the sender of the instruction information, the instruction information can be used to indicate the information to be indicated, and for the receiver of the instruction information, the instruction information can be used to determine the information to be indicated.

[0110] Fourth, in the embodiments of this application, descriptions such as "when," "under the circumstances," "if," and "if" all refer to the device making corresponding processing under certain objective circumstances, and are not limited to a specific time. They do not require the device to make a judgment action during implementation, nor do they imply any other limitations.

[0111] Fifth, the tables in the embodiments of this application are merely examples, and the values ​​of the information in each table are only examples and can be configured to other values. This application does not limit the scope of protection of this application. For example, appropriate modifications and adjustments can be made based on the tables mentioned above, such as splitting, merging, etc. For another example, the parameter names shown in the headings of each table can also use other names that the communication device can understand, and the values ​​or representations of the parameters can also be other values ​​or representations that the communication device can understand. For yet another example, there can be more or fewer parameters in each table. Furthermore, when implementing the above tables, other data structures can also be used, such as arrays, queues, containers, stacks, linear lists, pointers, linked lists, trees, graphs, structures, classes, heaps, hash tables, or hash tables, etc.

[0112] Sixth, in the embodiments of this application, the messages / information (such as the first model information) exchanged between devices / network elements / equipment can be forwarded through one or more intermediate network elements, or they can be directly exchanged between devices / network elements / equipment (i.e., without the need for intermediate network element forwarding). When the messages / information exchanged between devices / network elements / equipment are forwarded through one or more intermediate network elements, the content / information elements of the information / message remain unchanged, but the name and format of the information / message (such as the order of information elements / the number of information elements) may change.

[0113] Seventh, for ease of understanding, this document provides several examples of names, such as network element names like SMF network element, PCF network element, etc., and information names like policy information, QoS rules, PCC rules, etc. These examples are merely illustrations and should not constitute any limitation on this application. This application does not preclude the possibility of using other names for related functions or messages in future protocols.

[0114] Furthermore, this application does not limit the applicable communication system. Correspondingly, the various network elements and information involved in this application can be network elements / information in various types of communication systems. For example, the SMF network element can be a network element in a 5th generation (5G) system, or a session management function network element or module in a future communication system, etc. Other network elements involved in this application are similar, and will not be listed one by one here.

[0115] Eighth, the technical solutions of this application can be applied to various communication systems, such as long term evolution (LTE) systems, 5G systems, new radio (NR) systems, and future communication systems. The network architecture applicable to this application will be described in detail below with reference to Figure 1.

[0116] Figure 1 is a schematic diagram of the network architecture applicable to the communication method provided in the embodiments of this application.

[0117] As shown in Figure 1, this network architecture includes user equipment (UE), access network (AN), core network (CN), and data network (DN). The AN is used to implement wireless physical layer functions, resource scheduling and wireless resource management, wireless access control, and mobility management. The CN may include management devices and gateway devices. The management devices are mainly used for terminal device registration, security authentication, mobility management, and location management. The gateway devices are mainly used to establish channels with terminals and forward data packets between terminals and external data networks on these channels. The DN may include network devices (such as servers and routers), and the data network is mainly used to provide various data services to terminals.

[0118] It is understood that Figure 1 illustrates a 5G network architecture based on a service-oriented interface in a non-roaming scenario, where the AN can be the RAN. The RAN can include at least one RAN node, and the UE can connect to the RAN node wirelessly. The RAN node connects to the CN wirelessly or via a wired connection. The core network equipment in the CN and the RAN node in the RAN can be different physical devices, or they can be the same physical device integrating CN and RAN logical functions.

[0119] The following section will provide a detailed explanation of each part of the above network architecture. For details on the architecture and explanations of each part, please refer to 3GPP TS23.501.

[0120] I. RAN Node

[0121] RAN nodes can provide wireless communication services, enabling terminals to access the wireless network. RAN nodes can also be called RAN equipment, access network equipment, etc., and this application does not limit the terminology. RAN nodes can be base stations (BS), evolved NodeBs (eNodeBs), transmission reception points (TRPs), home evolved NodeBs (or home Node Bs, HNBs), Wi-Fi access points (APs), mobile switching centers, next-generation NodeBs (gNBs) in 5G mobile communication systems, or base stations in future communication systems. RAN nodes can also be devices that perform base station functions in device-to-device (D2D) communication systems, vehicle-to-everything (V2X) communication systems, machine-to-machine (M2M) communication systems, and internet-to-things (IoT) communication systems. RAN nodes can also be RAN nodes in non-terrestrial networks (NTNs), meaning they can be deployed on high-altitude platforms or satellites. RAN nodes can be macro base stations, micro base stations, indoor stations, relay nodes, donor nodes, or radio controllers in cloud radio access networks (CRAN) scenarios, or nodes in open radio access networks (O-RAN or ORAN) scenarios. Optionally, RAN nodes can also be servers, wearable devices, vehicles, or in-vehicle equipment. For example, in V2X technology, RAN nodes can be roadside units (RSUs). Of course, RAN nodes can also be nodes in the core network.

[0122] In another possible scenario, multiple RAN nodes collaborate to assist the terminal in achieving wireless access, with each RAN node performing a portion of the base station's functions. For example, a RAN node can be a CU, a distributed unit (DU), a CU-control plane (CP), a CU-user plane (UP), or a radio unit (RU). CUs and DUs can be configured separately or included in the same network element, such as a baseband unit (BBU). RUs can be included in radio frequency equipment or radio frequency units, such as remote radio units (RRUs), active antenna units (AAUs), or remote radio heads (RRHs).

[0123] In different systems, CU (or CU-CP and CU-UP), DU, or RU may have different names, but those skilled in the art will understand their meaning. For example, in an open RAN (ORAN or O-RAN) system, CU can also be called open CU (O-CU), DU can also be called open DU (O-DU), CU-CP can also be called open CU-CP (O-CU-CP), CU-UP can also be called open CU-UP (O-CU-UP), and RU can also be called open RU (O-RU).

[0124] Any one of the CU (or CU-CP, CU-UP), DU, and RU units can be implemented through software modules, hardware modules, or a combination of software and hardware modules. That is, the wireless access network device in this application can be a virtualized device, for example, implemented through general-purpose hardware and instantiated virtualization functions, or dedicated hardware and instantiated virtualization functions. The general-purpose hardware can be a server, such as a cloud server.

[0125] 2. UE

[0126] A UE can be a device that provides voice / data connectivity to a user, such as a handheld device or vehicle-mounted device with wireless connectivity. A UE can also be referred to as: terminal, mobile station (MS), mobile terminal (MT), access terminal, user unit, user station, mobile station, mobile terminal, remote station, remote terminal, mobile device, user terminal, terminal equipment, wireless communication equipment, user agent, or user equipment, etc. This application does not limit the scope of this name.

[0127] Currently, examples of terminals include: mobile phones, tablets, laptops, PDAs, mobile internet devices (MIDs), wearable devices, virtual reality (VR) devices, augmented reality (AR) devices, wireless terminals in industrial control, wireless terminals in self-driving vehicles, wireless terminals in remote medical surgery, wireless terminals in smart grids, wireless terminals in transportation safety, wireless terminals in smart cities, wireless terminals in smart homes, cellular phones, cordless phones, session initiation protocol (SIP) phones, wireless local loop (WLL) stations, personal digital assistants (PDAs), handheld devices with wireless communication capabilities, computing devices or other processing devices connected to wireless modems, in-vehicle devices, wearable devices, terminals in 5G networks, or future public land mobile communication networks. This application does not limit the scope to terminals in a network (PLMN), etc.

[0128] As an example and not a limitation, the terminal can be a terminal in an Internet of Things (IoT) system. The IoT is an important component of future information technology development. Its main technical characteristic is connecting objects to networks through communication technologies, thereby realizing an intelligent network of human-machine interconnection and object-to-object interconnection. Exemplarily, the terminal in this application embodiment can be a wearable device. Wearable devices, also known as wearable smart devices, are a general term for devices that utilize wearable technology to intelligently design and develop everyday wearables, such as glasses, gloves, watches, clothing, and shoes. Wearable devices are portable devices that can be worn directly on the body or integrated into a user's clothing or accessories. Wearable devices are not merely hardware devices; they can also achieve powerful functions through software support, data interaction, and cloud interaction. Broadly defined, wearable smart devices include those with comprehensive functions, large size, and the ability to achieve complete or partial functions without relying on a smartphone, such as smartwatches or smart glasses, as well as those focused on a specific application function and requiring the use of other devices such as smartphones, such as various smart bracelets and smart jewelry for vital sign monitoring.

[0129] By way of example and not limitation, the terminal can also be a terminal in machine-type communication (MTC). Furthermore, the terminal can also be an on-board module, on-board component, on-board chip, or on-board unit built into a vehicle as one or more components or units. The vehicle can implement the methods provided in this application through the built-in on-board module, on-board component, on-board chip, or on-board unit. Therefore, the embodiments of this application can also be applied to vehicle-to-everything (V2X) networks, such as vehicle-to-everything (V2X), long-term evolution-vehicle (LTE-V) technology, and vehicle-to-vehicle (V2V) technology.

[0130] III. CN

[0131] In the aforementioned 5G network architecture, the CN can include multiple functional units such as user plane function (UPF) network elements, authentication server function (AUSF) network elements, access and mobility management function (AMF) network elements, session management function (SMF) network elements, network slice selection function (NSSF) network elements, network exposure function (NEF) network elements, network repository function (NRF) network elements, policy control function (PCF) network elements, unified data management (UDM) network elements, application function (AF) network elements, service communication proxy (SCP) network elements, network slice specific authentication and authorization function (NSSAAF) network elements, network slice authentication and authorization control function (NSACF) network elements, and location management function (LMF) network elements.

[0132] Among them, the UPF network element is responsible for packet routing and forwarding, packet filtering, and QoS control related functions for external connections to the DN and user plane. The AUSF network element is responsible for UE authentication. The AMF network element is responsible for mobility management and access management services. The SMF network element is responsible for session management, UE address management and allocation, Dynamic Host Configuration Protocol (DHCP) functions, and user plane function selection and control. The NRF network element is responsible for service registration and service discovery, maintaining the NF text of available network function (NF) instances (also known as network function entities) and the services they support. The PCF network element is responsible for providing a unified policy framework for network behavior management, providing policy rules for control plane functions, and obtaining registration information related to policy decisions. The UDM network element supports authentication trust processing, user identity processing, access authorization, registration and mobility management, subscription management, and short message management in the 3GPP authentication and key negotiation mechanism. The AF network element can interact with the 3GPP core network to provide services, including interaction with the NEF and policy architecture. NF consumers and NF producers can communicate through SCP network element proxies to achieve converged connections, simplified connections, and other similar functions.

[0133] It is understood that in this application, the NF consumer can be understood as the party using the service, that is, the user of the service; and the NF producer can be understood as the party providing the service, that is, the provider of the service.

[0134] It should be noted that the above-mentioned functional units can work independently or be combined to implement certain control functions, such as access control and mobility management functions for terminal devices, including access authentication, security encryption, and location registration, as well as session management functions such as the establishment, release, and modification of user plane transmission paths.

[0135] In addition, the above-mentioned network elements can also be referred to as units, modules, entities, functions, functional entities, functional modules, functional units, etc. This application does not limit the names.

[0136] In the 5G core network, functional units can communicate with each other through the next generation (NG) interface. For example, the UE can transmit control plane messages with the AMF network element through NG interface 1 (N1); the AN / RAN equipment can establish a control plane signaling connection with the AMF network element through NG interface 2 (N2); the RAN equipment can establish a user plane data transmission channel with the UPF network element through NG interface 3 (N3); the UPF network element can interact with the SMF network element through NG interface 4 (N4); the UPF network element can interact with the DN network element through NG interface 6 (N6); UPF network elements can communicate with each other through NG interface 9 (N9); and the AMF network element can interact with the SMF network element through NG interface 11 (N11) (not shown in Figure 1), etc.

[0137] In service-oriented interface scenarios, some network elements in the 5G core network can be connected via a bus, as shown in Figure 1. AUSF, AMF, NSACF, SMF, NSSF, NEF, NRF, PCF, UDM, LMF, and AF network elements are interconnected via a bus. When these network elements are interconnected via the bus, they use service-oriented interfaces. For example, AUSF network elements connect to the bus via the Nausf interface, AMF network elements connect via the Namf interface, SMF network elements connect via the Nsmf interface, NSSF network elements connect via the Nnssf interface, NEF network elements connect via the Nnef interface, NRF network elements connect via the Nnrf interface, PCF network elements connect via the Npcf interface, UDM network elements connect via the Nudm interface, AF network elements connect via the Naf interface, and so on.

[0138] It should be noted that Figure 1 is only an exemplary architecture diagram. In addition to the functional units shown in Figure 1, the network architecture may also include other functional units, which are not limited in this application.

[0139] The above text provides specific details about this application and the network architecture to which this application applies. To facilitate understanding of the background technology related to this application described below, some technical terms involved in this application will be explained in detail below.

[0140] 1. QoS Rules: Used by terminals to mark and process data flows. QoS rules can include identifiers and traffic filtering conditions. Identifiers can include, for example, QoS flow identifiers (QFIs) and 5G QoS identifiers (5QIs). A QFI can be understood as an identifier for a specific QoS flow, which can be used to mark data packets in the Service Data Adaptation Protocol (SDAP) header or the GPRS tunneling protocol-user plane (GTP-U) header. A 5QI is an identifier associated with a specific QoS flow, used to define the quality of service characteristics (such as priority, latency, etc.) of that QoS flow. Traffic filtering conditions can include, for example, one or more of the following: source address, destination address, source port number, destination port number, or protocol type (or protocol characteristics). For example, traffic filtering conditions can be IP triples, such as source address, source port number, and protocol type. For example, traffic filtering conditions can be IP 5-tuples, such as source address, destination address, source port number, destination port number, and protocol type.

[0141] For a more detailed explanation of QoS rules, please refer to 3GPP TS23.501 and 23.503.

[0142] 2. QoS Profile: A QoS profile is a predefined set of settings used to describe the quality of service requirements for specific types of traffic. These requirements include, but are not limited to, requirements for one or more parameters: bandwidth, latency, jitter, or packet loss rate. The QoS profile is used to control the scheduling of different data streams by access network devices.

