Communication method and communication apparatus

Abstract

The present application provides a communication method and a communication apparatus. One or more computation models can be deployed on a computing node, and a computing capability supported by each computation model is reported to a network side, such as a computing management function. When a terminal device accesses a certain service, the network side can select, from the computation models supported by the computing node, a first computation model satisfying a computation delay requirement to execute a first computation task. Moreover, when the first computation model no longer satisfies the computation delay requirement, the network side can reselect a second computation model satisfying the computation delay requirement to execute a second computation task. In this way, a processing delay of the service can be controlled, thereby facilitating a delay requirement of the service to be satisfied, and improving user experience. In addition, address information of the computing node used for the first computation task is the same as address information of the computing node used for the second computation task, such that service migration can be avoided, thereby avoiding service interruption, and improving user experience.

Description

Communication method and communication apparatus

[0001] This application claims priority to the Chinese patent application No. 202411295777.5, filed on September 14, 2024, with the State Intellectual Property Office of China, and the Chinese patent application No. 202411295777.5 has the title of “Communication method and communication apparatus”, the content of which is incorporated herein by reference in its entirety. TECHNICAL FIELD

[0002] The present application relates to the field of communication, and in particular, to a communication method and a communication apparatus. BACKGROUND

[0003] Computing power is the ability of a computer to perform certain operations. The higher the computing power of a computer, the more data it can process and the faster it can process, which means that the computer has a stronger ability to complete complex computing tasks. Computing tasks are currently increasingly widely used in various businesses through computers, which may include, but are not limited to, picture rendering, automatic identification, etc.

[0004] In order to improve user experience, it is desirable to control the latency of a business. The processing latency of a business is an important factor affecting the latency of a business. Therefore, how to control the processing latency of a business to meet the latency requirement of the business is a technical problem to be solved. SUMMARY

[0005] The present application provides a communication method and a communication apparatus to control the processing latency of a business and meet the latency requirement of the business.

[0006] In a first aspect, a communication method is provided, which can be executed by a communication apparatus. The communication apparatus can be a computing management function (CMF), or can be a component in the CMF, such as a circuit, a processor, a chip, a chip system, etc., or can be a partial or whole function logic module or software of the CMF, etc. In the following, the method is described by taking the CMF as an example for the convenience of understanding and description, but this should not constitute any limitation on the present application.

[0007] Exemplarily, the method comprises: obtaining a computing latency requirement of a business; determining, according to node information of a first computing node, a first computing model satisfying the computing latency requirement from a plurality of computing models supported by the first computing node, the node information indicating an identification of each computing model in the plurality of computing models and a computing capability supported by each computing model in the plurality of computing models, the first computing model being used to execute a first computing task of the business; and sending the identification of the first computing model to the first computing node.

[0008] The first computing node can support multiple computing models, and the multiple computing models can support different computing capabilities. In other words, the first computing node can provide multiple different computing capabilities. The computing capability can be a computing latency, or can refer to other parameters capable of representing the computing capability, such as a computing speed. However, it can be understood that different computing capabilities can correspond to different computing latencies, and the faster the computing speed, the lower the computing latency. Therefore, the multiple computing models supporting different computing capabilities means that the multiple computing models have different computing latencies when processing the same computing task.

[0009] The computing latency requirement of the service can be understood as a range requirement of the service for the computing latency. As long as the computing latency of a computing model falls within the range requirement of the computing latency, it can be considered that the computing model meets the computing latency requirement of the service.

[0010] Based on the above technical solution, the CMF can find a first computing model that meets the computing latency requirement of the service according to the node information of the first computing node and the computing latency requirement of the service, and send an identifier of the first computing model to the first computing node, so that the first computing node adopts the first computing model to execute a first computing task of the service. In this way, the processing latency of the service is controlled, thereby facilitating control of the overall latency of the service to meet the latency requirement of the service.

[0011] In combination with the first aspect, in some possible implementation manners of the first aspect, in a case where the first computing model no longer meets the computing latency requirement, the method further includes: determining a second computing model that meets the computing latency requirement from the multiple computing models, the second computing model corresponding to a computing latency different from the computing latency corresponding to the first computing model, the second computing model being used to execute a second computing task of the service; and sending an identifier of the second computing model to the first computing node.

[0012] That is, once the computing latency requirement of the service is no longer met, the CMF can reselect a computing model to execute the computing task of the service. In other words, the CMF can switch the computing model in a case where the computing latency requirement is no longer met. Since the first computing node can provide multiple computing models supporting different computing capabilities, the service does not have to be bound to a computing model, and the CMF can dynamically adjust the computing model, thereby more effectively controlling the processing latency of the service and further controlling the overall latency of the service.

[0013] In combination with the first aspect, in some possible implementation manners of the first aspect, the address information of the first computing node for the first computing task is the same as the address information of the first computing node for the second computing task.

[0014] The address information of the first computing node for the first computing task can be understood as follows: after the CMF determines the first computing model for the terminal device to perform the first computing task, the CMF allocates the address information of the first computing node to the terminal device, so that the terminal device performs data interaction of the computing task with the first computing node. In other words, the address information is the address information of the first computing node that can be perceived by the terminal device.

[0015] The understanding of the address information of the first computing node for the second computing task can also refer to the above, and will not be described in detail.

[0016] As can be seen, regardless of whether the first computing model is used to perform the first computing task or the second computing model is used to perform the second computing task, the address information used by the first computing node to perform data interaction of the computing task with the terminal device is unchanged. The first computing model and the second computing model on the first computing node can share the same transport layer connection, and even if the computing model is switched, the current network connection can still be used. In other words, when the terminal device accesses the service, the address of the first computing node accessed is unchanged. Therefore, service migration caused by changes in computing resources can be avoided, service interruption can be avoided, and user experience can be improved. Especially for services that have strict requirements on time delay stability, better service experience can be obtained.

[0017] One possible case in which the first computing model no longer meets the computing time delay requirement is that the transmission time delay of the service changes, causing the computing time delay requirement to change, and further causing the computing model to be unable to meet the changed computing time delay requirement.

[0018] Optionally, the method further includes: obtaining a transmission time delay of the service, the transmission time delay being determined according to one or more of: an uplink maximum data amount of the service, a downlink maximum data amount of the service, an execution period of the service, or an uplink transmission time delay reported by the first computing node; and determining, according to the transmission time delay, that the first computing model no longer meets the computing time delay requirement.

[0019] Another possible case in which the first computing model no longer meets the computing time delay requirement is that the computing capability of the computing model changes, causing the computing model to be unable to meet the computing time delay requirement.

[0020] Optionally, the method further includes: receiving changed node information of the first computing node; and determining, according to the changed node information of the first computing node, that the first computing model no longer meets the computing time delay requirement.

[0021] In combination with the first aspect, in some possible implementation manners of the first aspect, the first computing node is a core network element or an application server (AS).

[0022] That is, the function of the first computing node can be deployed in a core network element, such as a user plane function (UPF), or can also be deployed in an AS, such as an edge application server (EAS). It can be understood that deploying the function of the first computing node in the core network element or the AS does not mean that the core network element or the AS deploying the function of the first computing node can only be used to perform computing tasks. The first computing node can be a module in an existing core network element or AS, or can also be a network element dedicated to performing computing tasks, and the present application does not limit this.

[0023] In a possible design, the first computing node is one of a plurality of computing nodes, and the plurality of computing nodes includes at least one core network element and / or at least one AS.

[0024] The plurality of computing nodes can constitute a computing power network (referred to as a computing network). The plurality of computing nodes can be deployed in different regions to provide computing services for terminals in different regions.

[0025] With reference to the first aspect, in some possible implementation manners of the first aspect, the first computing node is the core network element, and before the determining, from a plurality of computing models supported by the first computing node, a first computing model satisfying the computing latency requirement according to the node information of the first computing node, the method further includes: sending a discovery request to a network repository function (NRF), the discovery request carrying information of the computing latency requirement; and receiving, from the NRF, node information of the first computing node satisfying the computing latency requirement.

[0026] The information of the computing latency requirement is used to indicate the computing latency requirement.

[0027] It should be understood that the first computing node satisfying the computing latency requirement means that there is at least one model in the plurality of computing models supported by the first computing node satisfying the computing latency requirement. In other words, the first computing node satisfying the computing latency requirement mentioned above, that is, the computing node in which there is a computing model satisfying the computing latency requirement.

[0028] As described before, the first computing node can be one of the computing nodes in the computing network. The NRF can select a computing node satisfying the computing latency requirement from the computing network according to the computing latency requirement, and send node information of the computing node to the CMF.

[0029] Optionally, after the node information of the first computing node meeting the computing latency requirement is acquired, the method further includes: sending a first subscription request to the NRF, the first subscription request being used to subscribe to a change of the node information of the first computing node.

[0030] It can be understood that the changed change information of the first computing node can be sent by the NRF in response to the first subscription request. The NRF can send the changed node information of the first computing node to the CMF in the case that the changed node information of the first computing node is received.

[0031] Therefore, the CMF can obtain the node information of the first computing node in real time by subscribing to the change of the node information of the first computing node, so as to adjust the computing model in time in the case that the computing latency requirement is no longer met according to the changed node information of the first computing node, thereby more effectively controlling the processing latency of the service and further controlling the overall latency of the service.