[0143] 3. QoS Enforcement Rule (QER): This refers to the specific rules by which user plane function (UPF) network elements enforce QoS policies. QER includes, but is not limited to, one or more of the following: QFI, traffic shaping, and rate limiting parameters. Specifically, UPF network elements add QFI to the GTP-U header to identify and process data flows. Traffic shaping and rate limiting parameters indicate how to shape and rate-limit data flows to control traffic rates and network resource usage. For a more detailed explanation of QER, please refer to 3GPP TS23.501; it will not be elaborated upon here.

[0144] 4. Packet Detection Rule (PDR): Used to identify and process data flows. PDR can be associated with QER, which defines the QoS policy applied to the data flow. When a packet that meets the PDR requirements (also known as meeting traffic filtering conditions or packet filter set) is detected, the associated QER action is executed. Traffic filtering conditions are used to identify specific data flows, including but not limited to: source address, destination address, protocol type (or protocol characteristics, protocol features, etc.), source port number, or destination port number. The associated QER action executed by the UPF network element may include: when the UPF network element detects packets that meet different PDR requirements, associating these packets with the corresponding QoS flow (i.e., filling in the corresponding value in the QFI field of the GTP-U header).

[0145] The aforementioned QoS rules can be issued by the SMF network element to the terminal, the aforementioned QoS configuration file can be issued by the SMF network element to the access network device, and the aforementioned QER can be sent by the SMF network element to the UPF network element. The SMF network element is determined based on PCC rules, which will be explained in detail below. Furthermore, the SMF network element can also perform QoS flow binding based on PCC rules.

[0146] 5. PCC Rules: Sent from the PCF network element to the SMF network element, indicating the policies and charging control methods corresponding to different service data flows (SDFs) (hereinafter referred to as service flows). For a detailed explanation of PCC rules, please refer to 3GPP TS23.501 and 23.503, which will not be elaborated here.

[0147] For example, the PCF network element generates policy information for each UE / session / service based on pre-configured policies, or information such as UE subscription information and requirements provided by third parties, combined with local logic. PCC rules are one type of policy information. For instance, if a UE has subscribed to the acceleration service for game A, the PCC rules will have stricter QoS requirements for the SDF of that game service, such as lower packet latency budget and higher guaranteed bit rate.

[0148] After receiving a PCC rule, an SMF network element can perform QoS flow binding, or in other words, bind the received PCC rule to a QoS flow (or correspond it to a QoS flow), or determine the QoS flow corresponding to the PCC rule.

[0149] More specifically, a PCC rule includes one or more QoS parameters (or QoS features), where at least one QoS parameter (such as 5QI or ARP) is used for QoS flow binding. When multiple PCC rules use the same parameter for QoS flow binding, these multiple PCC rules can be bound to the same QoS flow. For a detailed explanation of QoS flow binding, please refer to 3GPP TS23.501 and 23.503, which will not be elaborated upon here.

[0150] In addition to binding QoS flows, SMF network elements can also generate QoS rules, QoS configuration files, and QERs based on PCC rules, and send the QoS rules to the terminal, the QoS configuration files to the access network device, and the QERs to the UPF network element, so that the terminal, access network device, and UPF can perform QoS control.

[0151] The following section, in conjunction with Figure 2, will provide a detailed explanation of the QoS control performed by the terminal, access network equipment, and UPF.

[0152] Figure 2 is a schematic diagram of a QoS architecture. It should be understood that Figure 2 uses the processing of downstream data packets as an example.

[0153] For example, as shown in Figure 2, the UPF network element can process service flows according to QER. When the UPF network element detects data packets that meet different PDR requirements, it associates these data packets with the corresponding QoS flows (that is, fills in the corresponding value in the QFI field of the GTP-U header). For example, service flows 1 and 2 are associated with QoS flow 1, service flow 3 is associated with QoS flow 2, and service flow 4 is associated with QoS flow 3. Further, the UPF network element can send the above data packets to the access network device, such as through a GTP-U tunnel. After receiving the above data packets, the access network device can schedule and control the data packets according to the QoS configuration file, and then send the data packets to the terminal through the data radio bearer (DRB). For example, DRB 1 is used to carry QoS flow 1, and DRB 2 is used to carry QoS flows 2 and QoS flows 3. After receiving the data packets, the terminal can associate the data packets with the corresponding QoS flows and process them based on QoS rules. A more detailed explanation of the above QoS architecture can be found in 3GPP TS23.501, and will not be elaborated here.

[0154] With the development of artificial intelligence, it has become possible to deploy AI models on user plane devices (such as UPF network elements, RAN devices, and UEs) and infer the arrival patterns of future data packets based on these models (e.g., a data packet of size x MB will arrive in 1 second). Such models can be called user plane models. It should be noted that "user plane model" is merely an exemplary name and does not constitute a limitation on the model's function or deployment location. Any AI model that can be used for the functions or transmission processes mentioned in this application should fall within the protection scope of this application.

[0155] Inference based on the user plane model allows access network devices to perform QoS control based on the inference results. For example, as shown in Figure 3, the model's input parameters (hereinafter referred to as input parameters) may include the size of the data packet, the arrival interval of the data packet, etc. (or the data packet itself may be used directly as the model's input parameters). The model performs inference to obtain the inference results, and the model's output parameters (hereinafter referred to as output parameters) may include the size of the next / multiple data packets, their arrival time, etc. Based on the above inference results, the access network device can perform resource scheduling (an example of QoS control).

[0156] After introducing the user plane model, access network devices can assist in QoS control based on the inference results, such as assisting access network devices in resource scheduling. However, the binding method of QoS flows by SMF network elements described above may not be compatible with the above mechanism (i.e., the inference results of the model assist access network devices in QoS control).

[0157] Therefore, it is desirable to provide a QoS flow binding method that can adapt to the mechanism of QoS control by assisting access network devices based on model-based inference results.

[0158] In view of this, this application provides a communication method in which the session management network element (such as the SMF network element) considers the model information corresponding to the service when determining the service quality channel (such as the QoS flow) corresponding to the service. In this way, the binding method of the service quality channel is adapted to the mechanism of QoS control assisted by the inference results of the model.

[0159] For example, as long as a service corresponds to model information, the service is assigned to a separate QoS channel. This ensures that the input parameters of the user plane model are separate data streams for the service, thereby improving model performance and correctly influencing the service through QoS streams.

[0160] For example, when different services share the same model identifier information, these different services can be mapped to the same Quality of Service (QoS) channel. This allows multiple services using the same model to share the QoS channel. Since services using the same model are likely to have similar data flow characteristics, the data flows of different services have less mutual influence, thus improving model performance. Furthermore, having multiple services using the same model share the QoS channel helps reduce resource consumption and simplifies QoS control.

[0161] For example, services with similar model characteristics can be assigned to the same Quality of Service (QoS) channel. Since the data flow characteristics of these services may have relatively small differences, assigning them to the same QoS channel minimizes the mutual influence between the data flows of different services, thus improving model performance. Furthermore, assigning services with similar model characteristics to the same QoS channel reduces resource consumption and simplifies QoS control.

[0162] The above implementation is merely an example. For specific implementation methods of determining the service quality channel corresponding to a business based on model information, please refer to the embodiments shown below, which will not be detailed here.

[0163] The communication method provided in this application will be described in detail below with reference to the accompanying drawings. The communication method provided in this application is described in detail below using a session management network element as an example, but this should not be construed as limiting the application in any way. The session management network element can also be replaced by components configured in the session management network element (such as chips, chip systems, processors, etc.), or logical modules or software capable of implementing all or part of the functions of the session management network element. For example, the method described below can be applied to the network architecture shown in Figure 1.

[0164] Figure 4 is a schematic flowchart of the communication method 400 provided in an embodiment of this application. The steps of method 400 are described in detail below.

[0165] In step 401, the session management network element obtains the first model information, which is used for the forwarding processing of the first service.

[0166] The session management network element can be used to manage sessions (such as establishing, modifying, and releasing sessions). It can also be used for UE address management and allocation, and user plane function selection and control. For example, the session management network element can be an SMF network element. This application does not limit the specific name of the session management network element, and does not exclude the possibility of using other names in future protocols. The session management network element can be a separate network element device or a module (software or hardware module) of a network element.

[0167] It should be noted that, in this application, the first service can refer to the data / data packet / service flow / data stream of the first service, or the data / data packet / service flow / data stream corresponding to the first service, or the data / data packet / service flow / data stream used for the first service, or the data / data packet / service flow / data stream sent or received by the application server corresponding to the first service. For example, the first service can be identified using a service data stream / service data stream filter (e.g., IP 5-tuple, IP 3-tuple), or using an application identifier, or using a domain name identifier (FQDN / URL). The second service described below is similar and will not be detailed here.

[0168] The first model information mentioned above may refer to the model information corresponding to the first business, and the second model information mentioned below may refer to the model information corresponding to the second business.

[0169] In this application, model information may include model identification information and / or model feature information. The model identification information and model feature information will be explained in detail below.

[0170] Model identification information is used to identify the model. Model identification information can take various forms, such as: a model identifier (e.g., a special string representing the model), model address information (e.g., accessing this address allows downloading or using the model or calling its API), model domain name information (e.g., URL, FQDN; accessing this domain name allows downloading or using the model or calling its API. Alternatively, a DNS lookup can be used to obtain the address corresponding to the domain name, and then accessing that address to download or call the model), or a model function (e.g., a function name; calling this function allows access to the model's API to obtain certain functions of the model or download the model by calling the API. Alternatively, the function can be used to call the device's local interface to access locally stored model functions). This application does not limit the form of model identification information; it can be explicit or implicit, as long as the information can identify the corresponding model, obtain the corresponding model, or call the model's functions.

[0171] For example, model feature information can be used to represent the attributes, characteristics, functions, or usage of a model. For instance, model feature information may include one or more of the following: model consumer, aspects of the model's impact on business flows, model function, model type, and analysis identifier. The above examples of model feature information are merely illustrative; model feature information may also include other information used to represent the attributes, characteristics, functions, or usage of a model, without limitation.

[0172] The parameters of the above model feature information will be explained in detail below.

[0173] In this context, a model consumer, also known as a model user, can be understood as a device / equipment that performs inference based on a model, or a device / equipment applicable to this model (meaning that the model is only effective when running on this device / equipment, or that the model can only run on this device / equipment). A model consumer can be, for example, one or more of the following: a terminal, a user plane function network element (such as a UPF network element), or an access network device. The term "model consumer" is not limited to "model user" in this application.

[0174] The direction in which a model affects a service flow can also be referred to as the model direction, the model's used direction, or whether it's used for uplink / downlink. Examples include, but are not limited to, uplink, downlink, or both. An uplink model's effect on a service flow means that the model is used to perform inference on uplink service flows (e.g., a terminal can perform inference on uplink service flows through the model); a downlink model's effect on a service flow means that the model is used to perform inference on downlink service flows (e.g., user plane function elements can perform inference on downlink service flows through the model); a model's effect on a service flow that is both uplink and downlink means that the model is used to perform inference on both uplink and downlink service flows.

[0175] Model functionality can be understood as the functions a model possesses. Model functionality can also be referred to as model role, model capability, or model usage / use case. Model functionality includes, but is not limited to, one or more of the following: burst prediction, encrypted flow detection, base station resource prediction, or congestion prediction.

[0176] Model Type: Indicates whether model information is used for service quality channel mapping. This information can be a "yes" or "no" indication. When the model type indicates that the model information is used for service quality channel mapping, services with the same model information correspond to the same service quality channel, and services with different model information correspond to different service quality channels. When the model type indicates that the model information is not used for service quality channel mapping, multiple services with different model information can still be bound to the same service quality channel. Regardless of whether the model information is used for service quality channel mapping, the session management network element can determine the service quality channel corresponding to the service based on other parameters, such as QoS parameters (when model information is used for service quality channel mapping, both model information and QoS parameters can be considered simultaneously, or only model information can be considered; this application does not limit this).

[0177] Analysis identifiers are used to identify the function of a model. 3GPP TS23.288 has defined a variety of analysis identifiers (analytics IDs). The model functions mentioned in this application can be used as new analysis identifiers, or a combination of one or more existing identifiers can be used to describe the model functions mentioned in this application.

[0178] The following is an example of model information based on Table 1. As shown in Table 1, the model information includes model identification information and model feature information.

[0179] Table 1

[0180] As an example, in the model information, model identification information is mandatory (this is merely an example and should not constitute any limitation on this application; for example, model identification information can also be optional, such as when model information includes model feature information), while model feature information is optional. In some scenarios, certain items of the model feature information may be mandatory. Several possible scenarios will be listed below.

[0181] Scenario 1: When a model with the same model identifier must be used simultaneously on multiple devices, the model consumer is mandatory. For example, a model (ID A) used for burst prediction needs to predict both uplink and downlink. Uplink inference is performed on the UE, and downlink inference is performed on the UPF network element. In this case, the same model identifier may represent a set of models, such as a model deployed on the UE for uplink inference and a model deployed on the UPF for downlink inference. Therefore, the model consumer is mandatory. The model information for inference performed on the UE is (Model ID: A, Model Consumer: UE), and on the UPF network element it is (Model ID: A, Model Consumer: UPF network element). In this situation, the UE and UPF need to determine the model to use based on the model identifier information and the model consumer / user.

[0182] Scenario 2: When models need to perform uplink and downlink inference simultaneously on the same device, and the models used for uplink and downlink inference are different but share the same model identification information, the direction of the model's action on the service flow is mandatory. For example, a model used for encrypted flow detection needs to detect both uplink and downlink flows on the UPF network element and infer the service type of the uplink and downlink encrypted flows. In this case, the model direction is mandatory (Model ID: A, Model Direction: Uplink) (Model ID: A, Model Direction: Downlink).

[0183] Scenario 3: When the device using the model supports models with multiple functions, the model function is mandatory. For example, if a UPF network element can use the model for both burst prediction and encrypted flow detection, then the UPF network element needs to be indicated as to which function the current model is used for (e.g., burst prediction).

[0184] It should be noted that when model feature information is used for the correspondence of quality of service channels, it can mean that all parameters in the model feature information are used for the correspondence of quality of service channels; or it can mean that some parameters in the model feature information are used for the correspondence of quality of service channels.

[0185] Regarding which one or more parameters in the above model feature information are used for the correspondence of the quality of service channel, one possible design is that the above model feature information can also include the attribute corresponding to each parameter, which indicates whether the parameter is used for the correspondence of the quality of service channel.