[0032] In combination with the first aspect, in some possible implementation manners of the first aspect, the first computing node is the AS, and before the first computing node meeting the computing latency requirement is determined according to the node information of the first computing node from a plurality of computing models supported by the first computing node and meeting the computing latency requirement, the method further includes: determining that the first computing node meets the computing latency requirement according to the node information of the first computing node.

[0033] As described before, the first computing node can be one computing node in a computing network. The CMF can select a computing node meeting the computing latency requirement from the computing network according to the computing latency requirement. It should be understood that since whether the first computing node meets the computing latency requirement can be determined according to whether there is a computing model meeting the computing latency requirement on the first computing node, in a specific implementation, determining that the first computing node meets the computing latency requirement according to the node information of the first computing node and determining the first computing model meeting the computing latency requirement from a plurality of computing models supported by the first computing node according to the node information of the first computing node can be implemented through one determination step, that is, determining the first computing model meeting the computing latency requirement from a plurality of computing models supported by the first computing node according to the node information of the first computing node, without necessarily performing the operation of determining twice. Determining that the first computing model supported by the first computing node meets the computing latency requirement is equivalent to determining that the first computing node meets the computing latency requirement.

[0034] Optionally, before the determining that the first computing node meets the computing latency requirement according to the node information of the first computing node, the method further includes: sending a second subscription request to a network exposure function (NEF), the second subscription request being used for subscribing to node information of the AS; and receiving the node information of the AS from the NEF.

[0035] For computing nodes deployed outside the core network, the NEF can provide them with an interface for interacting with the network. The NEF can forward the node information of the AS to the CMF, so that more computing nodes are used to provide computing services.

[0036] Optionally, the method further includes: receiving a session establishment request from a terminal device; and sending a session establishment response to the terminal device, the session establishment response carrying address information of the first computing node.

[0037] With reference to the first aspect, in some possible implementation manners of the first aspect, the obtaining of the computing latency requirement of the service includes: obtaining an end-to-end latency threshold of the service and a transmission latency; and determining the computing latency requirement according to the end-to-end latency of the service and the transmission latency.

[0038] The latency requirement of the service can mean that the end-to-end latency of the service does not exceed a predefined end-to-end latency threshold. The end-to-end latency of the service can include a processing latency of the service and a transmission latency. The processing latency of the service is the latency required for processing service data by a processing resource, for example, the computing latency required for computing by a computing model in the present application. The transmission latency of the service is the latency caused by transmission of service data in the network, including uplink transmission latency and downlink transmission latency. Therefore, the computing latency requirement of the service is determined according to the end-to-end latency threshold of the service and the transmission latency, and then a computing model with the computing latency requirement is selected to perform a computing task, which is beneficial to meeting the latency requirement of the service.

[0039] Optionally, the transmission latency is determined according to one or more of the following: an uplink maximum data amount of the service, a downlink maximum data amount of the service, or an execution period of the service.

[0040] The transmission latency can be obtained by model prediction. The uplink maximum data amount of the service, the downlink maximum data amount of the service, and the execution period of the service can be used to describe the characteristics of the service. Taking these parameters as inputs of a model, a more accurate transmission latency can be predicted, which is beneficial to obtaining a more accurate computing latency requirement and meeting the latency requirement of the service.

[0041] With reference to the first aspect, in some possible implementation of the first aspect, the method further includes: sending, to the first computing node, an identifier of the terminal device.

[0042] The same computing node can provide computing services for multiple terminal devices at the same time. Therefore, the computing node can distinguish computing tasks from different terminal devices through the identifiers of the terminal devices, so as to avoid routing the same computing task of the same terminal device to different computing models, and further avoid problems such as service interruption caused thereby.

[0043] The second aspect provides a communication method, which can be executed by a communication apparatus. The communication apparatus can be a computing node, or can be a device in which the computing node is deployed, or can be a component in the computing node, such as a circuit, a processor, a chip, a chip system, and the like, or can be a partial or whole functional logic module or software of the computing node, and the like. Hereinafter, for the convenience of understanding and description, the method is described by taking the computing node as an example, but this should not constitute any limitation to the present application.

[0044] Exemplarily, the method includes: determining an identifier of each of one or more computing models supported by the computing node, and a computing capability supported by each of the one or more computing models; and sending node information, the node information indicating the identifier of each of the one or more computing models, and the computing capability supported by each of the one or more computing models.

[0045] The computing node can load one or more computing models. The computing node can send the identifier of each of the computing models and the supported computing capability through the node information, so that when there is a service demand of the terminal device, a suitable computing node and computing model can be selected according to the computing capability of the computing node.

[0046] A possible design is that the computing node is a virtual machine or a container built on hardware resources (such as a graphics processing unit (GPU) or a central processing unit (CPU)), and the computing model can be a model loaded on the virtual machine or the container.

[0047] With reference to the second aspect, in some possible implementation of the second aspect, the computing node is a core network element or an AS.

[0048] With reference to the second aspect, in some possible implementation of the second aspect, the computing node supports a plurality of computing models, and the method further includes: receiving an identification of a first computing model, the first computing model being determined from the plurality of computing models according to the node information; and performing a first computing task of the service by using the first computing model.

[0049] Optionally, the method further includes: receiving an identification of a terminal device; and determining, according to the identification of the terminal device, that the first computing task is performed on the terminal device by using the first computing model.

[0050] With reference to the second aspect, in some possible implementation of the second aspect, the method further includes: receiving an identification of a second computing model, the second computing model being determined from the plurality of computing models according to the node information, and the second computing model supporting a different computing capability from the first computing model; and performing a second computing task of the service by using the second computing model.

[0051] Optionally, the method further includes: receiving an identification of a terminal device; and determining, according to the identification of the terminal device, that the second computing task is performed on the terminal device by using the second computing model.

[0052] With reference to the second aspect, in some possible implementation of the second aspect, address information of the first computing node for the first computing task is same as address information of the first computing node for the second computing task.

[0053] For the possible implementation and beneficial effects of the second aspect, refer to the related description of the first aspect, and no further repetition is made.

[0054] In a third aspect, a communication method is provided, which can be performed by a communication apparatus. The communication apparatus can be an NRF, or can be a component (such as a circuit, a processor, a chip, a chip system, etc.) in the NRF, or can be a partial or whole function logic module or software of the NRF, etc. Hereinafter, for the convenience of understanding and description, the method is described by taking the NRF as an example, but this should not constitute any limitation to the present application.

[0055] Exemplarily, the method comprises: receiving a discovery request from a CMF, the discovery request carrying information of a computing latency requirement; determining, according to node information of a first computing node, that the first computing node meets the computing latency requirement, the first computing node supporting a plurality of computing models, and computing capabilities of one or more computing models in the plurality of computing models meeting the computing latency requirement; and sending, to the CMF, node information of the first computing node, the node information indicating an identity of each computing model in the plurality of computing models and a computing capability supported by each computing model in the plurality of computing models.

[0056] With reference to the third aspect, in some possible implementation manners of the third aspect, the node information is used to determine the computing node and / or the computing model that meet the computing latency requirement.

[0057] With reference to the third aspect, in some possible implementation manners of the third aspect, the first computing node is a core network element.

[0058] With reference to the third aspect, in some possible implementation manners of the third aspect, the method further comprises: receiving a subscription request from the CMF, the subscription request being used to subscribe to a change in the node information of the first computing node.

[0059] With reference to the third aspect, in some possible implementation manners of the third aspect, the method further comprises: in a case where the changed node information of the first computing node is received, sending, to the CMF, the changed node information of the first computing node.

[0060] For the possible implementation manners and beneficial effects of the third aspect, refer to the related description of the first aspect, which will not be repeated here.

[0061] A fourth aspect provides a communication method, which can be executed by a communication apparatus. The communication apparatus can be an NEF, or can be a component in the NEF, such as a circuit, a processor, a chip, a chip system, etc., or can be a part or all of the functional logic modules or software of the NEF, etc. Hereinafter, for the convenience of understanding and description, the method is described by taking the NEF as an example, but this should not constitute any limitation to the present application.

[0062] Exemplarily, the method comprises: receiving node information of a computing node, the node information comprising: an identity of each computing model in one or more computing models supported by the computing node, and a computing capability supported by each computing model in the one or more computing models; receiving a subscription request from a CMF, the subscription request being used to request the node information of the computing node; and sending, to the CMF, the node information of the computing node.

[0063] In a possible implementation form of the fourth aspect, the node information is used to determine a computing node and / or a computing model that satisfies the computing latency requirement.

[0064] For the possible implementation forms and beneficial effects of the fourth aspect, refer to the related description of the first aspect.

[0065] In combination with the fifth aspect, a communication method is provided, which can be performed by a communication apparatus. The communication apparatus can be an application function (AF), or can be a component in the AF, such as a circuit, a processor, a chip, a chip system, etc., or can be a partial or whole function logic module or software of the AF, etc. Hereinafter, for the convenience of understanding and description, the method is described by taking the AF as an example, but this should not constitute any limitation to the present application.

[0066] Exemplarily, the method comprises: generating a request message carrying an end-to-end latency threshold corresponding to a supported service, the end-to-end latency threshold being used to determine a computing latency requirement of the service; and sending the request message.