[0186] For example, as shown in Table 2, each parameter corresponds to whether it is used for the correspondence of the Quality of Service (QoS) channel. For instance, model identification information is not used for the correspondence of the QoS channel, while model consumer information is used for the correspondence of the QoS channel, etc., which will not be listed here. It should be understood that whether the parameters shown in Table 2 are used for the correspondence of the QoS channel is only an example. For example, some parameters in the model feature information may also be used for the correspondence of the QoS channel.

[0187] Table 2

[0188] The aforementioned first model information is used for the forwarding processing of the first service. This can be understood as follows: the first model corresponding to the aforementioned first model information or the reasoning result obtained by performing reasoning on the data flow of the first service using the first model can be used for the forwarding processing of the first service; or, the aforementioned first model information can instruct the access network device to reserve / schedule resources for the first service, etc.

[0189] More specifically, the aforementioned first model information is used for the forwarding processing of the first service, specifically including: the aforementioned first model information can be used for one or more of the following: burst prediction (such as predicting the size and arrival time of the next / multiple data packets corresponding to the first service), encrypted flow detection, base station resource prediction, or congestion prediction.

[0190] Optionally, the aforementioned first model information used for forwarding processing of the first service can be replaced by the aforementioned first model information being used for service quality control, service quality management, customer experience, or traffic forwarding of the first service, etc., and this application does not limit this. Customer experience can also be referred to as user experience, etc.

[0191] It should be noted that, in this application, the aforementioned first model information is used for the service quality control / service quality management / customer experience / traffic forwarding of the first business. Here, "used for" can be replaced with "used to influence", "used to control", "used to assist", or "used to execute", etc.

[0192] Understandably, the first service described above is merely an example and should not constitute any limitation on this application. For instance, the session management network element can also obtain model information for other services, and the model information corresponding to each service can be used for the forwarding processing of that service.

[0193] In this application, the services may be, for example, multimedia services (such as video and audio services), real-time media services, cloud gaming services, SMS services, etc., wherein the video services may be video services of different applications.

[0194] The session management network element obtains first model information. Implementation method one: The session management network element receives first model information corresponding to the first service from the policy control network element. This implementation method one can be applied, for example, to scenarios where the session management network element and the policy control network element are deployed separately. The policy control network element can be responsible for providing a unified policy framework for network behavior management, providing policy rules for control plane functions, and obtaining registration information related to policy decisions. The policy control network element can be, for example, a PCF network element. This application does not limit the specific name of the policy control network element and does not exclude the possibility of using other names in future protocols. For example, the SMF network element receives first model information and the identifier of the first service from the PCF network element; this first model information is used for the forwarding processing of the first service.

[0195] It should be noted that, in this application, the identifier or information of the first service can be used to identify the first service. The aforementioned information / identifier of the first service may include, but is not limited to, one or more of the following: service identifier, application identifier corresponding to the service, IP triplet, or IP quintuple. The IP triplet may, for example, be source address, source port number, and protocol type; the IP quintuple may, for example, be source address, destination address, source port number, destination port number, and protocol type.

[0196] The session management network element obtains the first model information. Implementation method two: The session management network element obtains the first model information corresponding to the first service from local storage. The above implementation method can be applied to scenarios where the session management network element and the policy control network element are co-deployed. For an introduction to the policy control network element, please refer to implementation method one; it will not be repeated here.

[0197] The session management network element obtains the first model information, in implementation method three: the session management network element obtains the first policy information, which includes the aforementioned first model information and the aforementioned first service information. The aforementioned first policy information can indicate the policy and billing control method corresponding to the first service. The aforementioned first policy information can be, for example, the PCC rule corresponding to the first service, or the PCC rule of the first service, or include the PCC rule of the first service. For an explanation of the aforementioned first service information, please refer to implementation method one; it will not be repeated here.

[0198] The session management network element obtains the first policy information. One possible implementation is that the session management network element receives the first policy information from the policy control network element. This implementation method is applicable, for example, to scenarios where the session management network element and the policy control network element are deployed separately. Another possible implementation is that the session management network element obtains the first policy information from local storage. This implementation method is applicable, for example, to scenarios where the session management network element and the policy control network element are deployed centrally.

[0199] It should be noted that the aforementioned first model information and the aforementioned first service information can also be obtained separately. For example, one item (such as the first model information) can be pre-configured, that is, the first model information can be obtained from local storage, while the other item (such as the first service information) can be obtained from other network elements, such as the first service information being obtained from the policy control network element. The reverse is also true. Alternatively, the first model information and the first service information can be obtained from different network elements.

[0200] Regarding the implementation method of the session management network element receiving the first policy information from the policy control network element, the policy control network element can generate the first policy information and send the first policy information to the session management network element. Before that, the policy control network element can obtain the first model information in the first policy information.

[0201] One possible implementation for the policy control network element to obtain the first model information is that it obtains the first model information from local storage or database / storage function network elements (such as unified data repository (UDR) network elements). Optionally, it can also obtain the identifier of the first service corresponding to the first model information.

[0202] Another possible implementation is that the policy control network element receives first model information corresponding to the first service from the network data analysis function / application function network element. For example, the network data analysis function network element (such as the NWDAF network element) sends the identifier of the first service and the first model information corresponding to the first service to the policy control network element, and correspondingly, the policy control network element receives the first model information from the network data analysis function network element. Then, the policy control network element generates the aforementioned first policy information based on the first model information provided by the network data analysis function network element.

[0203] For example, an application function network element (such as an AF network element) sends the identifier of a first service and the first model information corresponding to the first service to the policy control network element. Correspondingly, the policy control network element receives the first model information from the application function network element. The application function network element may also send the terminal identifier, session attributes, etc. to the policy control network element.

[0204] It should be noted that the first model information corresponding to the first service mentioned above can be carried in the first message. The first message can be forwarded through one or more intermediate network elements. The content / information elements (such as the first model information) of the first message remain unchanged, but the name and format (such as the order of information elements / the number of information elements) of the first message may change.

[0205] In step 402, the session management network element determines the first quality of service channel corresponding to the first service based on the first model information mentioned above.

[0206] Traffic belonging to the aforementioned first Quality of Service (QoS) channel is provided with the same forwarding processing. A QoS channel can be, for example, the finest / smallest granularity of QoS differentiation, allowing the network to provide the same service / service experience / forwarding processing / QoS management / QoS control / traffic forwarding / customer experience for services / traffic / data packets belonging to the same QoS channel. The aforementioned QoS channel can be, for example, a QoS flow, bearer, radio bearer, or connection, etc.

[0207] The session management network element determines the first quality of service (QoS) channel corresponding to the first service based on the aforementioned first model information. Alternatively, the session management network element may perform the binding of the first service and the first QoS channel based on the first model information; or, the aforementioned first model information may be used for binding / corresponding QoS channels; or, the aforementioned first model information may be bound to the first QoS channel. The session management network element determines the first QoS channel corresponding to the first service based on the aforementioned first model information, including the following possible implementation methods:

[0208] Implementation Method A: When the first service corresponds to the first model information, the session management network element determines that the first service corresponds to the first quality of service (QoS) channel. This first QoS channel corresponds only to the first service, or in other words, this first QoS channel does not correspond to any other service besides the first service. Alternatively, the session management network element determining that the first service corresponds to the first QoS channel can be replaced by the session management network element binding the first service to the first QoS channel, or binding the first service to a specific / separate / dedicated QoS channel, or establishing a specific / separate / dedicated QoS channel for the first service, or determining that the first service corresponds to a separate / specific / dedicated QoS channel. Here, a separate / specific / dedicated QoS channel means that the QoS channel corresponds to / is bound to only one service.

[0209] In summary, as long as the service corresponds to the model information, or in other words, the terminal or the terminal's session supports / is authorized to perform service quality control / management based on the model-assisted service, the session management network element can bind the service with the corresponding model information to a separate service quality channel.

[0210] For example, if a terminal or its session supports / is authorized to perform service quality control / management based on model-assisted services, and the terminal / semester includes multiple services, at least one of which corresponds to model information, then each of these services is bound to a separate service quality channel. For instance, if the terminal / semester supports / is authorized to perform service quality control / management based on model-assisted services, and the session includes service 1, service 2, service 3, and service 4, where service 1 and service 2 both correspond to model information, while service 3 and service 4 do not, then service 1 is bound to a separate service quality channel, and service 2 is also bound to a separate service quality channel. Alternatively, service 3 and service 4 can, for example, determine whether to bind to the same service quality channel based on service quality parameters.

[0211] It should be noted that in this application, "model-assisted service quality control / management" can be replaced with "AI-assisted service quality control / management," "model-enhanced service quality control / management," or "AI-enhanced service quality control / management." These will not be explained further below.

[0212] For example, if service 1 corresponds to model information 1 and service 2 corresponds to model information 2, then the session management network element determines that these two services correspond to separate quality of service (QoS) channels, such as service 1 corresponding to QoS channel 1 and service 2 corresponding to QoS channel 2. Both QoS channel 1 and QoS channel 2 are separate QoS channels. For instance, QoS channel 1 only includes service 1 and does not include any other services. In other words, as long as a service corresponds to model information, regardless of whether the model information is the same or whether the QoS parameters (such as 5QI, ARP, etc.) are the same, the service will be mapped to a separate QoS channel.

[0213] It should be noted that when the first service does not correspond to model information, the session management network element can determine the QoS channel corresponding to the first service based on the QoS parameters. For specific details, please refer to the existing introduction to QoS flow binding methods above; they will not be repeated here.

[0214] In addition, in this application, "correspondence" and "binding" have the same meaning and can be replaced. "The session management network element binds the service to the quality of service channel" can be replaced with "The session management network element corresponds the service to the quality of service channel".

[0215] Implementation Method B: Services with the same model information are assigned to the same Quality of Service (QoS) channel. In other words, services with the same model information are bound to the same QoS channel. Alternatively, when different services have the same model information, these services can be assigned to / bound to the same QoS channel. Conversely, services with different model information are assigned to different QoS channels. In other words, when different services have different model information, these services can be assigned to / bound to different QoS channels. Alternatively, the QoS channel for a service is bound based on its model information.

[0216] Taking the first and second services as examples, the above implementation method B will be explained in detail. Method 400 shown in Figure 4 further includes: the session management network element obtaining second model information, which is used for forwarding processing of the second service. For an explanation of the second model information and how it is obtained, please refer to the explanation of the first model information; it will not be repeated here.

[0217] When the first model information and the second model information are the same, the session management network element determines that the second service corresponds to the aforementioned first quality of service channel; that is, the first service and the second service correspond to the same quality of service channel. Alternatively, the first quality of service channel corresponds to one or more services that use the first model information, including the first service; or, the aforementioned first quality of service channel corresponds to services that use the first model information (such as the first service and the second service).

[0218] When the first model information and the second model information are different, the session management network element determines that the second service corresponds to the second quality of service channel. This second quality of service channel is different from the first quality of service channel. That is, the first service and the second service correspond to different quality of service channels.

[0219] For example, for service A, after the session management network element obtains the model information of service A, it can determine whether there is a service quality channel among the existing service quality channels that matches the model information of service A. If there is (such as service quality channel A, which has already been associated with / bound to service B), the session management network element can map service A to the aforementioned service quality channel A; if not, the session management network element can create a new service quality channel (such as service quality channel B) and map service A to the aforementioned service quality channel B.

[0220] Alternatively, the model information for the aforementioned service quality channel binding can include the following possible designs:

[0221] Design 1 uses model identification information for service quality channel binding; Design 2 uses model feature information; Design 3 uses both model identification information and model feature information. The following section will describe in detail the service quality channel binding using these three possible designs.

[0222] Implementation method B1 (corresponding to Design 1 above): Services with the same model identifier information correspond to the same Quality of Service (QoS) channel; services with different model identifier information correspond to different QoS channels. Taking the first and second services mentioned above as examples, implementation method B1 will be explained in detail.

[0223] When the identification information of the first model (the model corresponding to the first service) is the same as the identification information of the second model (the model corresponding to the second service), the session management network element determines that the second service corresponds to the aforementioned first quality of service channel; that is, the first service and the second service correspond to the same quality of service channel. Alternatively, the aforementioned first quality of service channel corresponds to one or more services using the first model identification information, including the first service; or, the aforementioned first quality of service channel corresponds to services using the first model identification information (such as the first service and the second service).

[0224] For example, if the model identifier information corresponding to service A is the same as that corresponding to service B, then service A and service B can correspond to the same quality of service channel.

[0225] When the identification information of the first model (the model corresponding to the first service) is different from the identification information of the second model (the model corresponding to the second service), the session management network element determines that the second service corresponds to the second quality of service channel. This second quality of service channel is different from the first quality of service channel. That is, the first service and the second service correspond to different quality of service channels.

[0226] Implementation method B2 (corresponding to Design 2 above): Services with the same model feature information correspond to the same quality of service channel; services with different model feature information correspond to different quality of service channels.

[0227] The same model feature information can mean that all corresponding model feature information is the same, or that the model feature information used for binding the quality of service channel is the same (for example, each feature information has a corresponding attribute, i.e., whether the feature information is used for binding the quality of service channel).

[0228] Using the first and second business processes mentioned above as examples, the implementation method B2 will be explained in detail.

[0229] When the model feature information corresponding to the first model (the model corresponding to the first service) is the same as the model feature information corresponding to the second model (the model corresponding to the second service), the session management network element determines that the second service corresponds to the aforementioned first quality of service channel, that is, the first service and the second service correspond to the same quality of service channel.

[0230] When the model feature information corresponding to the first model (the model corresponding to the first service) is different from the model feature information corresponding to the second model (the model corresponding to the second service), the session management network element determines that the second service corresponds to the second quality of service channel. This second quality of service channel is different from the first quality of service channel. That is, the first service and the second service correspond to different quality of service channels.

[0231] Implementation method B3 (corresponding to Design 3 above): Businesses with the same model identification information and the same model feature information correspond to the same quality of service channel; businesses with different model identification information and / or different model feature information correspond to different quality of service channels.

[0232] The identical model feature information can mean that all model feature information is identical, or that the model feature information used for binding the Quality of Service (QoS) channel is identical. Taking the first and second services mentioned above as examples, the implementation method B3 will be explained in detail.

[0233] When the model feature information corresponding to the first model (the model corresponding to the first service) is the same as the model feature information corresponding to the second model (the model corresponding to the second service), and the identification information of the first model is the same as the identification information of the second model, the session management network element determines that the second service corresponds to the aforementioned first quality of service channel, that is, the first service and the second service correspond to the same quality of service channel.

[0234] When the model feature information corresponding to the first model (the model corresponding to the first service) is different from the model feature information corresponding to the second model (the model corresponding to the second service), and / or the identification information of the first model is different from the identification information of the second model, the session management network element determines that the second service corresponds to the second quality of service channel. This second quality of service channel is different from the first quality of service channel, that is, the first service and the second service correspond to different quality of service channels.

[0235] Implementation method C: Businesses with the same model information and the same service quality parameters correspond to the same service quality channel; businesses with different model information and / or different service quality parameters correspond to different service quality channels.

[0236] Service quality parameters can include, for example, 5QI and ARP, which will not be listed here. Same service quality parameters mean that the values ​​for the same service quality parameter are identical, such as 5QI being 456 for all parameters, or ARP being 123 for all parameters.

[0237] Taking the first and second services as examples, the above implementation method C will be explained in detail. The method 400 shown in Figure 4 further includes: obtaining the first service quality parameter corresponding to the first service and the service quality parameter corresponding to the second service; determining the service quality channel corresponding to the second service based on the second model information, including: when the first model information and the second model information are the same, and the first service quality parameter and the second service quality parameter are the same, determining that the second service corresponds to the first service quality channel. When the first model information and the second model information are different, and / or the first service quality parameter and the second service quality parameter are different, determining that the second service corresponds to the second service quality channel.

[0238] As previously mentioned, the model information used for the Quality of Service (QoS) channel includes Design 1 to Design 3. The following section will detail the implementation of the QoS channel binding method when the model information corresponds to different designs.

[0239] Implementation method C1 (corresponding to Design 1 above): Businesses with the same model identification information and the same service quality parameters correspond to the same service quality channel; businesses with different model identification information and / or different service quality parameters correspond to different service quality channels.

[0240] For a detailed explanation of why the model identification information is the same, please refer to implementation method B1, which will not be repeated here.

[0241] Using the first and second business processes mentioned above as examples, the implementation method C1 will be explained in detail.

[0242] When the identification information of the first model (the model corresponding to the first service) is the same as the identification information of the second model (the model corresponding to the second service), and the first service quality parameter and the second service quality parameter are the same, the session management network element determines that the second service corresponds to the aforementioned first service quality channel, that is, the first service and the second service correspond to the same service quality channel.

[0243] When the identification information of the first model (the model corresponding to the first service) is different from the identification information of the second model (the model corresponding to the second service), or the first service quality parameter is different from the second service quality parameter, or the identification information of the first model is different from the identification information of the second model and the first service quality parameter is different from the second service quality parameter, the session management network element determines that the second service corresponds to the second service quality channel. This second service quality channel is different from the first service quality channel. That is, the first service and the second service correspond to different service quality channels.

[0244] Implementation method C2 (corresponding to Design 2 above): Businesses with the same model feature information and the same service quality parameters correspond to the same service quality channel; businesses with different model feature information and / or different service quality parameters correspond to different service quality channels.

[0245] Using the first and second business processes mentioned above as examples, the implementation method C2 will be explained in detail.

[0246] When the model feature information corresponding to the first model (the model corresponding to the first service) is the same as the model feature information corresponding to the second model (the model corresponding to the second service), and the first service quality parameter and the second service quality parameter are the same, the session management network element determines that the second service corresponds to the aforementioned first service quality channel, that is, the first service and the second service correspond to the same service quality channel.

[0247] When the model feature information corresponding to the first model (the model corresponding to the first service) is different from the model feature information corresponding to the second model (the model corresponding to the second service), or the first service quality parameter is different from the second service quality parameter, or the first service quality parameter is different from the second service quality parameter and the model feature information corresponding to the first model is different from the model feature information corresponding to the second model, the session management network element determines that the second service corresponds to the second service quality channel. This second service quality channel is different from the first service quality channel. That is, the first service and the second service correspond to different service quality channels.

[0248] For a detailed explanation of the identical model feature information, please refer to implementation method B2, which will not be repeated here.

[0249] Implementation method C3 (corresponding to Design 3 above): Businesses with the same model identification information, the same model feature information, and the same service quality parameters correspond to the same service quality channel; businesses with different model identification information, model feature information, or service quality parameters correspond to different service quality channels.

[0250] For an explanation of why the model identification information and the model feature information are the same, please refer to implementation method B3, which will not be repeated here.

[0251] Using the first and second business processes mentioned above as examples, the implementation method C3 will be explained in detail.

[0252] When the model feature information corresponding to the first model (the model corresponding to the first service) is the same as the model feature information corresponding to the second model (the model corresponding to the second service), and the identification information of the first model is the same as the identification information of the second model, and the first service quality parameter and the second service quality parameter are the same, the session management network element determines that the second service corresponds to the aforementioned first service quality channel, that is, the first service and the second service correspond to the same service quality channel.

[0253] When one or more of the following are different, the session management network element determines that the second service corresponds to the second quality of service channel, which is different from the first quality of service channel: the first quality of service parameter and the second quality of service parameter, the identification information of the first model and the identification information of the second model, the model feature information corresponding to the first model and the model feature information corresponding to the second model.

[0254] The above section used the binding of service and quality of service (QoS) channels as an example to explain the binding methods of QoS channels in detail. In another possible implementation, policy information includes services, or, alternatively, policy information includes services and model information. Services and policy information correspond, and one policy information can correspond to one or more service information, but the QoS parameters and model information in the same policy information are identical. The session management network element can bind / correspond to policy information and QoS channels, or in other words, bind policy information to QoS channels, thus binding services and QoS channels. For example, in the third implementation method described above for the session management network element to obtain the first model information, the session management network element can obtain the first policy information, which includes the aforementioned first model information and the aforementioned first service information. Then, the session management network element can determine the first QoS channel corresponding to the first policy information based on the first model information.

[0255] Optionally, the session management network element determines the first quality of service channel corresponding to the first policy information based on the first model information, including the following possible implementation methods:

[0256] Implementation method X: When the first policy information includes the first model information, the session management network element determines that the first policy information corresponds to the first quality of service channel. The first quality of service channel only corresponds to the first policy information, or in other words, the first quality of service channel does not correspond to any other policy information besides the first policy information.

[0257] The session management network element determines that the first policy information corresponds to the first quality of service channel. Alternatively, the session management network element may bind the first policy information to the first quality of service channel, or bind the first policy information to a specific / separate / dedicated quality of service channel, or establish a specific / separate / dedicated quality of service channel for the first policy information, or determine that the first policy information corresponds to a separate / specific / dedicated quality of service channel. Here, a separate / specific / dedicated quality of service channel means that the quality of service channel corresponds to / is bound to only one policy information.

[0258] In summary, as long as the policy information includes model information, the session management network element can bind the policy information to a separate quality of service (QoS) channel. That is, as long as the policy information includes model information, regardless of whether the model information is the same or the QoS parameters (such as 5QI, ARP, etc.) are the same, the policy information will be mapped to a separate QoS channel.

[0259] For example, as shown in Table 3, the policy information includes service information, quality of service (QoS) parameters, and model information. The service information may be, for example, an IP triplet or IP quintuple; the QoS parameters may include, for example, 5QI and ARP; and the model information may include, for example, model identification information and / or model feature information. For instance, policy information 1 is used for forwarding service 1 and includes model information 1; policy information 2 is used for forwarding service 2 and includes model information 2. The session management network element determines that these two policy information entries correspond to separate QoS channels, such as policy information 1 corresponding to QoS channel 1 and policy information 2 corresponding to QoS channel 2. Both QoS channels 1 and QoS channel 2 are separate QoS channels. Model information 1 and model information 2 may be the same or different.

[0260] Table 3

[0261] It should be noted that when the first policy information does not include the first model information, the session management network element can determine the QoS channel corresponding to the first policy information based on the QoS parameters. For specific details, please refer to the existing introduction to QoS flow binding methods above; they will not be repeated here.

[0262] Implementation Method Y: Policy information containing the same model information corresponds to the same Quality of Service (QoS) channel; or, policy information containing the same model information is bound to the same QoS channel. Alternatively, when different policy information contains the same model information, these policy information can correspond to / be bound to the same QoS channel. Conversely, policy information containing different model information corresponds to different QoS channels; or, policy information containing different model information is bound to different QoS channels. Alternatively, the QoS channel of policy information is bound based on the model information.

[0263] For a detailed explanation of implementation method Y, please refer to implementation method B above. When the model information includes model identification information and / or model feature information, the specific implementation of implementation method Y can be found in B1 to B3, except that the correspondence / binding between services and QoS channels is replaced by the correspondence / binding between policy information and QoS channels. Policy information is used to instruct services on how to perform forwarding processing or to instruct the forwarding processing policy of a service. Policy information includes services; one policy information can correspond to one or more service information, but the QoS parameters and model information within the same policy information are identical. The session management network element binds the policy information and the QoS channel, which is equivalent to binding the services and the QoS channel.

[0264] Implementation method Z: Strategy information with the same model information and the same service quality parameters corresponds to the same service quality channel; strategy information with different model information and / or different service quality parameters corresponds to different service quality channels.

[0265] For a detailed explanation of the above implementation method Z, please refer to implementation method C above. When the model information includes model identification information and / or model feature information, the specific implementation of implementation method Z can be found in C1 to C3. The only difference is that the correspondence / binding between the business and the quality of service channel is replaced by the correspondence / binding between the policy information and the quality of service channel. The policy information is used to instruct the business on how to perform forwarding processing or to instruct the forwarding processing policy of the business.

[0266] After the session management network element determines the quality of service channel corresponding to the service, it can distribute model information to enable other devices to perform model inference. In this application, either the terminal or user plane function network element can perform inference based on the model (denoted as Design A), or the access network device can perform inference based on the model (denoted as Design B). The method by which the session management network element distributes model information differs for these two possible designs, which will be explained in detail below.

[0267] For Design A: Inference is performed on the terminal or user plane function network element. The above method 400 further includes: the session management network element sending a first rule corresponding to the first service to the terminal and / or user plane function network element, the first rule being used to control the traffic forwarding of the quality of service channel, the first rule including first model information.

[0268] Traffic forwarding can also be replaced by forwarding processing, customer experience, service quality control or service quality management, etc., and this application does not limit it in this regard.

[0269] The first rule mentioned above can instruct terminal and / or user plane function network elements to enable model-assisted quality of service control / quality of service management / traffic forwarding / forwarding processing / customer experience.

[0270] One possible design for the first rule is that it includes QFI, service flow identifier, and model information, wherein the model information includes model identifier information and / or model feature information.

[0271] One example is that the session management network element performs service and quality of service (QoS) channel binding based on implementation method A, or performs policy information and QoS channel binding based on implementation method X. In this case, taking service 1 and service 2 as examples, the session management network element sends the first rule corresponding to service 1 and the first rule corresponding to service 2 to the terminal and / or user plane function network element, such as QFI 1: SDF 1, model identification information, and model feature information; and QFI 2: SDF 2, model identification information, and model feature information. Here, QFI 1 is used to identify the QoS channel corresponding to service 1, and QFI 2 is used to identify the QoS channel corresponding to service 2. These two services correspond to different QoS channels. SDF 1 is used to identify service 1, and SDF 2 is used to identify service 2. The model feature information in the rules sent by the session management network element to the terminal may include the model direction: uplink, and the model feature information in the rules sent to the user plane function network element may include the model direction: downlink.

[0272] Another example is that the session management network element performs service and quality of service (QoS) channel binding based on implementation method B1. In this case, taking service 1 and service 2 as examples, the session management network element sends the first rule corresponding to service 1 and the first rule corresponding to service 2 to the terminal and / or user plane function network element, such as QFI 1: SDF 1, SDF 2, model identification information, and model feature information. QFI 1 is used to identify the QoS channels corresponding to service 1 and service 2. These two services are bound to the same QoS channel, and these two services correspond to the same model. For an explanation of the model feature information, please refer to the previous example; it will not be repeated here. It should be understood that the parameters included in the above first rule are merely examples and should not constitute any limitation on this application.

[0273] Another example is that the session management network element performs service and quality of service (QoS) channel binding based on implementation method B2. In this case, taking service 1 and service 2 as examples, the session management network element sends the first rule corresponding to service 1 and the first rule corresponding to service 2 to the terminal and / or user plane function network element, such as QFI 1: SDF 1 - Model Identifier Information 1, SDF 2 - Model Identifier Information 2, Model Feature Information. Here, QFI 1 is used to identify the QoS channels corresponding to service 1 and service 2. These two services are bound to the same QoS channel. These two services correspond to different models (e.g., service 1 corresponds to model identifier information 1, and service 2 corresponds to model identifier information 2), but they correspond to the same model feature information. For an explanation of the model feature information, please refer to the previous example, which will not be repeated here. It should be understood that the parameters included in the above first rule are merely examples and should not constitute any limitation on this application.

[0274] In the above example, the session management network element can also indicate the quality of service channel corresponding to the service to the terminal / user plane function network element, such as through QFI.

[0275] For example, the session management network element sends first information to the terminal / user plane function network element, the first information indicating the quality of service channel corresponding to the first service and the model information corresponding to the quality of service channel. Correspondingly, the terminal / user plane function network element receives the first information and performs inference on the data packets in the quality of service channel through the first model. The first model is determined based on the first information and is the model corresponding to the quality of service channel.

[0276] In this way, terminal / user plane function network elements can determine the correspondence between services, quality of service channels, and model information based on the first information received, thereby facilitating inference of data packets in the quality of service channel based on the model corresponding to the quality of service channel.

[0277] It should be noted that a Quality of Service (QoS) channel can correspond to model information; for example, one QoS channel can correspond to one piece of model information (or model). However, this should not constitute any limitation on this application. Model information (or models) can also correspond one-to-one with a business; for example, businesses with the same model feature information correspond to the same QoS channel. In this case, one QoS channel may correspond to multiple pieces of model information (or models), but one business corresponds to one piece of model information. Correspondingly, in the above cases, the first model can refer to the model corresponding to the first business.

[0278] The aforementioned session management network element sends a first rule to the terminal. One possible design is that the session management network element sends a QoS rule corresponding to the first service to the terminal. This QoS rule includes first model information, which may include, for example, model identification information and / or model feature information. The parameters included in the model feature information can be referred to above and will not be detailed here. The first model information included in this QoS rule and the first model information included in the first policy information can be the same. Optionally, the QoS rule may also include QFI and QoS parameters. Correspondingly, the terminal receives the aforementioned QoS rule. Furthermore, the terminal can also enable the corresponding model for the first service based on the aforementioned first model information. This model can be locally stored or obtained from other network elements; this application does not limit this.