[0067] The AF can send the end-to-end latency threshold of the service to a policy control function (PCF) through the request message, so as to provide the end-to-end latency threshold corresponding to the service to the CMF when the service is accessed, so that the CMF determines the computing latency requirement, thereby facilitating the CMF to select a computing model satisfying the computing latency requirement to perform a computing task, and further facilitating the latency requirement of the service to be met.

[0068] In a possible implementation form of the fifth aspect, the request message further carries one or more of the following corresponding to the service: a data network name (DNN), a single network slice selection assistance information (S-NSSAI), a maximum data volume, a minimum data volume, or an execution period.

[0069] The DNN and / or S-NSSAI correspond to the service, and carrying the DNN and / or S-NSSAI in the request message can determine the end-to-end latency threshold in the request message. The maximum data volume, the minimum data volume, and the execution period can be used to describe the characteristics of the service, and can be used as inputs of the model to predict the transmission latency.

[0070] In a sixth aspect, a communication apparatus is provided, which can implement the method in the first to fifth aspects and / or any possible implementation of the first to fifth aspects. The apparatus includes one or more function units or modules for performing the corresponding method steps. The function units or modules included in the apparatus can be implemented by software and / or hardware.

[0071] In a seventh aspect, a communication apparatus is provided, which includes a processor configured to perform the method in the first to fifth aspects and / or any possible implementation of the first to fifth aspects.

[0072] Optionally, the apparatus can further include a memory configured to store instructions and data. The memory is coupled to the processor, and the processor is configured to execute the instructions stored in the memory, thereby implementing the method described in the aspects.

[0073] Optionally, the apparatus can further include a communication interface configured to enable the apparatus to communicate with other devices. The communication interface can be, for example, a transceiver, a circuit, a bus, a module, or other types of communication interfaces.

[0074] In an eighth aspect, a chip system is provided, which includes at least one processor configured to support the implementation of the functions related to the first to fifth aspects and / or any possible implementation of the first to fifth aspects, such as receiving or processing the data and / or information related to the method.

[0075] In a possible design, the chip system further includes a memory configured to store program instructions and data, and the memory is located in or out of the processor.

[0076] In a possible design, the chip system further includes an interface circuit configured to transmit data and / or a power supply circuit configured to supply power to the chip system.

[0077] The chip system can be composed of a chip, or can include a chip and other discrete devices.

[0078] In a ninth aspect, a communication system is provided, which includes one or more of the following: a CMF, a computing node, an NRF, an NEF, an AF, or a terminal device. The CMF can be configured to implement the method in any possible implementation of the first aspect. The computing node can be configured to implement the method in any possible implementation of the second aspect. The NRF can be configured to implement the method in any possible implementation of the third aspect. The NEF can be configured to implement the method in any possible implementation of the fourth aspect. The AF can be configured to implement the method in any possible implementation of the fifth aspect. Optionally, the computing node is a UPF. Optionally, the computing node is an AS.

[0079] In a tenth aspect, a computer-readable storage medium is provided, which includes a computer program, when the computer program is run on a computer, causes the computer to implement the method in the first to fifth aspects and any possible implementation of the first to fifth aspects.

[0080] In an eleventh aspect, a computer program product is provided, which includes a computer program (also referred to as code or instructions), when the computer program is run, causes a computer to execute the method in the first to fifth aspects and any possible implementation of the first to fifth aspects.

[0081] It should be understood that the sixth aspect to the eleventh aspect of the present application correspond to the technical solutions of the first aspect to the fifth aspect of the present application, and the beneficial effects achieved by each aspect and the corresponding possible implementation are similar, which will not be repeated. BRIEF DESCRIPTION OF DRAWINGS

[0082] FIG. 1 is a schematic diagram of a 5th generation mobile communication technology (5G) network architecture based on a service-oriented architecture, which is applicable to the communication method provided by the embodiments of the present application;

[0083] FIG. 2 is a schematic diagram of a 5G network architecture based on a point-to-point interface, which is applicable to the communication method provided by the embodiments of the present application;

[0084] FIG. 3 is a schematic diagram of a scenario, which is applicable to the communication method provided by the embodiments of the present application;

[0085] FIG. 4 is a schematic flowchart of the communication method provided by the embodiments of the present application;

[0086] FIG. 5 is a schematic flowchart of obtaining a prediction model provided by the embodiments of the present application;

[0087] FIG. 6 is a schematic diagram of the communication method provided by the embodiments of the present application applied to different scenarios;

[0088] FIG. 7 is a schematic flowchart of a communication method according to another embodiment of the present application;

[0089] FIG. 8 shows a possible flow of a communication method according to another embodiment of the present application;

[0090] FIG. 9 shows another possible flow of a communication method according to another embodiment of the present application;

[0091] FIG. 10 is a schematic diagram of a communication method according to yet another embodiment of the present application;

[0092] FIG. 11 shows a possible flow of a communication method according to yet another embodiment of the present application;

[0093] FIG. 12 shows another possible flow of a communication method according to yet another embodiment of the present application;

[0094] FIG. 13 is a schematic block diagram of a communication apparatus according to an embodiment of the present application;

[0095] FIG. 14 is another schematic block diagram of a communication apparatus according to an embodiment of the present application. DETAILED DESCRIPTION

[0096] The technical solutions in the present application will be described below with reference to the accompanying drawings.

[0097] For the convenience of understanding, the following points are explained before introducing the solutions of the present application.

[0098] First, in the present application, indication includes explicit indication (also referred to as direct indication) and implicit indication (also referred to as indirect indication). Wherein, explicit indication of information A means including the information A; implicit indication of information A means indicating the information A through the correspondence between the information A and information B and the direct indication of information B, the correspondence between the information A and information B can be predefined, pre-stored, pre-burned, or pre-configured; or, it can also mean indicating the information A through the information B and a preset rule.

[0099] Second, in the present application, information C is used for determining information D, which includes that information D is determined based on information C only, and includes that information D is determined based on information C and other information. In addition, information C used for determining information D can also include the case of indirect determination, such as the case that information D is determined based on information E, and information E is determined based on information C.

[0100] Third, in this application, "at least one" means one or more, and "multiple" means two or more. The expression " / " is used to represent the "or" relationship of the objects before and after it; for example, A / B can represent: A or B. The expression "and / or" is used to represent the "and" relationship or the "or" relationship of the objects before and after it, for example, A and / or B, which can represent: the existence of A alone, the existence of A and B, and the existence of B alone, where A and B can be singular or plural. The character " / " generally represents an "or" relationship between the objects before and after it, but does not exclude the case of representing an "and" relationship between the objects before and after it. The specific meaning can be understood in combination with the context. "At least one" or similar expressions refer to any combination of these items, including any combination of single or multiple items. For example, at least one of a, b, or c can represent: a, b, c; a and b; a and c; b and c; or a and b and c. Where a, b, and c can be single or multiple.

[0101] Fourth, in this application, the use of prefixes such as "first", "second", and the like is only for the convenience of distinguishing and describing different things belonging to the same name category, and does not constrain the order, size, or quantity of the things. For example, "first computing model" and "second computing model" are only different computing models, and there is no time sequence, size relationship or priority relationship between them; for example, "first computing task" and "second computing task" are only different computing tasks, and there is no time sequence, size relationship or priority relationship between them. It should be understood that the objects thus described can be interchanged under appropriate circumstances to describe solutions other than the embodiments of the application.

[0102] Fifth, in this application, "send" and "receive" represent the direction of signal transmission. For example, "sending information to XX" can be understood as the destination of the information being XX, which can include direct transmission through the air interface, or indirect transmission through the air interface by other units or modules. "Receiving information from YY" can be understood as the source of the information being YY, which can include direct reception from YY through the air interface, or indirect reception from YY through 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. In other words, sending and receiving can be between devices, such as between terminal devices and computing nodes, or within a device, such as between components, modules, chips, software modules or hardware modules within a device through a bus, wire or interface.

[0103] Sixth, in the embodiments of the present application, "when", "if" and "provided that" all refer to the device making corresponding processing under certain objective conditions, and are not limited in time, and do not require the device to have a judgment action when implemented, nor does it mean that there are other limitations.

[0104] Seventh, in the present application, "example", "exemplarily", "for example" or "such as" means to take as an example, illustration or description. Any embodiment or design scheme described as "example", "exemplarily", "for example" or "such as" in the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the use of "example", "exemplarily", "for example" or "such as" is intended to present the relevant concept in a specific manner.

[0105] Eighth, in various embodiments of the present application, the terms and / or descriptions of different embodiments are consistent and can be referred to each other if there is no special description and logical conflict, and the technical features in different embodiments can be combined to form new embodiments according to their inherent logical relationship.

[0106] The method provided by the present application can be applied to various communication systems, for example: long term evolution (LTE) system, LTE frequency division duplex (FDD) system, LTE time division duplex (TDD) system, 5G mobile communication system or new radio access technology (NR). Among them, the 5G mobile communication system can include non-standalone (NSA) and / or standalone (SA).

[0107] The technical solutions provided in the present application can also be applied to machine type communication (MTC), long term evolution-machine (LTE-M), device-to device (D2D) network, machine to machine (M2M) network, internet of things (IoT) network or other network. The IoT network may, for example, include a vehicle network. In the vehicle network system, the communication modes are collectively referred to as vehicle to X (V2X) system, for example, the V2X may include vehicle to vehicle (V2V) communication, vehicle to infrastructure (V2I) communication, vehicle to pedestrian (V2P) communication or vehicle to network (V2N) communication, etc. The technical solutions provided in the present application can also be applied to future communication systems. The present application does not limit this.

[0108] For the convenience of understanding, first, the network architecture applicable to the method provided in the embodiments of the present application will be described in more detail.

[0109] FIG. 1 is a schematic diagram of a 5G network architecture based on serving based architecture (SBA) applicable to the communication method provided in the embodiments of the present application. The 5G network architecture shown in FIG. 1 can include three parts, which are terminal device part, data network (DN) and operator network part.

[0110] FIG. 2 is a schematic diagram of a 5G network architecture based on point-to-point interface applicable to the communication method provided in the embodiments of the present application. The functions of the network elements in FIG. 2 can refer to the functions of the corresponding network elements in FIG. 1, and will not be described again. The main difference between FIG. 1 and FIG. 2 is that the interfaces between the network elements in FIG. 2 are point-to-point interfaces, rather than service interfaces.

[0111] Among them, the operator network can include one or more of the following network elements: network slice specific authentication and authorization function (NSSAAF) network element, network slice selection function (NSSF) network element, network slice admission control function (NSACF) network element, CMF network element, authentication server function (AUSF) network element, network exposure function (NEF) network element, policy control function (PCF) network element, unified data management (UDM) network element, unified data repository (UDR), network repository function (NRF) network element, application function (AF) network element, access and mobility management function (AMF) network element, session management function (SMF) network element, radio access network (RAN), and user plane function (UPF) network element, etc. Among the above operator network, except for the radio access network part, the part can be referred to as the core network part. In the following, for the convenience of description, when describing each core network element, the word "network element" is omitted, such as AF network element is referred to as AF, UDM network element is referred to as UDM, CMF network element is referred to as CMF, etc.

[0112] The functions of the network elements involved in FIG. 1 and FIG. 2 are briefly described below.

[0113] Terminal device: a device with wireless transceiver function. The terminal device can communicate with one or more core network (CN) devices (or core devices) through the access network device (or access device) in the wireless access network. The terminal device can be deployed on land, including indoor or outdoor, handheld or vehicle-mounted; can also be deployed on water (such as ships, etc.); can also be deployed in the air (such as airplanes, balloons and satellites, etc.).

[0114] The terminal device can also be a terminal device in an internet of things (IoT) system, which can also be referred to as an IoT node. The IoT is an important component of future information technology development, and its main technical feature is to connect objects through communication technology and network, so as to realize the intelligent network of man-machine interconnection and object-object interconnection. The connection can be through broadband technology or narrow band (NB) technology. IoT technology can achieve mass connection, deep coverage, and power saving of terminal devices through, for example, narrow band technology.

[0115] In an embodiment of the present application, the apparatus for implementing the function of the terminal device can be a terminal device, or an apparatus capable of supporting the terminal device to implement the function, such as a chip system, which can be installed in the terminal device or used in matching with the terminal device. In an embodiment of the present application, the chip system can be composed of a chip, or can include a chip and other discrete devices. In the present application, only the apparatus for implementing the function of the terminal device is taken as an example for description, and the scheme of the present application is not limited.

[0116] The terminal device in the present application can be a hardware device, or a software function running on a special hardware, or a software function running on a general hardware, or a virtualized device, such as a general hardware and an instantiated virtualization function, or a special hardware and an instantiated virtualization function. The general hardware can be a server, such as a cloud server.

[0117] RAN: used to implement wireless related functions. The RAN can be a network composed of multiple RAN nodes, which implements wireless physical layer functions, resource scheduling and radio resource management, radio access control, and mobility management functions. The 5G-RAN can be connected to the UPF through a user plane interface (i.e. N3 interface) and used to transmit data of the terminal device; the 5G-RAN establishes a control plane signaling connection with the AMF through a control plane interface (i.e. N2 interface), and is used to implement radio access bearer control and other functions.

[0118] The RAN node can provide wireless communication function services and access the terminal device to the wireless network. The RAN node can also be referred to as a RAN device, or an access network device, etc.

[0119] In one possible scenario, the RAN node can be a base station, an evolved Node B (eNodeB), an access point (AP), a transmission reception point (TRP), a next generation NodeB (gNB), or a base station in a future mobile communication system. The RAN node can be a macro base station, a micro base station, or an indoor station, a relay node or a donor node, or a radio controller in a cloud radio access network (CRAN) scenario. Optionally, the RAN node can also be a server.

[0120] In another possible scenario, a terminal device is assisted by multiple RAN nodes to implement wireless access, and different RAN nodes respectively implement part of functions of a base station. For example, the RAN node can be a central unit (CU), a distributed unit (DU), a CU-control plane (CP), a CU-user plane (UP), or a radio unit (RU), etc. The CU and the DU can be separately arranged, or can be included in the same network element, such as a baseband unit (BBU). The RU can be included in a radio frequency device or a radio frequency unit, such as a remote radio unit (RRU), an active antenna unit (AAU), or a remote radio head (RRH).

[0121] In different systems, the CU (or CU-CP and CU-UP), DU, or RU can also have different names, but those skilled in the art can understand their meanings. For example, in an open RAN (O-RAN or ORAN) system, the CU can also be referred to as an open CU (O-CU), the DU can also be referred to as an O-DU, the CU-CP can also be referred to as an O-CU-CP, the CU-UP can also be referred to as an O-CU-UP, and the RU can also be referred to as an O-RU. For the convenience of description, the CU, CU-CP, CU-UP, DU, and RU are taken as examples for description in this application. Any one of the CU (or CU-CP, CU-UP), DU, and RU in this application can be implemented by a software module, a hardware module, or a combination of a software module and a hardware module.

[0122] In the embodiments of this application, the device for implementing the function of the RAN node can be the RAN node itself; or can be a device capable of supporting the RAN node to implement the function, such as a chip system, a hardware circuit, a software module, or a hardware circuit plus a software module. The device can be installed in the RAN node or used in the RAN node. In the embodiments of this application, only the device for implementing the function of the RAN node is taken as an example to illustrate the RAN node, and the scheme of the embodiments of this application is not limited.

[0123] The RAN node in this application can be a hardware device, or a software function running on a dedicated hardware, or a software function running on a general-purpose hardware, or a virtualized device, such as a general-purpose hardware and an instantiated virtualization function, or a dedicated hardware and an instantiated virtualization function. The general-purpose hardware can be a server, such as a cloud server.

[0124] AMF: for mobility management function, responsible for user mobility management, including mobile state management, allocation of user temporary identity, authentication and authorization of user;

[0125] SMF: mainly responsible for UPF network element selection, UPF network element reselection, internet protocol (IP) address allocation, responsible for establishment, modification and release of bearer, quality of service (QoS) control;

[0126] UDM: mainly responsible for managing subscription data, and responsible for notifying the corresponding network element when the subscription data is modified.

[0127] UDR: mainly used for storing and retrieving subscription data, policy data and public architecture data, etc.; for UDM, PCF and NEF to obtain related data. UDR should be able to have different data access authentication mechanisms for different types of data such as subscription data and policy data, to ensure the security of data access; UDR should be able to return a failure response with appropriate reason value for illegal service operation or data access request.

[0128] NEF: used for interaction between other internal network elements of the core network and application function (AF) network elements corresponding to AS outside the core network, to provide network opening capability to AF, or provide information provided by AF to core network elements.

[0129] AF: provides certain application layer services to the UE, and has requirements for quality of service (QoS) policy and charging policy when providing services to terminal devices, and needs to inform the network. At the same time, the AF also needs application-related information fed back by the core network.

[0130] UPF: supports all or part of the following functions: interworking between protocol data unit (PDU) sessions and data networks; packet routing and forwarding (for example, supporting uplink classifier traffic forwarding to data networks, supporting branching points to support multi-homed PDU sessions; packet detection. In the embodiments of the present application, the UPF can act as a computing node to provide computing services.

[0131] CMF: a control plane network element for computing management, responsible for managing computing power resources, intelligent scheduling of computing power, selecting or reselecting computing nodes, etc. The CMF can be a newly added network element in the core network (not shown in FIG. 2); it can also be integrated into an existing core network element, for example, the CMF can also be an SMF or an AMF with added computing power management functions, or the CMF can also be a module in an SMF, an AMF, or an access network element; it can also be used to replace an existing core network element, for example, the CMF can be used to replace one or more of an SMF, an AMF, or an access network element. It can be understood that the CMF can have the functions of computing management control and one or more of an SMF, an AMF, or an access network element. As a control plane network element, the CMF can interact with other core network elements of the core network, such as AMF, SMF, etc. In addition, the CMF can also interact with the access network (such as access network devices, terminal devices, etc.), computing nodes.

[0132] It should be understood that FIG. 1 is only an example showing a possible deployment location of the CMF as a newly added network element, but this should not constitute any limitation on the present application. The present application does not limit the product form and deployment location of the CMF.