[0279] The aforementioned session management network element sends a first rule to the user plane function network element. One possible design is that the session management network element sends a QoS enforcement rule (such as QER) corresponding to the first service to the user plane function network element. This QoS enforcement rule includes first model information, which may include, for example, model identification information and / or model feature information. The parameters included in the model feature information can be referred to above and will not be detailed here. The first model information included in this QoS enforcement rule and the first model information included in the first policy information can be the same. Optionally, the QoS enforcement rule may also include QFI and QoS parameters. Correspondingly, the user plane function network element receives the aforementioned QoS enforcement rule. Furthermore, the user plane function network element can also enable the corresponding model for the first service based on the aforementioned first model information. This model can be locally stored or obtained from other network elements, and this application does not limit this.

[0280] It should be noted that, in this application, the sending of model information from the session management network element to the user plane function network element, so that the user plane function network element can obtain the corresponding model based on the model information, is merely an example and should not constitute any limitation on this application. In this application, the session management network element may also directly send the model to the user plane function network element. Similarly, the sending of model information from the application function network element to the policy control network element, and then the policy control network element instructing the session management network element on the model information, is also merely an example and should not constitute any limitation on this application. For example, the application function network element may also send the model to the policy control network element, and then the policy control network element instructing the session management network element on the model.

[0281] Optionally, for design A, the method 400 further includes: the session management network element sending first indication information to the access network device, the first indication information being used to instruct the first quality of service channel to enable model-assisted quality of service control. In this way, the access network device can determine that model-assisted quality of service control is enabled, prepare to receive inference results, and thus facilitate the execution of quality of service control based on the inference results.

[0282] The aforementioned first indication information is used to instruct the first quality of service channel to enable model-assisted quality of service control. One possible design is that the session management network element instructs the access network device to enable model-assisted quality of service control for the first quality of service channel through one bit. Another possible design is that when the first indication information indicates the first model information, the access network device is instructed by default to enable model-assisted quality of service control for the first quality of service channel.

[0283] Optionally, the aforementioned first indication information is also used to indicate first model information. For example, the aforementioned first indication information may also indicate the function of the model corresponding to the first service, such as one or more of burst prediction, encrypted flow detection, base station resource prediction, or congestion prediction.

[0284] In one possible implementation, the session management network element can indicate the aforementioned first model information to the access network device through a QoS profile. For example, the session management network element sends a QoS profile to the access network device, which includes the first model information and QoS parameters.

[0285] Optionally, for design A, the above method 400 further includes: the terminal / user plane function network element sending the inference result to the access network device.

[0286] After performing inference on the data flow on the Quality of Service (QoS) channel using a model, the terminal / user plane function network element can send the inference results to the access network equipment. The access network equipment can then reserve resources, perform QoS control, or perform QoS management based on the inference results.

[0287] Scenario for user plane function network elements performing inference (taking burst prediction as an example): Downlink data packets arrive at the user plane function network element, such as when a server sends a downlink data packet to the user plane function network element. After receiving the downlink data packet, the user plane function network element determines which service the data packet belongs to and the corresponding quality of service channel. Then, it performs inference on the downlink data packet using the corresponding model (which can be determined based on model information), obtaining the inference result, such as one or more of the following: the size of one or more data packets, arrival time, or end time. The user plane function network element then sends the inference result to the access network device. For example, the user plane function network element can add the inference result to the GTP-U header of the data packet to notify the access network device. The access network device reserves resources / performs quality of service control / performs quality of service management based on the inference result. Further, the access network device forwards the downlink data packet to the terminal.

[0288] In a scenario where the terminal performs inference (taking burst prediction as an example): the terminal's uplink data packets arrive at the mobile terminal (MT) module from the terminal equipment (TE) module. The MT module performs inference on the uplink data packets based on the model (similar to the user plane function network element side). The TE module is used to configure applications and the operating system, while the MT module can be a modem chip. The TE and MT modules can be deployed on the same chip or on different chips; this application does not limit this. After obtaining the inference result, the terminal sends it to the access network device. For example, the terminal can send the inference result to the access network device via a status report (SR) (i.e., the inference result is carried in the SR), or via a buffer status report (BSR) (i.e., the inference result is carried in the BSR), or via other messages; this application does not limit this either. Based on the inference result, the access network device reserves resources and receives the aforementioned data packets through these reserved resources, then forwards them to the user plane function network element.

[0289] For Design B: The access network device performs inference. The method 400 further includes: the session management network element sending a second rule corresponding to the first quality of service channel to the access network device, the second rule including the first model information. Correspondingly, the access network device receives the second rule.

[0290] The second rule mentioned above can instruct access network devices to enable model-assisted quality of service control / quality of service management / traffic forwarding / forwarding processing / customer experience. For example, the second rule could be a QoS profile, which includes the first model information and QoS parameters.

[0291] After receiving the second rule, the access network device can activate the corresponding model for the first service based on the first model information. This model can be stored locally or obtained from other network elements; this application does not limit this. After obtaining the inference result, the access network device can perform quality of service control based on the inference result.

[0292] It should be noted that the second rule mentioned above includes model information, but this is merely an example and should not constitute any limitation on this application. For example, the session management network element can send a rule corresponding to a Quality of Service (QoS) channel X (an example of a third QoS channel) to the access network device. This rule includes one or more model information pieces, which are used for forwarding processing of QoS channel X. Correspondingly, the access network device receives the rule corresponding to QoS channel X. Furthermore, the access network device can execute / control the forwarding processing of QoS channel X based on the one or more model information pieces mentioned above.

[0293] The scenario in which the above rules include one model information could be, for example, a scenario where the session management network element performs service quality channel binding based on implementation method A, implementation method B1, or implementation method B3. The scenario in which the above rules include multiple model information could be, for example, a scenario where the session management network element performs service quality channel binding based on implementation method B2.

[0294] The aforementioned access network device, based on one or more of the aforementioned model information, performs / controls the forwarding process of the aforementioned Quality of Service (QoS) channel X, which may include: performing the forwarding process of the aforementioned QoS channel X based on the inference results corresponding to the one or more of the aforementioned model information.

[0295] The inference results corresponding to the above one or more model information can be understood as the inference results of one or more models corresponding to the above one or more model information.

[0296] In one possible implementation of the method 400 shown in Figure 4, before performing step 401, the method 400 further includes: the session management network element receiving a request message from the terminal, the request message being used to request model-assisted quality of service control.

[0297] This request message can be a service quality control mechanism that assists in the request model of a specific session / service / terminal. For example, the aforementioned request message could be a session establishment request.

[0298] One possible design is that the terminal can directly request model-assisted quality of service control from the session management element. For example, the terminal can request model-assisted quality of service control from the session management element using only one bit.

[0299] Another possible design is that the terminal can request model-assisted quality of service (QoS) control from the session management network element via session attributes. For example, one or more session attributes correspond to a request for model-assisted QoS control. The terminal can request this session attribute, such as data network information or slice information, in the session establishment request. The data network information identifies the specific data network the terminal wishes to connect to, and the slice information is used to select a specific network slice. Data network information can be, for example, a data network name (DNN), and slice information can be, for example, single network slice selection assistance information (S-NSSAI). The session management network element can determine the model-assisted QoS control request for that session / service / terminal based on the aforementioned session attributes.

[0300] Optionally, the terminal can also indicate desired / expected model functions to the session management network element, such as for burst prediction, encrypted stream detection, etc.

[0301] Optionally, the terminal can also indicate to the session management network element the services that need to use the model.

[0302] In one possible implementation of the method 400 shown in Figure 4, prior to step 401, the method 400 further includes: receiving second instruction information from a unified data management network element, the second instruction information being used to instruct authorized model-assisted quality of service control.

[0303] The second instruction information may be an instruction from the unified data management network element to assist in quality of service control for a specific session / service / terminal authorization model.

[0304] For example, after receiving the aforementioned request message from the terminal, the session management network element can query the UDM network element for the subscription information corresponding to the terminal. This subscription information indicates whether the terminal is authorized / allowed for model-assisted quality of service control. Correspondingly, after receiving the query request, the UDM network element can instruct the session management network element to authorize model-assisted quality of service control.

[0305] To better understand the communication method shown in Figure 4, the following will provide a detailed flowchart of possible implementations of the method shown in Figure 4, in conjunction with Figures 5 to 8. The differences between the embodiments shown in Figures 5 to 7 are as follows: In the embodiment shown in Figure 5, the session management network element performs the binding of policy information and quality of service channels based on whether the policy information includes model information; in the embodiment shown in Figure 6, the session management network element performs the binding of policy information and quality of service channels based on model identifier information; and in the embodiment shown in Figure 7, the session management network element performs the binding of policy information and quality of service channels based on model feature information. The difference between Figure 8 and the embodiments shown in Figures 5 to 7 is that in the embodiment shown in Figure 8, the access network device performs model inference, while in the embodiments shown in Figures 5 to 7, the UPF network element and the terminal perform model inference. The embodiment shown in Figure 8 uses the example of binding policy information and quality of service channels based on whether the policy information includes model information; other implementation methods can be found in Figures 6 and 7.

[0306] The following description details the communication method provided in this application using the interaction between session management network elements, policy control network elements, application function network elements, unified data management network elements, network data analysis function network elements, user plane function network elements, terminals, and access network devices as examples, but this should not be construed as limiting this application in any way. For example, the session management network element can also be replaced by components configured in the session management network element (such as chips, chip systems, processors, etc.), or logical modules or software capable of implementing all or part of the functions of the session management network element, etc. Other network elements are similar and will not be described in detail here. Exemplarily, the method described below can be applied to the network architecture shown in Figure 1.

[0307] Figure 5 is a detailed flowchart of the communication method provided in the embodiments of this application.

[0308] In step 501, the terminal sends indication information 1 to the session management network element, which requests model-assisted quality of service control. Correspondingly, the session management network element receives the aforementioned indication information 1.

[0309] The term "request" can also be replaced with "demand," "expectation," "hope," or "support." The aforementioned instruction information 1 can be forwarded through one or more intermediate network elements. For example, the terminal sends instruction information 1 to the access network device, the access network device receives instruction information 1, and forwards it to the session management network element.

[0310] Optionally, the aforementioned indication information 1 may be a service quality control mechanism assisted by a specific session / service / terminal request model.

[0311] Optionally, the aforementioned indication information 1 can be carried in a session establishment request, which is used to request the establishment of a session. The aforementioned request message can be, for example, a session establishment request.

[0312] For a detailed description of the possible design of the request message for service quality control assisted by the terminal request model, please refer to Figure 4, which will not be repeated here.

[0313] Optionally, the terminal can also indicate desired / expected model functions to the session management network element, such as for burst prediction, encrypted stream detection, etc.

[0314] Optionally, the terminal can also indicate to the session management network element the services that need to use the model.

[0315] In step 502, the session management network element queries the subscription information from the unified data management network element.

[0316] For example, the session management network element can send request message 1 to the unified data management network element (such as the UDM network element), which requests to query the subscription information of the aforementioned terminal. Correspondingly, the unified data management network element receives the aforementioned request message 1. Furthermore, the unified data management network element sends response message 1 to the session management network element, which can indicate whether the aforementioned terminal / session / service is subject to authorized / allowed model-assisted quality of service control. In one scenario, response message 1 indicates that the aforementioned terminal / session / service is subject to authorized / allowed model-assisted quality of service control. In another scenario, response message 1 indicates that the aforementioned terminal / session / service is not subject to authorized / allowed model-assisted quality of service control. It is understood that the steps described below can be performed, for example, when the aforementioned terminal / session / service is subject to authorized / allowed model-assisted quality of service control.

[0317] It should be noted that step 502 is optional. When the session management network element does not execute step 502, the service quality control assisted by the above-mentioned terminal / session / service authorization / allowance model can be used by default.

[0318] In step 503, the session management network element sends instruction information 2 to the policy control network element, which instructs the use of model-assisted quality of service control. Correspondingly, the policy control network element receives the instruction information 2.

[0319] In this context, instruction information 2 may indicate the activation / use of model-assisted Quality of Service (QoS) for a specific session / service / terminal. The granularity of instruction information 2 can be the same as that of response message 1. For example, if response message 1 indicates that model-assisted QoS is authorized for a specific service, then instruction information 2 indicates that model-assisted QoS is enabled for that service. The granularity of instruction information 2 and instruction information 1 can also be the same or different. For instance, if instruction information 1 indicates a request for model-assisted QoS for a specific terminal, but response message 1 indicates that model-assisted QoS is authorized / allowed for a specific service of that terminal (meaning not all services of that terminal are authorized for model-assisted QoS), then instruction information 2 indicates that model-assisted QoS is used for that specific service of the terminal.

[0320] After receiving the aforementioned instruction information 2, the policy control network element can obtain model information, which is used for service forwarding processing. Forwarding processing can be replaced by traffic forwarding, customer experience, service quality management, or service quality control, etc. The policy control network element can obtain model information in either step 504 or step 505; this application does not limit this. For example, the policy control network element can also obtain model information from local storage.

[0321] It should be noted that steps 501 and 503 are also optional. For example, the policy control network element can determine the use of model-assisted quality of service control based on the attributes of sessions such as DNN / S-NSSAI.

[0322] In step 504, the application function network element sends indication information 3 to the policy control network element, which indicates the model information corresponding to the service. Correspondingly, the policy control network element receives the aforementioned indication information 3.

[0323] The aforementioned instruction information 3 can also instruct a certain service / session / terminal to enable model-assisted quality of service control.

[0324] For example, the application function network element sends a service identifier, a terminal identifier, and model information to the policy control network element. The service identifier identifies the service corresponding to the model information, the terminal identifier identifies the terminal, and the model information includes, but is not limited to, one or more of the following: model identifier information, model feature information, or the model itself. A description of the model feature information can be found in the explanation in Figure 4, and will not be repeated here.

[0325] The aforementioned instruction information 3 may be sent directly from the application function network element to the policy control network element, or it may be forwarded through the NEF network element. This application does not limit this.

[0326] In step 505, the network data analysis function network element indicates the model information corresponding to the service to the policy control network element. Correspondingly, the policy control network element receives the aforementioned model information corresponding to the service.