[0133] NRF: supports service discovery functions, receives NF discovery requests from network function (NF) instances, and provides information about discovered NF instances (discovered) to NF instances.

[0134] DN: mainly provides service services for users. The AS deployed in the DN can be used to run applications (APPs), provide service services for users, for example, the user can be provided with an access interface of the application through a client, and the access request of the user is processed. In the embodiments of the present application, the AS can serve as a computing node to provide computing services.

[0135] Computing node: also referred to as a computing unit, an intelligent node, an intelligent unit, etc. The computing node can be used to provide computing services, and through the services, one or more functions of connection, calculation or data storage are provided for users. Optionally, the computing node can be deployed in the core network, for example, it can be a newly added core network element (as shown in FIG. 1), or it can also be integrated in an existing core network element (such as a UPF). Optionally, the computing node can also be deployed outside the core network, for example, it can be deployed in the DN, and it can be a newly added network element in the DN, or it can also be integrated in an existing network element (such as an AS). The specific form and deployment location of the computing node are not limited in the present application.

[0136] The network elements communicate with each other through interfaces. For example, in FIG. 2, the interface between the terminal and the AMF is the N1 interface, the interface between the (R)AN and the AMF is the N2 interface, the interface between the (R)AN and the UPF is the N3 interface, the interface between the SMF and the UPF is the N4 interface, and the interface between the UPF and the DN is the N6 interface. Some network elements can communicate based on a service interface, for example, Nnssf, Nnef, Nnrf, Npcf, Namf, Nsmf, Ncmf, etc. in FIG. 1. Among them, the interface Ncmf is only one possible name, and the name of the service interface corresponding to the CMF is not limited in the present application.

[0137] It should be noted that the AS is used to provide service content, for example, including but not limited to, multimedia conference, converged communication, etc. Among them, the AS can be deployed at the edge side, such as multi-access edge computing (MEC), and the AS deployed at the edge side can be referred to as EAS (edge application server). The control plane function presented to the outside by the EAS and the MEC platform can be referred to as AF. The AF is a network element defined in the 3rd generation partnership project (3GPP) service & system aspects (SA) working group 2. When the AS needs to interact with the core network element through signaling, such as the AS sending a subscription request to the core network element, the AS can realize the forwarding process of the subscription request to the core network element through the AF; conversely, when the core network element needs to interact with the AS to deliver some parameters on the AS side, such as the load condition on the AS side, the AS can realize the interaction process between the core network element through the AF. The AF and the AS can be two physical devices separated in hardware form, or can be integrated in one physical device. The specific form of the AF and the AS is not limited in the present application.

[0138] FIG. 3 is a schematic diagram of a scenario suitable for the communication method provided by the embodiments of the present application. FIG. 3(a) and (b) respectively show two examples of deploying a computing node in a core network element and in an application server. Among them, the core network element includes but is not limited to UPF; the application server includes but is not limited to a server deployed in a multi-access edge computing (MEC) platform, such as EAS.

[0139] Deploying a computing node in a core network element can be understood as that the function of the computing node is realized through the core network element, for example, it can directly interact with other network elements of the core network such as SMF, etc. Further, the core network element not only has its original function as a core network element, but also has the function of the computing node, which can be used to perform computing tasks. For example, deploying a computing node in a UPF, the UPF not only supports its user plane function, for example, all or part of the functions described above: interworking PDU session and data network; packet routing and forwarding (for example, supporting uplink classification and forwarding to data network, supporting branch nodes to support multi-homing PDU session), etc., but also has the function of the computing node, which can be used to perform computing tasks. Deploying a computing node in a core network element, that is, the core network element can be regarded as a computing node in the entire computing network (or called computing power network, algorithm network). Therefore, deploying a computing node in a core network element can also be referred to as that the computing node is the core network element.

[0140] Similarly, deploying a computing node on an application server can be understood as that the function of the computing node is implemented through the application server. That is, the application server has the function of the computing node in addition to its original function, and can be used to perform a computing task. Deploying a computing node on an application server means that the application server can be regarded as a computing node in the entire computing network. Therefore, deploying a computing node on an application server can also be referred to as that the computing node is the application server.

[0141] In (a) of FIG. 3, the UPF can be used as a computing node as an example of a core network element. After receiving a service request sent by a terminal device, the UPF can process the service message and return the processing result to the terminal device. It can be understood that the UPF not only has the message forwarding capability, but also has the processing capability of the service message. Alternatively, it can be understood that the traditional service processing server and the UPF are integrated in the same physical node, and after the data message sent by the terminal device reaches the UPF, the UPF integrated with the service processing capability can process the message.

[0142] In (b) of FIG. 3, the application server is used as a computing node. After receiving a service request sent by a terminal device, the application server can process the service message and return the processing result to the terminal device.

[0143] It should be understood that the deployment of the CMF in FIG. 3 is only an example, and the deployment of the CMF can be combined with the description of the deployment of the CMF in FIGS. 1 and 2.

[0144] Computing power refers to the ability of a computer to perform certain operations, which is usually measured by the speed of floating point operations (usually in floating point operations per second (FLOPS) units). The higher the computing power, the more data the computer can process and the faster the processing speed, which also means that the computer has stronger ability to complete complex computing tasks.

[0145] The central processing unit (CPU), also known as the central processor, is an important component of a computer system, used to execute various instructions and control the operation of the computer. The CPU is located on the computer motherboard and is one of the most important components of the computer, bearing a large amount of calculation and computing tasks. Generally speaking, CPUs are suitable for performing general-purpose computing tasks, and their computing power is usually low.

[0146] A graphics processing unit (GPU), also known as a graphics processor, is a type of processor designed specifically for efficient processing of images and graphics. It is a type of processor in a computer system that can perform parallel computing and is suitable for large-scale parallel processing tasks. Currently, GPUs have been widely used in scientific computing, computer vision, deep learning, graphics rendering, and other fields. Compared with CPUs, GPUs have more cores and higher memory bandwidth, and can process a large amount of data in a short time, making them an ideal choice for training deep learning models.

[0147] Currently, the execution of computing tasks by computers has been increasingly widely applied in various different businesses, involving computing tasks such as, but not limited to, image rendering, automatic identification, etc.

[0148] In the aspect of artificial intelligence (AI) task reasoning, by increasing computing resources, the time delay of internal processing of the server can be reduced.

[0149] Taking target recognition as an example, you only look once (YOLO) is a target detection algorithm for computer vision tasks. YOLO version 8 (v8) has 5 basic deep neural network (DNN) models, namely yolov8n, yolov8s, yolov8m, yolov8l, and yolov8x, and their parameters are shown in Table 1:

[0150] Table 1

[0151] Among them, FLOPS (floating point operations per second) represents the number of floating point operations per second, which is understood as the calculation speed.

[0152] From the values of the model parameter quantity and the required computing power, it can be seen that in the order of the models from n, s, m, l, and x, the model parameter quantity is arranged in the order from small to large, and the calculation time delay is also arranged in the order from small to large.

[0153] Currently, the execution of computing tasks by computers has been increasingly widely applied in various different businesses, involving computing tasks such as, but not limited to, image rendering, automatic identification, etc. In order to improve user experience, it is desirable to control the time delay of the business. The processing time delay of the business is an important factor affecting the time delay of the business. Therefore, how to control the processing time delay of the business to meet the time delay requirements of the business is a technical problem to be solved.

[0154] Therefore, the method provided in the present application is provided. In the method, a computing node can deploy one or more computing models, and report computing capabilities supported by each computing model to a network side. When a terminal device accesses a service, the network side can select a computing model whose computing capability meets a computing delay requirement of the service according to the computing delay requirement of the service, to perform a computing task. In this way, the processing delay of the service is controlled, and the delay requirement of the service is met, and user experience is improved.

[0155] The method provided in the present application will be described in detail below with reference to the accompanying drawings.

[0156] FIG. 4 is a schematic flowchart of a communication method provided in an embodiment of the present application. The method 400 shown in FIG. 4 can include steps 410 to 430. Optionally, steps 440 to 470 are further included. The steps of the method 400 will be described in detail below.

[0157] In step 410, the CMF obtains a computing delay requirement of a service.

[0158] The CMF can obtain the computing delay requirement of the service in the case that there is an access requirement for the service. For example, the CMF can receive a session establishment request from a terminal device, the session establishment request being used to request access to a service. Exemplarily, the session establishment request is a computing task session establishment request, which can be used to request access to a service and perform a computing task.

[0159] The computing delay requirement of the service can be understood as a range requirement of the service on a computing delay. The computing delay requirement of the service can be obtained by the CMF in advance, such as from a PCF, or from a NEF, or determined by the CMF itself, and the like, which are not limited in the present application.

[0160] The computing delay requirement of the service can be determined according to a delay requirement of the service. Exemplarily, the delay requirement of the service can be embodied by an end-to-end delay threshold of the service, for example, the delay of the service does not exceed the end-to-end delay threshold. The end-to-end delay can be a round-trip time (RTT) for example, which refers to a time required for data to be transmitted from one end of a network to the other end. In the embodiment of the present application, taking a computing task as an example, the end-to-end delay of the service can refer to a time required for data of the computing task to be transmitted from a terminal device to a result of the computing task being received.

[0161] Taking the two scenarios shown in (a) and (b) of FIG. 3 as examples.