[0327] An example of the process by which a policy control network element obtains model information is that a terminal requests model-assisted quality of service control from a session management network element. The session management network element determines, by querying the subscription information, that a certain service of the terminal is authorized to be subject to model-assisted quality of service control, and then instructs the policy control network element that the service is authorized to be subject to model-assisted quality of service control. The policy control network element then obtains the model information corresponding to the service from the network data analysis function network element.

[0328] It should be noted that steps 504 and 505 above are optional, and the policy control network element can also obtain model information through other means.

[0329] In step 506, the policy control network element determines to enable model-assisted quality of service control for the aforementioned services.

[0330] For example, the policy control network element determines to enable model-assisted quality of service control for the above services based on the instructions of the application function network element / session management network element.

[0331] It should be noted that step 506 above is optional.

[0332] In step 507, the policy control network element generates policy information, which includes model information and service information.

[0333] The aforementioned strategy information could be, for example, PCC rules. An explanation of the model information and business information can be found in Figure 4, and will not be repeated here.

[0334] For example, the policy control network element generates policy information 1, which includes model information 1 and an identifier for service 1, wherein model information 1 is the model information corresponding to service 1. It can be understood that the policy control network element can generate corresponding policy information for different services. For example, for service 2, the policy control network element can also generate policy information 2, which includes model information 2 and an identifier for service 2, wherein model information 2 is the model information corresponding to service 2.

[0335] In step 508, the policy control network element sends policy information to the session management network element, and correspondingly, the session management network element receives the policy information. In Figure 5, taking the policy control network element sending policy information corresponding to service 1 and policy information corresponding to service 2 to the session management network element as an example, step 508 includes steps 508a and 508b. It should be understood that the policy control network element may also send more or fewer pieces of policy information to the session management network element, and this application does not limit this.

[0336] In step 508a, the policy control network element sends policy information 1 to the session management network element. The policy information 1 includes model information 1 and the identifier of service 1.

[0337] In step 508b, the policy control network element sends policy information 2 to the session management network element. The policy information 2 includes model information 2 and the identifier of service 2.

[0338] Policy information 1 and policy information 2 can be carried in the same signaling or in different signaling; this application does not limit this.

[0339] In step 509, the session management network element performs the binding of policy information and quality of service channels based on the model information.

[0340] For example, when the policy information includes model information, the session management network element determines that the policy information corresponds to a specific quality of service channel. This quality of service channel only corresponds to this policy information, or in other words, this quality of service channel does not correspond to any other policy information besides this policy information.

[0341] Regarding the implementation method of binding the session management network element with the execution policy information and the quality of service channel based on the model information, please refer to the implementation method X described above, which will not be repeated here.

[0342] In step 510, the session management network element sends QoS enforcement rules to the user plane function network element. Correspondingly, the user plane function network element receives the QoS enforcement rules. The QoS enforcement rules include model information corresponding to the service.

[0343] For example, step 510 includes steps 510a and 510b.

[0344] In step 510a, the session management network element sends QoS enforcement rule 1 to the user plane function network element. Correspondingly, the user plane function network element receives QoS enforcement rule 1. QoS enforcement rule 1 includes model information 1 corresponding to service 1.

[0345] For example, QoS enforcement rule 1 includes: QFI 1: service 1 (such as SDF 1) and model information 1.

[0346] In step 510b, the session management network element sends QoS enforcement rule 2 to the user plane function network element. Correspondingly, the user plane function network element receives QoS enforcement rule 2. QoS enforcement rule 2 includes model information 2 corresponding to service 2.

[0347] For example, QoS enforcement rule 2 includes: QFI 2: service 2 (such as SDF 2) and model information 2.

[0348] It should be understood that the aforementioned QoS enforcement rule 1 and QoS enforcement rule 2 can be carried in the same signaling or in different signaling, and this application does not limit this.

[0349] It should also be understood that the above QoS enforcement rules 1 and QoS enforcement rules 2 take the binding of service 1 and service 2 to different quality of service channels as an example, but this should not constitute any limitation on this application. The implementation of the binding of quality of service channels is different, and the rules sent by the session management network element to the user plane function network element may be different. For a detailed explanation, please refer to the relevant explanation in Figure 4, which will not be repeated here.

[0350] In step 511, the user plane function network element obtains the model corresponding to the model information based on the model information.

[0351] For example, a user plane function network element can obtain the corresponding model based on the model identification information, such as obtaining the corresponding model from other network elements based on the model identification information, or obtaining the corresponding model from local storage based on the model identification information.

[0352] Another example is that user plane function network elements can obtain the corresponding model based on the model feature information, such as obtaining the corresponding model from other network elements / local storage based on the model feature information.

[0353] It should be noted that in this application, the sending of model information from the session management network element to the user plane function network element, and the user plane function network element obtaining the corresponding model based on the model information, is merely an example and should not constitute any limitation on this application. In this application, the session management network element may also directly send the model to the user plane function network element. Similarly, the sending of model information from the application function network element to the policy control network element, and then the policy control network element instructing the session management network element on the model information, is also merely an example and should not constitute any limitation on this application. For example, the application function network element may also send the model to the policy control network element, and then the policy control network element instructing the session management network element on the model. The embodiments shown below are similar and will not be described again below.

[0354] In step 512, the session management network element sends indication information 4 to the access network device, which instructs the quality of service channel to perform model-assisted quality of service control. Correspondingly, the access network device receives the aforementioned indication information 4.

[0355] For example, the session management network element instructs the access network device to perform model-assisted quality of service control for quality of service channel 1.

[0356] It should be noted that this application does not limit the execution time of step 512. For example, in Figure 5, step 512 is executed before step 513. However, this should not constitute any limitation on this application, as long as step 512 is executed before the access network device performs model-assisted quality of service control, such as step 512 being executed before step 518 or step 521.

[0357] In step 513, the session management network element sends QoS rules to the terminal. Correspondingly, the terminal receives the QoS rules. The QoS rules include model information corresponding to the service.

[0358] For example, step 513 includes steps 513a and 513b.

[0359] In step 513a, the session management network element sends QoS rule 1 to the terminal. Correspondingly, the terminal receives QoS rule 1. QoS rule 1 includes model information 1 corresponding to service 1.

[0360] In step 513b, the session management network element sends QoS rule 2 to the terminal. Correspondingly, the terminal receives QoS rule 2. QoS rule 2 includes model information 2 corresponding to service 2.

[0361] It should be understood that the aforementioned QoS rule 1 and QoS rule 2 can be carried in the same signaling or in different signaling, and this application does not limit this.

[0362] It should also be understood that the above QoS rules 1 and QoS rules 2 take the binding of service 1 and service 2 to different quality of service channels as an example, but this should not constitute any limitation on this application. The implementation of binding quality of service channels is different, and the rules sent by the session management network element to the terminal may be different. For a detailed explanation, please refer to the relevant explanation in Figure 4, which will not be repeated here.

[0363] In step 514, the terminal obtains the model corresponding to the model information based on the model information. The method by which the terminal obtains the model corresponding to the model information is similar to that of the user plane function network element; for a detailed explanation, please refer to step 511, which will not be elaborated here.

[0364] It is understandable that user plane functional network elements can infer downlink service flows based on the model, and terminals can infer uplink service flows based on the model. The inference process performed by user plane functional network elements and terminals can be seen in the explanation in Figure 4, which will not be repeated here.

[0365] After obtaining the model, the terminal and user plane function network elements can perform inference based on the model, obtain and report the inference results to the access network equipment, so that the access network equipment can reserve resources / perform quality of service control / perform quality of service management based on the inference results. Steps 515 to 518 below are the downlink data packet processing process, and steps 519 to 521 are the uplink data packet processing process.

[0366] In step 515, the server sends data packets to the user plane function network element. Correspondingly, the user plane function network element receives the data packets.

[0367] In step 516, the user plane function network element performs inference based on the model to obtain the inference result.

[0368] The inference result described above could be one or more of the following: the size of the next / multiple data packets, their arrival time, or their end time.

[0369] In step 517, the user plane function network element sends the inference result to the access network device. Correspondingly, the access network device receives the inference result.

[0370] For example, user plane function network elements can add inference results to the GTP-U header of data packets to notify access network devices.

[0371] In step 518, the access network device reserves resources / performs quality of service control / performs quality of service management based on the inference results.

[0372] For example, upon receiving indication information 4 (which instructs the QoS channel to perform model-assisted QoS control), the access network device can reserve resources / perform QoS control / perform QoS management based on the inference result. It is understood that the above example should not constitute any limitation on this application. For instance, the QoS channel can also default to performing model-assisted QoS control, in which case the access network device, upon receiving the inference result, can reserve resources / perform QoS control / perform QoS management based on the inference result.

[0373] Furthermore, access network devices can also forward data packets to terminals.

[0374] In step 519, the terminal performs inference based on the model and obtains the inference result.

[0375] The process of terminal inference can be referred to the relevant description in Figure 4, which will not be elaborated here.

[0376] In step 520, the terminal sends the inference result to the access network device. Correspondingly, the access network device receives the inference result.

[0377] For a detailed explanation of how the terminal sends inference results to the access network equipment, please refer to the relevant description in Figure 4, which will not be elaborated here.

[0378] In step 521, the access network device reserves resources / performs quality of service control / performs quality of service management based on the inference results. A detailed explanation of step 521 can be found in step 518, and will not be repeated here.

[0379] Furthermore, access network devices can also forward data packets to user plane function network elements.

[0380] Figure 6 is another detailed flowchart of the communication method provided in the embodiments of this application.

[0381] In step 601, the terminal sends indication information 1 to the session management network element, which requests model-assisted quality of service control. Correspondingly, the session management network element receives the aforementioned indication information 1.

[0382] In step 602, the session management network element queries the subscription information from the unified data management network element.

[0383] In step 603, the session management network element sends instruction information 2 to the policy control network element, which instructs the use of model-assisted quality of service control. Correspondingly, the policy control network element receives the instruction information 2.

[0384] In step 604, the application function network element sends indication information 3 to the policy control network element, which indicates the model information corresponding to the service. Correspondingly, the policy control network element receives the aforementioned indication information 3.

[0385] In step 605, the network data analysis function network element indicates the model information corresponding to the service to the policy control network element. Correspondingly, the policy control network element receives the aforementioned model information corresponding to the service.

[0386] In step 606, the policy control network element determines to enable model-assisted quality of service control for the aforementioned services.

[0387] In step 607, the policy control network element generates policy information, which includes model information and service information.

[0388] In step 608, the policy control network element sends policy information to the session management network element, and correspondingly, the session management network element receives the policy information. In Figure 6, taking the policy control network element sending policy information corresponding to service 1 and policy information corresponding to service 2 to the session management network element as an example, step 608 includes steps 608a and 608b. It should be understood that the policy control network element may also send more or fewer pieces of policy information to the session management network element, and this application does not limit this.

[0389] In step 608a, the policy control network element sends policy information 1 to the session management network element. The policy information 1 includes model information 1 and the identifier of service 1.

[0390] In step 608b, the policy control network element sends policy information 2 to the session management network element. The policy information 2 includes model information 2 and the identifier of service 2.

[0391] For a detailed explanation of steps 601 to 608, please refer to steps 501 to 508 in Figure 5, which will not be repeated here.

[0392] In step 609, the session management network element performs the binding of policy information and quality of service channels based on the model identification information.

[0393] For example, policy information including the same model identifier corresponds to the same quality of service (QoS) channel; policy information including different model identifiers corresponds to different QoS channels. Alternatively, policy information corresponding to the same model corresponds to the same QoS channel; policy information corresponding to different models corresponds to different QoS channels.

[0394] Regarding the implementation method of binding policy information and quality of service channels based on model identification information, please refer to the implementation method B1 described above. The only difference is that the binding of service and quality of service channels is replaced with the binding of policy information and quality of service channels.

[0395] Optionally, the session management network element can bind policy information and quality of service channels based on model identification information and model feature information. For a detailed explanation, please refer to implementation method B3 shown above; simply replace the binding of service and quality of service channels with the binding of policy information and quality of service channels.

[0396] Optionally, the session management network element can perform the binding of policy information and quality of service channels based on model identification information and quality of service parameters. For a detailed explanation, please refer to implementation method C1 shown above, except that the binding of services and quality of service channels is replaced with the binding of policy information and quality of service channels.

[0397] Optionally, the session management network element can bind policy information and quality of service channels based on model identification information, model feature information, and quality of service parameters. For a detailed explanation, please refer to implementation method C3 shown above; simply replace the binding of services and quality of service channels with the binding of policy information and quality of service channels.

[0398] In step 610, the session management network element sends QoS enforcement rules to the user plane function network element. Correspondingly, the user plane function network element receives the QoS enforcement rules. The QoS enforcement rules include model information corresponding to the service.

[0399] For example, the session management network element sends QoS enforcement rule 1 to the user plane function network element. Correspondingly, the user plane function network element receives QoS enforcement rule 1. QoS enforcement rule 1 includes: QFI 1: service 1 (e.g., SDF 1), service 2 (e.g., SDF 2), model identification information, and model feature information.

[0400] It should be understood that the above example uses the example of Business 1 and Business 2 being bound to the same Quality of Service channel, but this should not constitute any limitation on this application.

[0401] In step 611, the user plane function network element obtains the model corresponding to the model information based on the model information.

[0402] In step 612, the session management network element sends indication information 4 to the access network device, which instructs the quality of service channel to perform model-assisted quality of service control. Correspondingly, the access network device receives the aforementioned indication information 4.

[0403] In step 613, the session management network element sends QoS rules to the terminal. Correspondingly, the terminal receives the QoS rules. The QoS rules include model information corresponding to the service.

[0404] For example, the session management network element sends QoS rule 1 to the terminal. Correspondingly, the terminal receives QoS rule 1. For example, QoS rule 1 includes service 1 (such as SDF 1), service 2 (such as SDF 2), model identification information, and model feature information.

[0405] In step 614, the terminal obtains the model corresponding to the model information based on the model information.

[0406] After obtaining the model, the terminal and user plane function network elements can perform inference based on the model, obtain and report the inference results to the access network equipment, so that the access network equipment can reserve resources / perform quality of service control / perform quality of service management based on the inference results. Steps 615 to 618 below are the downlink data packet processing process, and steps 619 to 621 are the uplink data packet processing process.

[0407] In step 615, the server sends data packets to the user plane function network element. Correspondingly, the user plane function network element receives the data packets.