[0162] If the computing node is a UPF, i.e. as shown in (a) of FIG. 3, the end-to-end latency of the service can include: the time for the data of the computing task to be sent from the terminal device to the UPF via the air interface, the time for the UPF to perform the computing task as a computing node, and the time for the UPF to send the output result to the terminal device. Among them, the time for the data of the computing task to be sent from the terminal device to the UPF is the uplink transmission latency, the time for the UPF to send the output result to the terminal device is the downlink transmission latency, and the time for the UPF to perform the computing task as a computing node is the processing latency. Since the processing latency of the computing node performing the processing task is also the latency of performing the computing task, the processing latency can also be referred to as the computing latency.

[0163] If the computing node is an AS, i.e. as shown in (b) of FIG. 3, the end-to-end latency of the service can include: the time for the data of the computing task to be sent from the terminal device to the AS via the network (including via the air interface and the UPF), the time for the AS to perform the computing task as a computing node, and the time for the AS to send the output result to the terminal device via the network (including via the UPF and the air interface). Among them, the time for the data of the computing task to be sent from the terminal device to the AS via the network is the uplink transmission latency, the time for the AS to send the output result to the terminal device via the mobile network is the downlink transmission latency, and the time for the AS to perform the computing task as a computing node is the processing latency, i.e. the computing latency.

[0164] The range of the computing latency of the service can be determined according to the end-to-end latency threshold of the service and the transmission latency. For example, the range of the computing latency of the service can be obtained by subtracting the range of the transmission latency from the end-to-end (E2E) latency threshold of the service. For example, the end-to-end latency threshold of the service is T E2E , the range of the transmission latency of the service is [T trans1 , T trans2 ], and then the range of the computing latency of the service can be [T E2E -T trans2 , T E2E -T trans1 ]. It can be understood that the smaller the computing latency, the greater the transmission latency left for the service, and therefore, the upper limit of the range of the computing latency of the service can be set, and the lower limit can not be set. For example, in this example, the range of the computing latency of the service can be less than or equal to (T E2E -T trans1 ).

[0165] Among them, the end-to-end latency threshold of the service can be obtained from the AS or the AF providing the service. The transmission latency of the service can be obtained by prediction through a prediction model.

[0166] In a possible implementation, the CMF can obtain a prediction model from a network data analysis function (NWDAF) in advance, and then use the prediction model to predict the transmission delay of a service when determining the computation delay requirement of the service.

[0167] FIG. 5 exemplarily shows a process of obtaining the prediction model.

[0168] As shown in the figure, the NWDAF can collect RAN node granularity air interface environment information and transmission delay information from RAN nodes in the network. The air interface environment information and the transmission delay information can be used as samples for model training to obtain the prediction model. The air interface environment information and the transmission delay information corresponding to the same RAN node can be used as a group of samples, where the air interface environment information can be used as a feature parameter and the transmission delay information can be used as a label. The samples corresponding to the RAN nodes are input into a pre-constructed model, and after parameter training of the prediction model, a trained prediction model can be obtained.

[0169] Exemplarily, the air interface environment information can include one or more of the following: a signal to noise ratio (SNR), a modulation coding scheme (MCS), an uplink transmission rate, a downlink transmission rate, a total number of resource blocks (RBs) of uplink channel resources, a total number of RBs of downlink channel resources, a RB retransmission rate, or a path loss (PL). The uplink transmission rate and the downlink transmission rate can be represented by parameters such as bit rate.

[0170] The NWDAF can send the trained prediction model to the CMF. It should be understood that one possible implementation of sending the prediction model to the CMF is to send parameters of the trained prediction model to the CMF, so that the CMF constructs the prediction model based on the received parameters.

[0171] It should be understood that the flow shown in FIG. 5 is one possible way of obtaining the prediction model, and should not constitute any limitation on the present application. The present application does not limit the specific implementation of obtaining the prediction model.

[0172] It should also be understood that the flow shown in FIG. 5 can be performed before the flow shown in FIG. 4, or before step 601 in FIG. 7 or step 1001 in FIG. 10 shown later. For brevity, this will not be repeated hereinafter.

[0173] In step 420, the CMF determines, according to the node information of the first computing node, a first computing model satisfying the computing latency requirement from a plurality of computing models supported by the first computing node, the first computing model being used to perform the first computing task of the service.

[0174] As mentioned above, a computing node can support one or more computing models. In one possible design, a computing node is a virtual machine or a container built on hardware resources, and a computing model can be a model loaded on the virtual machine or the container.

[0175] In this application, for the convenience of understanding and description, it is assumed that the first computing node is a computing node satisfying the computing latency requirement, and the first computing node can support a plurality of computing models. Wherein, the first computing node satisfying the computing latency requirement means that there is at least one computing model satisfying the computing latency requirement in the plurality of computing models supported by the first computing node. In other words, as long as there is at least one computing model satisfying the computing latency requirement on a certain computing node, it can be considered that the computing node satisfies the computing latency requirement.

[0176] The CMF can determine, according to the node information of the first computing node, a first computing model satisfying the computing latency requirement from a plurality of computing models supported by the first computing node, the first computing model being used to perform the first computing task of the service.

[0177] Exemplarily, the node information of the first computing node can indicate an identity of each computing model in the plurality of computing models supported by the first computing node, and a computing capability supported by each computing model in the plurality of computing models.

[0178] Wherein, the identity of each computing model can be used to identify a computing model. The identities of the computing models can be different from each other on a computing node, or in other words, the identities of the computing models are unique within a computing node.

[0179] The computing capability can be represented by a computing latency or a computing speed, which is not limited in this application. It can be understood that the smaller the computing latency is, the stronger the computing capability is. The faster the computing speed is, the stronger the computing capability is. The node information can indicate the computing capability supported by each computing model by the computing latency and / or the computing speed. It can be understood that the computing latency and the computing speed are inversely proportional in the case of a certain amount of data processing. Therefore, if the computing capability indicated by the node information is indicated by the computing speed, the computing latency can also be obtained based on the computing speed.

[0180] The computing capability supported by each computing model can be embodied by the correspondence between the identity (ID) of the computing model and one or more of the following: computing latency or computing speed. The identity of the computing model is used to distinguish different computing models on the same computing node.

[0181] Tables 2 and 3 below show two examples of the correspondence between the computing model and the computing capability. The correspondence shown in Table 2 is the correspondence between the identity of the computing model and the computing latency, and the correspondence shown in Table 3 is the correspondence between the identity of the computing model and the computing speed.

[0182] Table 2

[0183] Table 3

[0184] The model ID #1 to the model ID #5 in the table are respectively the identities of the computing models. It can be seen that the computing capabilities of different computing models on the same computing node are different.

[0185] It should be understood that the above indication of the computing capability supported by each computing model by means of the table is only one possible way, and other data structures can also be used in specific implementation, such as array, queue, container, stack, linear list, pointer, linked list, tree, graph, structure, class, heap, hash table, etc., without limitation.

[0186] Taking the correspondence between the computing model and the computing latency in Table 2 as an example. Assuming that the computing latency requirement is less than or equal to 40 ms, and the models 1, 2 and 3 in Table 2 meet the computing latency requirement, the computing node to which the computing model shown in Table 2 belongs also meets the computing latency requirement.

[0187] It can be seen that one or more computing models on a computing node can meet the computing latency requirement. In the case where multiple computing models on the first computing node meet the computing latency requirement, the CMF can select the computing model with the lowest computing latency from the multiple computing models meeting the computing latency requirement as the first computing model, so as to leave a certain margin when the computing resources of the first computing node change or the transmission latency of the service changes, thereby avoiding frequent model switching; or the CMF can select the computing model closest to the critical value of the computing latency requirement from the multiple computing models meeting the computing latency requirement as the first computing model according to the computing latency requirement, so as to maximize the resource utilization; or the CMF can randomly select one from the multiple computing models meeting the computing latency requirement as the first computing model. The present application does not limit this.

[0188] It should be understood that the identification of each computing model of the first computing node, and the computing capability supported by each computing model, can be obtained in a granularity of the computing node, or in a granularity of the computing model. In this case, the node information of the first computing node can be replaced by information in a granularity of the computing model, such as information of each computing model on the first computing node. The present application does not limit this.

[0189] In a possible design, the first computing node is one of the at least one computing node. Therefore, each computing node of the at least one computing node can support one or more computing models. Each computing node can be preloaded with a computing model, and obtain the identification of each computing model and the correspondence between each computing model and a computing capability.

[0190] Optionally, the node information of the first computing node further includes identification of the first computing node.

[0191] The identification of the first computing node can be used to indicate the first computing node. For example, the identification of the first computing node is address information of the first computing node, such as an IP address and a port number.

[0192] As described above, the first computing node can be one of the at least one computing node, and each computing node of the at least one computing node can be distinguished by a unique identification.

[0193] By including the identification of the first computing node in the node information of the first computing node, after the CMF selects the first computing node, the computing node that can provide computing services for the terminal device can be indicated by the identification of the first computing node, thereby facilitating subsequent data interaction between the terminal device and the first computing node for a computing task.

[0194] Optionally, the node information of the first computing node further includes location information of the first computing node.