[0408] In step 616, the user plane function network element performs inference based on the model to obtain the inference result.

[0409] In step 617, the user plane function network element sends the inference result to the access network device. Correspondingly, the access network device receives the inference result.

[0410] In step 618, the access network device reserves resources / performs quality of service control / performs quality of service management based on the inference results.

[0411] In step 619, the terminal performs inference based on the model and obtains the inference result.

[0412] In step 620, the terminal sends the inference result to the access network device. Correspondingly, the access network device receives the inference result.

[0413] In step 621, the access network device reserves resources / performs quality of service control / performs quality of service management based on the inference results. A detailed explanation of steps 610 to 621 can be found in steps 510 to 521 of Figure 5, and will not be repeated here.

[0414] Figure 7 is another detailed flowchart of the communication method provided in the embodiments of this application.

[0415] In step 701, the terminal sends indication information 1 to the session management network element, which requests model-assisted quality of service control. Correspondingly, the session management network element receives the aforementioned indication information 1.

[0416] In step 702, the session management network element queries the subscription information from the unified data management network element.

[0417] In step 703, the session management network element sends instruction information 2 to the policy control network element, which instructs the use of model-assisted quality of service control. Correspondingly, the policy control network element receives the instruction information 2.

[0418] In step 704, the application function network element sends indication information 3 to the policy control network element, which indicates the model information corresponding to the service. Correspondingly, the policy control network element receives the aforementioned indication information 3.

[0419] In step 705, the network data analysis function network element indicates the model information corresponding to the service to the policy control network element. Correspondingly, the policy control network element receives the aforementioned model information corresponding to the service.

[0420] In step 706, the policy control network element determines to enable model-assisted quality of service control for the aforementioned services.

[0421] In step 707, the policy control network element generates policy information, which includes model information and service information.

[0422] In step 708, the policy control network element sends policy information to the session management network element, and correspondingly, the session management network element receives the policy information. In Figure 7, taking the policy control network element sending policy information corresponding to service 1 and policy information corresponding to service 2 to the session management network element as an example, step 708 includes steps 708a and 708b. It should be understood that the policy control network element may also send more or fewer pieces of policy information to the session management network element, and this application does not limit this.

[0423] In step 708a, the policy control network element sends policy information 1 to the session management network element. The policy information 1 includes model information 1 and the identifier of service 1.

[0424] In step 708b, the policy control network element sends policy information 2 to the session management network element. The policy information 2 includes model information 2 and the identifier of service 2.

[0425] For a detailed explanation of steps 701 to 708, please refer to steps 501 to 508 in Figure 5, which will not be repeated here.

[0426] In step 709, the session management network element performs the binding of policy information and quality of service channels based on model feature information.

[0427] For example, policy information including the same model feature information corresponds to the same quality of service channel; services including different model feature information correspond to different quality of service channels.

[0428] Regarding the implementation method of binding policy information and quality of service channels based on model feature information, please refer to the implementation method B2 described above. The only difference is that the binding of service and quality of service channels is replaced with the binding of policy information and quality of service channels.

[0429] Optionally, the session management network element can bind policy information and quality of service channels based on model feature information and quality of service parameters. For a detailed explanation, please refer to implementation method C2 shown above; simply replace the binding of services and quality of service channels with the binding of policy information and quality of service channels.

[0430] In step 710, the session management network element sends QoS enforcement rules to the user plane function network element. Correspondingly, the user plane function network element receives the QoS enforcement rules. The QoS enforcement rules include model information corresponding to the service.

[0431] For example, the session management network element sends QoS enforcement rule 1 to the user plane function network element. Correspondingly, the user plane function network element receives QoS enforcement rule 1. QoS enforcement rule 1 includes: QFI 1: service 1 (e.g., SDF 1) corresponds to model 1, service 2 (e.g., SDF 2) corresponds to model 2, and model feature information. In this example, the model feature information included in policy information 1 and policy information 2 are the same, but this should not constitute any limitation to this application. For example, the model feature information included in policy information 1 and policy information 2 can also be different; in this case, policy information 1 and policy information 2 are bound to different quality of service channels.

[0432] In step 711, the user plane function network element obtains the model corresponding to the model information based on the model information.

[0433] In step 712, the session management network element sends indication information 4 to the access network device, which instructs the quality of service channel to perform model-assisted quality of service control. Correspondingly, the access network device receives the aforementioned indication information 4.

[0434] In step 713, the session management network element sends QoS rules to the terminal. Correspondingly, the terminal receives the QoS rules. The QoS rules include model information corresponding to the service.

[0435] For example, the session management network element sends QoS rule 1 to the terminal. Correspondingly, the terminal receives QoS rule 1. Rule 1 includes service corresponding to model 1, service 2 corresponding to model 2, and model feature information. In this example, it is taken that the model feature information included in policy information 1 and policy information 2 are the same (both are bound to the same quality of service channel), but this should not constitute any limitation to this application. For example, the model feature information included in policy information 1 and policy information 2 may also be different. In this case, policy information 1 and policy information 2 are bound to different quality of service channels.

[0436] In step 714, the terminal obtains the model corresponding to the model information based on the model information.

[0437] After obtaining the model, the terminal and user plane function network elements can perform inference based on the model, obtain and report the inference results to the access network equipment, so that the access network equipment can reserve resources / perform quality of service control / perform quality of service management based on the inference results. Steps 715 to 718 below are the downlink data packet processing process, and steps 719 to 721 are the uplink data packet processing process.

[0438] In step 715, the server sends data packets to the user plane function network element. Correspondingly, the user plane function network element receives the data packets.

[0439] In step 716, the user plane functional network element performs inference based on the model to obtain the inference result.

[0440] In step 717, the user plane function network element sends the inference result to the access network device. Correspondingly, the access network device receives the inference result.

[0441] In step 718, the access network device reserves resources / performs quality of service control / performs quality of service management based on the inference results.

[0442] In step 719, the terminal performs inference based on the model and obtains the inference result.

[0443] In step 720, the terminal sends the inference result to the access network device. Correspondingly, the access network device receives the inference result.

[0444] In step 721, the access network device reserves resources / performs quality of service control / performs quality of service management based on the inference results.

[0445] For a detailed explanation of steps 710 to 721, please refer to steps 510 to 521 in Figure 5, which will not be repeated here.

[0446] Figure 8 is a further detailed flowchart of the communication method provided in the embodiments of this application.

[0447] In step 801, the terminal sends indication information 1 to the session management network element, which requests model-assisted quality of service control. Correspondingly, the session management network element receives the aforementioned indication information 1.

[0448] In step 802, the session management network element queries the subscription information from the unified data management network element.

[0449] In step 803, the session management network element sends instruction information 2 to the policy control network element, which instructs the use of model-assisted quality of service control. Correspondingly, the policy control network element receives the instruction information 2.

[0450] In step 804, the application function network element sends indication information 3 to the policy control network element, which indicates the model information corresponding to the service. Correspondingly, the policy control network element receives the aforementioned indication information 3.

[0451] In step 805, the network data analysis function network element indicates the model information corresponding to the service to the policy control network element. Correspondingly, the policy control network element receives the aforementioned model information corresponding to the service.

[0452] In step 806, the policy control network element determines to enable model-assisted quality of service control for the aforementioned services.

[0453] In step 807, the policy control network element generates policy information, which includes model information and service information.

[0454] In step 808, the policy control network element sends policy information to the session management network element, and correspondingly, the session management network element receives the policy information. In Figure 8, taking the policy control network element sending policy information corresponding to service 1 and policy information corresponding to service 2 to the session management network element as an example, step 808 includes steps 808a and 808b. It should be understood that the policy control network element may also send more or fewer pieces of policy information to the session management network element, and this application does not limit this.

[0455] In step 808a, the policy control network element sends policy information 1 to the session management network element. The policy information 1 includes model information 1 and the identifier of service 1.

[0456] In step 808b, the policy control network element sends policy information 2 to the session management network element. The policy information 2 includes model information 2 and the identifier of service 2.

[0457] For a detailed explanation of steps 801 to 808, please refer to steps 501 to 508 in Figure 5, which will not be repeated here.

[0458] In step 809, the session management network element performs the binding of policy information and quality of service channels based on the model information.

[0459] For a detailed explanation of step 809, please refer to step 509 in Figure 5, which will not be elaborated here.

[0460] In step 810, the session management network element sends a QoS configuration file to the access network device. Correspondingly, the access network device receives the QoS configuration file.

[0461] For example, the QoS configuration file includes QFI, the model identifier information corresponding to QFI, and model feature information.

[0462] In step 811, the access network device obtains the model corresponding to the model information based on the model information.

[0463] The implementation method of the access network device obtaining the corresponding model based on the model information can be seen in step 511 of Figure 5, which will not be repeated here.

[0464] It is understood that the QoS configuration file in step 810 above is merely an example and should not constitute any limitation on this application. For example, a more general rule is to send a rule corresponding to the Quality of Service (QoS) channel to the access network device, the rule including one or more model information, the one or more model information being used for the forwarding processing of the aforementioned QoS channel; after receiving the rule, the access network device performs the forwarding processing of the aforementioned QoS channel according to the one or more model information.

[0465] Forwarding can be replaced by traffic forwarding, customer experience, service quality control, or service quality management.

[0466] One possible design is that the above rules include one model information, meaning that the above quality of service channel corresponds to one model information. In this case, the binding method between the quality of service channel and the business can be, for example, implementation method A, implementation method B1, or implementation method B3. Another possible design is that the above rules include multiple model information, meaning that the above quality of service channel corresponds to multiple model information (or multiple models). In this case, the binding method between the quality of service channel and the business can be, for example, implementation method B2.

[0467] The access network device performs forwarding processing of the above-mentioned quality of service channel based on one or more of the above-mentioned model information. For example, the access network device determines one or more corresponding models based on the above-mentioned model information, and performs inference on the data flow on the above-mentioned quality of service channel through the one or more models. Based on the inference results, the access network device performs traffic forwarding / forwarding processing / quality of service control / management on the above-mentioned quality of service channel, as described in step 813 below.

[0468] In step 812, the access network device performs inference based on the model and obtains the inference result.

[0469] For downlink data packets, the server sends the data packets to the user plane function network element, which then forwards the data packets to the access network device. After receiving the data packets, the access network device performs inference based on the model to obtain the inference result.

[0470] For uplink data packets, the terminal sends data packets to the access network device. After receiving the data packets, the access network device performs inference based on the model and obtains the inference result.

[0471] It is understandable that when a Quality of Service (QoS) channel corresponds to multiple models, the access network device uses multiple models to infer all data packets in the entire QoS channel and comprehensively considers the inference results. When a QoS channel corresponds to one model, the access network device uses that model to infer all data packets in the QoS channel and obtains the inference result.

[0472] In step 813, the access network device reserves resources / performs quality of service control / performs quality of service management based on the inference results.

[0473] It should be noted that the steps in the embodiments shown in Figures 5 to 8 are merely examples and should not constitute any limitation on this application. For example, in practical applications, more or fewer steps may be included, and the order of the steps may also be different. For example, the order of steps 510, 512, and 513 is not limited, and the order of steps 512 and 514 is not limited.

[0474] The methods provided in the embodiments of this application have been described in detail above with reference to the accompanying drawings. The apparatus provided in the embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0475] It should be understood that the devices shown in Figures 9 and 10 can be used to implement the functions of the session management network element, policy control network element, access network equipment, user plane function network element or terminal in the above method embodiments, and thus can also achieve the beneficial effects of the above method embodiments.

[0476] Figure 9 is a schematic block diagram of a communication device 900 provided in an embodiment of this application.

[0477] The communication device 900 can be used to implement the method described in any of the method embodiments shown in Figures 4 to 8 above. Specifically, it can be used to implement the functions of the session management network element, policy control network element, access network equipment, user plane function network element, or terminal in the above method embodiments.

[0478] The modules included in the communication device 900 can be implemented through software and / or hardware.

[0479] For example, the communication device 900 may include modules or units that correspond one-to-one with the methods / operations / steps / actions described in any of the method embodiments shown in Figures 4 to 8. The modules or units may be hardware circuits, software, or a combination of hardware circuits and software.

[0480] For example, when the communication device 900 is used to implement the function of the session management network element in the method embodiment shown in FIG4, the communication device 900 may include an acquisition module 910 and a processing module 920, wherein the acquisition module 910 is used to acquire first model information, which is used for forwarding processing of the first service; the processing module 920 is used to determine the first quality of service channel corresponding to the first service according to the first model information, and traffic belonging to the first quality of service channel is provided with the same forwarding processing.

[0481] Optionally, the acquisition module 910 is specifically used to acquire first strategy information, which includes the aforementioned first model information and first service information.

[0482] Optionally, the processing module 920 is specifically used to determine the first quality of service channel corresponding to the first strategy information based on the aforementioned first model information.

[0483] Optionally, the aforementioned first model information is used to indicate the forwarding process of the first service, including: the aforementioned first model information is used for any one or more of the following of the first service: burst prediction, encrypted flow detection, base station resource prediction, or congestion prediction.

[0484] Optionally, the aforementioned first service quality channel corresponds only to the aforementioned first service; or, the aforementioned first service quality channel does not correspond to any other service besides the aforementioned first service.

[0485] Optionally, the acquisition module 910 is further configured to acquire second model information, which is used for forwarding processing of the second service; the processing module 920 is further configured to determine the quality of service channel corresponding to the second service based on the second model information.

[0486] Optionally, the processing module 920 is specifically used to determine the first service quality channel corresponding to the second service when the first model information and the second model information are the same.

[0487] Optionally, the above processing module 920 is specifically used to determine that the second service corresponds to the second quality of service channel when the first model information and the second model information are different, and the second quality of service channel is different from the first quality of service channel.

[0488] Optionally, the first model information includes the identification information of the first model and / or the model feature information corresponding to the first model, and the second model information includes the identification information of the second model and / or the model feature information corresponding to the second model.

[0489] Optionally, the first model information and the second model information mentioned above are the same, including: the identification information of the first model and the identification information of the second model are the same; or, the model feature information corresponding to the first model and the model feature information corresponding to the second model are the same; or, the identification information of the first model and the identification information of the second model are the same, and the model feature information corresponding to the first model and the model feature information corresponding to the second model are the same.

[0490] Optionally, the acquisition module 910 is further configured to acquire the first service quality parameter corresponding to the first service and the second service quality parameter corresponding to the second service; the processing module 920 is specifically configured to determine the first service quality channel corresponding to the second service when the first model information and the second model information are the same, and the first service quality parameter and the second service quality parameter are the same.

[0491] Optionally, the above model feature information may include one or more of the following: model consumer, the direction of the model's effect on the business flow, model function, model type, analysis identifier, or whether the model feature information is used for the corresponding service quality channel.

[0492] Optionally, the above-mentioned device 900 further includes a transceiver module 930, which is used to send a first rule corresponding to the first service to the terminal and / or user plane function network element. The first rule is used to control the traffic forwarding of the quality of service channel, and the first rule includes first model information.

[0493] Optionally, the transceiver module 930 is further configured to send a first indication message to the access network device, the first indication message being used to instruct the first quality of service channel to enable model-assisted quality of service control.

[0494] Optionally, the aforementioned first indication information is also used to indicate first model information.

[0495] Optionally, the transceiver module 930 is further configured to send a second rule corresponding to the first quality of service channel to the access network device, the second rule including the first model information.

[0496] Optionally, the transceiver module 930 is further configured to receive a request message from the terminal, which is used to request model-assisted quality of service control.

[0497] Optionally, the transceiver module 930 is further configured to receive second instruction information from the unified data management network element, the second instruction information being used to instruct authorized model-assisted quality of service control.

[0498] For example, when the communication device 900 is used to implement the function of the policy control network element in the method embodiment shown in FIG4, the communication device 900 may include an acquisition module 910 and a transceiver module 930, wherein the acquisition module 910 is used to acquire first model information, which is used for forwarding processing of the first service; the transceiver module 930 is used to send first policy information, which includes information of the first service and first model information, and the first model information is used to determine the quality of service channel corresponding to the first policy information.

[0499] Optionally, the acquisition module 910 is specifically used to acquire the first model information from local storage; or to receive the first model information corresponding to the first service from the network data analysis function / application function network element.

[0500] Optionally, the first model information includes the identification information of the first model and / or the model feature information corresponding to the first model. The model feature information includes one or more of the following: model consumer, the direction of the model's effect on the business flow, model function, model type, and whether the analysis identifier or model feature information is used for the corresponding service quality channel.

[0501] For example, when the communication device 900 is used to implement the function of the access network device in the method embodiment shown in FIG4, the communication device 900 may include a processing module 920 and a transceiver module 930, wherein the transceiver module 930 is used to receive the rules corresponding to the third quality of service channel, the rules including one or more model information, the one or more model information being used for the forwarding processing of the third quality of service channel; the processing module 920 is used to perform the forwarding processing of the third quality of service channel according to the one or more model information.

[0502] Optionally, the processing module 920 is specifically used to perform forwarding processing of the third quality of service channel based on the inference results corresponding to one or more of the above model information.

[0503] For example, when the communication device 900 is used to implement the functions of the terminal / user plane function network element in the method embodiment shown in FIG4, the communication device 900 may include a processing module 920 and a transceiver module 930, wherein the transceiver module 930 is used to receive first information, the first information indicating the quality of service channel corresponding to the first service and the model information corresponding to the quality of service channel; the processing module 920 is used to perform inference on the data packets in the quality of service channel through the first model, the first model being determined based on the first information, and the first model being the model corresponding to the quality of service channel.

[0504] Optionally, the transceiver module 930 is also used to send a request message for request model-assisted quality of service control.

[0505] Optionally, the transceiver module 930 is also used to indicate the inference results corresponding to the first model to the access network device.

[0506] More detailed descriptions of the above modules can be obtained directly from the relevant descriptions in the above method embodiments, and will not be repeated here.

[0507] It should be understood that the module division in the embodiments of this application is illustrative and only represents a logical functional division. In actual implementation, there may be other division methods. Furthermore, the functional modules in the various embodiments of this application can be integrated into a single processor, exist as separate physical entities, or be integrated into a single module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0508] Figure 10 is another schematic block diagram of the communication device 1000 provided in an embodiment of this application.

[0509] The communication device 1000 can be a chip system, or an apparatus configured with a chip system to implement the methods described in the above-described method embodiments. In the embodiments of this application, the chip system can be composed of chips, or it can include chips and other discrete devices.

[0510] As shown in Figure 10, the communication device 1000 may include a processor 1010, which can be used to execute computer programs or instructions in the memory to implement the steps performed by the session management network element, policy control network element, access network device, user plane function network element or terminal in any of the embodiments shown in Figures 4 to 8.

[0511] In one possible implementation, the communication device 1000 further includes a communication interface 1020. The communication interface 1020 can be used to communicate with other devices via a transmission medium, thereby enabling the communication device 1000 to communicate with other devices. The communication interface 1020 can be, for example, a transceiver, interface, pin, bus, circuit, or a device capable of transmitting and receiving functions. The processor 1010 can use the communication interface 1020 to input and output data, and to implement the steps executed by the session management network element, policy control network element, access network device, user plane function network element, or terminal in any of the embodiments shown in Figures 4 to 8.

[0512] In one possible implementation, the communication device 1000 further includes at least one memory 1030 for storing program instructions and / or data. The memory 1030 is coupled to the processor 1010. The coupling in this embodiment is an indirect coupling or communication connection between devices, units, or modules, and can be electrical, mechanical, or other forms, used for information exchange between devices, units, or modules. The processor 1010 may operate in conjunction with the memory 1030. The processor 1010 may execute program instructions stored in the memory 1030. At least one of the at least one memory may be included in the processor.

[0513] It should be understood that the coupling in the embodiments of this application is an indirect coupling or communication connection between devices, units, or modules, which can be electrical, mechanical, or other forms, used for information interaction between devices, units, or modules. The processor 1010 may operate in conjunction with the memory 1030. The specific connection medium between the processor 1010, communication interface 1020, and memory 1030 is not limited in the embodiments of this application. Optionally, the processor 1010, communication interface 1020, and memory 1030 are connected via a bus 1040. The bus 1040 is represented by a thick line in Figure 10. The connection methods between other components are only illustrative and not intended to be limiting. The bus can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used in Figure 10, but this does not indicate that there is only one bus or one type of bus.

[0514] This application also provides a computer program product, which includes a computer program (also referred to as code or instructions) that, when run, can implement the steps executed by the session management network element, policy control network element, access network device, user plane function network element or terminal in any of the embodiments shown in Figures 4 to 8.

[0515] This application also provides a computer-readable storage medium storing a computer program (also referred to as code or instructions). When the computer program is run, it can implement the steps executed by the session management network element, policy control network element, access network device, user plane function network element, or terminal in any of the embodiments shown in Figures 4 to 8.

[0516] It should be understood that the processor in the embodiments of this application can be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method embodiments can be completed by the integrated logic circuitry in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a microprocessor unit (MPU), a microcontroller unit (MCU), a graphics processing unit (GPU), an artificial intelligence processor (AI processor) or a neural processing unit (NPU), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or a combination of one or more discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software module can reside in a mature storage medium in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0517] It should also be understood that the memory in the embodiments of this application can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be a cache, random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (SLDRAM), and direct rambus RAM (DR RAM). It should be noted that the memory used in the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0518] The terms "unit," "module," etc., used in this specification can be used to refer to computer-related entities, hardware, firmware, combinations of hardware and software, software, or software in execution. In the embodiments of this application, "unit" and "module" have the same meaning and can be used interchangeably.

[0519] Those skilled in the art will recognize that the various illustrative logical blocks and steps described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software 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 application. In the several embodiments provided in this application, it should be understood that the disclosed apparatus, devices, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the shown or discussed mutual couplings or direct couplings or communication connections may be through some interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0520] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0521] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0522] In the above embodiments, the functions of each functional unit can be implemented entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions (programs). When the computer program instructions (programs) are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., digital video discs, DVDs), or semiconductor media (e.g., solid-state drives, SSDs), etc.

[0523] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the technology, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0524] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology 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 communication method, characterized in that, include: Obtain first model information, which is used for forwarding processing of the first service; Based on the first model information, the first quality of service channel corresponding to the first service is determined, and traffic belonging to the first quality of service channel is provided with the same forwarding processing.

2. The method as described in claim 1, characterized in that, The acquisition of the first model information includes: Obtain first strategy information, which includes the first model information and the first service information.

3. The method as described in claim 2, characterized in that, The step of determining the first quality of service channel corresponding to the first service based on the first model information includes: Based on the first model information, determine the first quality of service channel corresponding to the first strategy information.

4. The method according to any one of claims 1 to 3, characterized in that, The first model information is used to indicate the forwarding processing of the first service, including: The first model information is used for any one or more of the following in the first service: burst prediction, encrypted flow detection, base station resource prediction, or congestion prediction.

5. The method according to any one of claims 1 to 4, characterized in that, The first quality of service channel corresponds only to the first service; or, the first quality of service channel does not correspond to any other service besides the first service.

6. The method according to any one of claims 1 to 4, characterized in that, The method further includes: Obtain the second model information, which is used for forwarding processing of the second service; Based on the information from the second model, determine the quality of service channel corresponding to the second service.

7. The method as described in claim 6, characterized in that, The step of determining the quality of service channel corresponding to the second service based on the second model information includes: When the first model information and the second model information are the same, it is determined that the second service corresponds to the first quality of service channel.

8. The method as described in claim 6 or 7, characterized in that, The step of determining the quality of service channel corresponding to the second service based on the second model information includes: When the first model information and the second model information are different, it is determined that the second service corresponds to the second quality of service channel, and the second quality of service channel is different from the first quality of service channel.

9. The method according to any one of claims 6 to 8, characterized in that, The first model information includes the identification information of the first model and / or the model feature information corresponding to the first model, and the second model information includes the identification information of the second model and / or the model feature information corresponding to the second model.

10. The method as described in claim 9, characterized in that, The first model information and the second model information are the same, including: The identification information of the first model is the same as the identification information of the second model; or, The model feature information corresponding to the first model is the same as the model feature information corresponding to the second model; or, The identification information of the first model is the same as that of the second model, and the model feature information corresponding to the first model is the same as that corresponding to the second model.

11. The method according to any one of claims 6 to 10, characterized in that, The method further includes: Obtain the first service quality parameter corresponding to the first service and the second service quality parameter corresponding to the second service; The step of determining the quality of service channel corresponding to the second service based on the second model information includes: When the first model information and the second model information are the same, and the first service quality parameter and the second service quality parameter are the same, it is determined that the second service corresponds to the first service quality channel.

12. The method as described in claim 9 or 10, characterized in that, The model feature information includes one or more of the following: model consumer, the direction of the model's effect on the business flow, model function, model type, analysis identifier, or whether the model feature information is used for the corresponding service quality channel.

13. The method according to any one of claims 1 to 12, characterized in that, The method further includes: Send the first rule corresponding to the first service to the terminal and / or user plane function network element. The first rule is used to control the traffic forwarding of the quality of service channel. The first rule includes the first model information.

14. The method according to any one of claims 1 to 13, characterized in that, The method further includes: Send a first indication message to the access network device, the first indication message being used to instruct the first quality of service channel to enable model-assisted quality of service control.

15. The method as described in claim 14, characterized in that, The first indication information is also used to indicate the first model information.

16. The method according to any one of claims 1 to 12, characterized in that, The method further includes: Send the second rule corresponding to the first quality of service channel to the access network device, wherein the second rule includes the first model information.

17. The method according to any one of claims 1 to 16, characterized in that, The method further includes: Receive a request message from the terminal, the request message being used to request model-assisted quality of service control.

18. The method according to any one of claims 1 to 17, characterized in that, The method further includes: Receive a second instruction from the unified data management network element, the second instruction being used to instruct authorized model-assisted quality of service control.

19. A communication method, characterized in that, include: Obtain first model information, which is used for forwarding processing of the first service; Send first policy information, which includes information about the first service and information about the first model. The first model information is used to determine the quality of service channel corresponding to the first policy information.

20. The method as described in claim 19, characterized in that, The step of obtaining the first model information corresponding to the first service includes: Retrieve the first model information from local storage; or, Receive the first model information corresponding to the first service from the network data analysis function / application function network element.

21. The method as described in claim 19 or 20, characterized in that, The first model information includes the identification information of the first model corresponding to the first service and the model feature information corresponding to the first model; the model feature information includes one or more of the following: model consumer, the direction of the model's effect on the business flow, model function, model type, analysis identifier, or whether the model feature information is used for the corresponding service quality channel.

22. A communication method, characterized in that, include: Receive the rules corresponding to the third quality of service channel, wherein the rules include one or more model information, and the one or more model information is used for the forwarding processing of the third quality of service channel; Based on the one or more model information, perform the forwarding processing of the third quality of service channel.

23. The method as described in claim 22, characterized in that, The step of performing the forwarding process of the third quality of service channel based on the one or more model information includes: Based on the inference results corresponding to the one or more model information, the forwarding process of the third quality of service channel is executed.

24. A communication method, characterized in that, include: Receive first information, the first information indicating the quality of service channel corresponding to the first service and the model information corresponding to the quality of service channel; Inference is performed on data packets in the Quality of Service channel using a first model, which is determined based on the first information and is the model corresponding to the Quality of Service channel.

25. The method as described in claim 24, characterized in that, The method further includes: Send a request message, which is used to request model-assisted quality of service control.

26. The method as described in claim 24 or 25, characterized in that, The method further includes: Indicate the inference result corresponding to the first model to the access network device.

27. A communication device, characterized in that, include: A module for performing the method as described in any one of claims 1 to 18, or the method as described in any one of claims 19 to 21, or the method as described in claims 22 or 23, or the method as described in any one of claims 24 to 26.

28. A communication device, characterized in that, include: A processor, when invoked from a computer program in memory, causes the apparatus to perform the method as claimed in any one of claims 1 to 18, or the method as claimed in any one of claims 19 to 21, or the method as claimed in claims 22 or 23, or the method as claimed in any one of claims 24 to 26.

29. A computer-readable storage medium, characterized in that, Used to store computer programs, which, when run on a computer, cause the computer to implement the method as described in any one of claims 1 to 18, or the method as described in any one of claims 19 to 21, or the method as described in claims 22 or 23, or the method as described in any one of claims 24 to 26.

30. A computer program product, the computer program product comprising instructions, characterized in that, When the instructions are executed on a computer, the computer causes the computer to implement the method as described in any one of claims 1 to 18, or the method as described in any one of claims 19 to 21, or the method as described in claims 22 or 23, or the method as described in any one of claims 24 to 26.

Images