[0195] The location information of the first computing node is used to indicate the geographic location of the first computing node. For example, the location information is a data network access identity (DNAI). The DNAI can be used to indicate an entry to access a data network, which is equivalent to providing the geographic location of the data network deployed in the network.

[0196] By including the location information of the first computing node in the node information of the first computing node, it is beneficial for the CMF to select a computing node that is relatively close to the geographic location of the terminal device to provide computing services.

[0197] Optionally, the node information of the first computing node further includes address information of the first computing node.

[0198] The address information of the first computing node is used to indicate a communication address of the first computing node. Exemplarily, the address information includes an IP address and a port number. The address information can be used for communication between other network elements, such as a CMF, a terminal device, and the like, and the first computing node. In an embodiment of the present application, the address information can also be used to establish a user plane between the terminal device and the first computing node.

[0199] In an implementation manner, the first computing node is an AS.

[0200] Since the AS itself can be used to run an application program to provide services for users, for example, an access interface of the application program can be provided for the users through a client to process access requests of the users. The AS as a computing node means that the application program itself can provide computing services. It can be understood that different ASs can run different application programs, and thus different ASs can support different services. In other words, the AS has a binding relationship with the service. For example, a certain AS can support one or more services. In an embodiment of the present application, it is assumed that the AS as the first computing node can support services accessed by the terminal device.

[0201] Exemplarily, the AS is an EAS deployed in a data network (DN), such as an EAS deployed in a multi-access edge computing (MEC) platform.

[0202] Optionally, the node information of the first computing node further includes one or more of the following: a fully qualified domain name (FQDN) of the AS, a maximum data volume of a service supported by the AS, a minimum data volume of the service supported by the AS, an execution period of the service supported by the AS, an end-to-end delay threshold of the service supported by the AS, a DNN corresponding to the service supported by the AS, or an S-NSSAI corresponding to the service supported by the AS.

[0203] The FQDN can be used for domain name system (DNS) resolution to obtain an IP address.

[0204] The DNN and / or S-NSSAI corresponding to the service supported by the AS can be used to identify the service. The DNN and / or S-NSSAI can be an identifier agreed by the AS and an operator in advance. The node information of the first computing node includes the DNN and / or S-NSSAI, which means that the node information of the first computing node indicates what service the AS supports as the first computing node.

[0205] The maximum data volume and the minimum data volume of the service supported by the AS can be used to describe the data volume size of the service supported by the AS, and are characteristic parameters of the service. The execution period of the service supported by the AS, that is, how often the service is executed, is another characteristic parameter used to describe the service. The maximum data volume, the minimum data volume and the execution period of the service supported by the AS can be regarded as the service model of the AS. It should be understood that the service model is different from the aforementioned calculation model or prediction model, and the service model aims to describe the characteristics of the service. The node information of the first computing node includes the maximum data volume, the minimum data volume and the execution period of the service, which means that the node information of the first computing node indicates the service model supported by the AS as the first computing node.

[0206] The end-to-end delay threshold of the service supported by the AS can be used to indicate the delay requirement of the service supported by the AS, and can be used to determine the aforementioned calculation delay requirement.

[0207] In another implementation manner, the first computing node is a core network element.

[0208] Unlike the AS, the core network element is generally not used to run an application program to provide service access, and therefore the core network element has no fixed binding relationship with the service. In the present application, the core network element can be used to provide a calculation service, and does not mean that the core network element runs an application program of a certain service.

[0209] Exemplarily, the first computing node is a UPF. In this case, the UPF not only has the data forwarding capability of the calculation task, but also has the processing capability of the data. The first computing node is a UPF, which can also be understood as that a traditional service processing server (used to implement the function of the computing node) and a UPF (used to implement the function of user plane forwarding) are combined in the same physical node.

[0210] In the case where the first computing node is a core network element, the related characteristics of the service can be obtained from the AF. The AF can send a request message to the CMF, which carries one or more of the following: DNN, S-NSSAI, maximum data volume of the service, minimum data volume of the service, execution period of the service or end-to-end delay threshold of the service.

[0211] In step 430, the CMF sends the identifier of the first calculation model to the first computing node.

[0212] The CMF can send the identifier of the first calculation model to the first computing node after determining the first calculation model that meets the calculation delay requirement. In this way, the first computing node can execute the first calculation task of the service by using the received first calculation model. Specifically, the first computing node determines to execute the first calculation task by using the first calculation model according to the identifier of the first calculation model.

[0213] Optionally, the method further comprises: the first computing node sending an acknowledgement (ACK) to the CMF.

[0214] The ACK can be regarded as a response of the first computing node to the identification of the first computing model sent by the CMF.

[0215] The first computing node can send the ACK to the CMF after confirming that the first computing model can perform the first computing task, such as after completing the loading of the first computing model, or after confirming that the load of the first computing model is small, or after confirming that the computing capability of the first computing model has not changed or has changed little compared with the computing capability in the node information sent previously.

[0216] Thereafter, the first computing node can perform the first computing task after receiving the data of the first computing task from the terminal device.

[0217] Based on the above technical solution, the CMF can find a first computing model that meets the computing delay requirement of the service according to the node information of the first computing node and the computing delay requirement of the service, and send the identification of the first computing model to the first computing node, so that the first computing node adopts the first computing model to perform the first computing task of the service. In this way, the processing delay of the service is controlled, thereby facilitating the control of the overall delay of the service to meet the delay requirement of the service.

[0218] It should be understood that the CMF can perform the above steps 410 to 430 after receiving the session establishment request of the terminal device and determining the service requested to be accessed by the terminal device, and then determine the computing model for performing the computing task of the service for the terminal device. Since the same computing node can provide computing services for multiple terminal devices at the same time, after determining the first computing model for the terminal device, the CMF can send the identification of the terminal device to the first computing node at the same time as sending the identification of the first computing model, so as to bind the first computing task of the terminal device to the first computing model.

[0219] Optionally, the method further comprises: the CMF sending the identification of the terminal device to the first computing node.

[0220] The identification of the terminal device can include address information of the terminal device.

[0221] In a possible design, the identification of the first computing model and the identification of the terminal device are carried in the same signaling.

[0222] In the case that the first computing node is a core network element (such as a UPF), the present application further provides another implementation to bind the first computing task with the first computing model. Since the terminal device requests to access a service through a session establishment request, the PDU session ID can be carried in the session establishment request. Therefore, the CMF can send the PDU session ID to the first computing node at the same time when sending the identifier of the first computing model, so as to bind the PDU session of the terminal device with the first computing model. In a possible design, the identifier of the first computing model and the PDU session ID are carried in the same signaling.

[0223] Optionally, the method further includes: the CMF sending address information of the first computing node to the terminal device.

[0224] The address information of the first computing node can be used to identify the location of the first computing node in the network. Based on the address information of the first computing node, data can be routed to the first computing node. Exemplarily, the address information includes an internet protocol (IP) address and a port number.

[0225] In a possible design, the address information of the first computing node is carried in a session establishment response. In other words, the CMF indicates the address information of the first computing node to the terminal device through the session establishment response. The session establishment response is a response message to the session establishment request of the terminal device. Exemplarily, as a response to the computing task session establishment of the terminal device, the session establishment response is a computing task session establishment response.

[0226] Based on the address information of the first computing node, the terminal device can send uplink data of the first computing task. For example, the terminal device can carry the IP address and the port number of the first computing node in the packet header of a data packet, and the payload of the data packet can be the uplink data of the first computing task.

[0227] Optionally, the method 400 further includes step 440: obtaining node information of the first computing node.

[0228] As described above, the computing node can be a core network element or an AS. Therefore, the first computing node in the present application can be a core network element or an AS. The specific implementation of step 440 is described below respectively for the two cases.

[0229] Case 1: The first computing node is a core network element.

[0230] The core network element as the computing node can send node information of the computing node to the NRF after the computing node (or the function of the computing node) is online. Therefore, the NRF can store the node information of each computing node in the network.

[0231] After determining the computing latency requirement, the CMF can send a discovery request to the NRF, and the discovery request can carry information of the computing latency requirement. The NRF can determine that the first computing node meets the computing latency requirement in response to the discovery request, and then send node information of the first computing node to the CMF.

[0232] Case two, the first computing node is an AS:

[0233] Since the AS can interact with the core network element through the NEF, the AS as the computing node can send node information of the computing node to the NEF after the computing node (or the function of the computing node) is online. The NEF can save the node information of the computing node from each AS locally, or send the node information to the UDR for saving.

[0234] After determining the computing latency requirement, the CMF can send a subscription request to the NEF, and the subscription request can be used to request node information of each computing node. The NEF sends the node information of each computing node to the CMF in response to the subscription request. The CMF can determine that the first computing node meets the computing latency requirement according to the node information of each computing node, and thus obtain the node information of the first computing node.

[0235] It can be understood that, regardless of whether the first computing node is a core network element or an AS, the node information of the first computing node obtained by the CMF is sent by the first computing node. The figure only shows an example in which the node information of the first computing node reaches the CMF from the first computing node. In the case where the first computing node is a core network element, a more detailed process in which the CMF obtains the node information of the first computing node can be referred to the description in steps 607 to 609 of method 600 below. In the case where the first computing node is an AS, a more detailed process in which the CMF obtains the node information of the first computing node can be referred to the description in steps 1002 to 1003 and 1008 to 1009 of method 1000 below, which will not be described here in detail. It should be understood that, when referring to method 600, UPF#1 can be replaced with the first computing node; when referring to method 1000, EAS#1 can be replaced with the first computing node.

[0236] Optionally, the method 400 further includes: step 450, the CMF determines that the first computing model no longer meets the computing latency requirement.

[0237] In a possible case, the transmission delay of the service changes due to changes in the environment of the transmission link, so that the calculated delay requirement determined according to the end-to-end delay threshold of the service and the transmission delay also changes. For example, the transmission delay of the service becomes larger, and then the calculated delay requirement of the service can be correspondingly reduced, which makes the first calculation model that originally meets the calculated delay requirement no longer meet the changed calculated delay requirement.

[0238] Exemplarily, in the process of executing the first calculation task, the CMF can predict the downlink transmission delay through the prediction model, and can collect the uplink transmission delay of the first calculation node statistics from the first calculation node. In this way, the CMF can learn the change of the transmission delay of the service in time, and then adjust the calculated delay requirement in time when the transmission delay changes. For example, the CMF can adjust the calculated delay requirement when the change amount of the transmission delay of the service exceeds a preset threshold. After adjusting the calculated delay requirement, the CMF can determine whether the first calculation model meets the adjusted calculated delay requirement according to the calculation capability of the first calculation model.

[0239] In another possible case, since the first calculation model can be used to execute multiple calculation tasks at the same time, the occupancy rate of the resource is high, the load is large, and the calculation capability is affected, so that the calculation capability no longer meets the calculated delay requirement of the service.

[0240] Exemplarily, the CMF can subscribe to the changes of the node information of the first calculation node to the NRF. The first calculation node can send the changed node information of the first calculation node to the CMF when the calculation capability of the first calculation model changes. The CMF can obtain the calculation capability of the first calculation model from the changed node information of the first calculation node, and then determine whether the calculation capability of the first calculation model meets the calculated delay requirement of the service.

[0241] More details of step 450 can be found in steps 618-620 in flow 600A (corresponding to the case that the computing latency requirement changes) shown in FIG. 8, steps 627-630 in flow 600B (corresponding to the case that the computing capability of the first computing model changes) shown in FIG. 9, and steps 1018-1021 in flow 1000A (corresponding to the case that the computing latency requirement changes) shown in FIG. 11 and steps 1029-1031 in flow 1000B (corresponding to the case that the computing capability of the first computing model changes) shown in FIG. 12, which will not be described herein. It should be understood that when step 450 is understood in combination with FIG. 8 and FIG. 9, UPF#1 can be replaced by the first computing node, the node information of UPF#1 can be replaced by the node information of the first computing node, and the changed node information #1 of UPF#1 can be replaced by the changed node information of the first computing node. When step 450 is understood in combination with FIG. 11 and FIG. 12, EAS#1 can be replaced by the first computing node. The node information of EAS#1 can be replaced by the node information of the first computing node, and the changed node information #1 of EAS#1 can be replaced by the changed node information of the first computing node.

[0242] It should be understood that the two cases listed above can exist alone or simultaneously, which is not limited in the present application.

[0243] Optionally, the method 400 further includes: step 460, the CMF determines a second computing model satisfying the computing latency requirement from the plurality of computing models, the second computing model being used to execute a second computing task of the service.

[0244] Since the service can include one or more computing tasks, the one or more computing tasks should all satisfy the latency requirement of the service, and thus the one or more computing tasks should all satisfy the computing latency requirement determined based on the latency requirement of the service.

[0245] The CMF can determine a second computing model satisfying the computing latency requirement from the plurality of computing models of the first computing node in a case that the first computing model no longer satisfies the computing latency requirement. It can be understood that since the second computing model satisfies the computing latency requirement, the computing capability supported by the second computing model is different from the computing capability supported by the first computing model, for example, the computing latency supported by the second computing model is less than the computing latency supported by the first computing model.

[0246] It can be understood that the CMF determines the second computing model to be used to execute the second computing task, which means that the computing model used to process the service of the terminal device is switched. Since the second computing model is a computing model supported by the first computing node, the model switching is a model switching on the same computing node.

[0247] Optionally, the method 400 further comprises: step 470, the CMF sends an identification of the second computing model to the first computing node.

[0248] The CMF can send the identification of the second computing model to the first computing node after determining that the second computing model meets the computing latency requirement. In this way, the first computing node can determine to execute the second computing task by using the second computing model according to the identification of the second computing model.

[0249] One possible implementation of the step 470 is that the CMF sends a switching indication to the first computing node, and the switching indication can be used to explicitly instruct the first computing node to switch the computing model. Another possible implementation of the step 470 is that the CMF sends the identification of the second computing model to the first computing node while sending the identification of the terminal device or the PDU session identification, for example, carrying the identification of the second computing model in the same signaling, and optionally, carrying at least one of the identification of the terminal device or the PDU session identification, thereby the first computing node can infer that the second computing model is used to execute another computing task of the terminal device or the PDU session, thereby implicitly instructing the first computing node to switch the computing model.

[0250] After that, the CMF can execute the second computing task after receiving the data of the second computing task from the terminal device.

[0251] After determining to execute the second computing task by using the second computing model, the first computing node can continue to execute the first computing task by using the first computing model and send the result of the first computing task to the terminal device, or can stop executing the first computing task by using the first computing model. The present application does not limit this.

[0252] For example, the first computing node can distinguish the data of the first computing task and the data of the second computing task according to the data packets received from the terminal device and the execution period of the service.

[0253] For example, the data received from the terminal device in the period 1 is the data of the first computing task, and the data received from the terminal device in the period 2 is the data of the second computing task.

[0254] For another example, the first computing node can also distinguish the first computing task and the second computing task according to the received data packets. Taking image recognition as an example, the terminal device can deliver N pictures obtained by the terminal device to the computing node through a data packet for image recognition. The N pictures in the data packet are the data of a computing task.

[0255] It should be understood that the first computing task and the second computing task do not affect each other and have no dependency relationship.

[0256] Optionally, the address information of the first computing node for the first computing task is the same as the address information of the second computing node for the second computing task.

[0257] The address information of the computing node for a certain computing task can be understood as the address information allocated to the terminal device for accessing the computing node. That is, the terminal device can interact with the computing node for data of the certain computing task based on the address information.

[0258] In the embodiments of the present application, the address information of the first computing node for the first computing task, i.e., the address information for the terminal device to interact with the first computing node for data of the first computing task, is the same as the address information of the first computing node for the second computing task, i.e., the address information for the terminal device to interact with the first computing node for data of the first computing task. The two are the same, which means that no matter whether the first computing task or the second computing task is executed, the address information used by the first computing node to interact with the terminal device for data of the computing task does not change, and the terminal device can interact with the first computing node for data of the service based on the same address information.

[0259] From the perspective of the first computing node, the address information when the first computing model executes the first computing task is the same as the address information when the second computing model executes the second computing task. The first computing model and the second computing model can share the same transport layer connection, and even if the computing model is switched, the current network connection can still be used. In other words, when the terminal device accesses the service, the address of the first computing node accessed by the terminal device does not change, and therefore the terminal device does not perceive the change of the computing model (or computing resource). Therefore, service migration caused by changes in computing resources can be avoided, service interruption can be avoided, and user experience can be improved. Especially for services that have strict requirements on time delay stability, better service experience can be obtained.

[0260] In order to better understand the method provided above, the above process is simply illustrated below in combination with two different scenarios in (a) and (b) in FIG. 3.

[0261] The scenario shown in (a) of FIG. 6 corresponds to (a) of FIG. 3, and shows a case where the computing node is a UPF. The UPF as the computing node can report the CMF with the identities of the computing models supported by the UPF and the correspondence between the computing models and the computing capabilities. When the terminal device has a demand for accessing the target recognition service, the terminal device initiates a session establishment request to the CMF through the RAN. The CMF selects, based on the session establishment response, a first computing model that meets the computing latency requirement from the multiple computing models supported by the UPF for the terminal device, and sends the identity of the first computing model that meets the computing latency requirement to the UPF. The terminal device can perform data interaction of a first computing task with the UPF via the RAN. The CMF reselects, in a case where the first computing model no longer meets the computing latency requirement, a second computing model that meets the computing latency requirement from the multiple computing models supported by the UPF for the terminal device, and sends the identity of the second computing model to the UPF. The terminal device can perform data interaction of a second computing task with the UPF via the RAN.

[0262] The scenario shown in (b) of FIG. 6 corresponds to (b) of FIG. 3, and shows a case where the computing node is an EAS. The EAS as the computing node can report the CMF with the identities of the computing models supported by the EAS and the correspondence between the computing models and the computing capabilities. When the terminal device has a demand for accessing the target recognition service, the terminal device initiates a session establishment request to the CMF through the RAN. The CMF selects, based on the session establishment response, a first computing model that meets the computing latency requirement from the multiple computing models supported by the EAS for the terminal device, and sends the identity of the first computing model that meets the computing latency requirement to the EAS. The terminal device can perform data interaction of a first computing task with the EAS via the RAN and the UPF. The CMF reselects, in a case where the first computing model no longer meets the computing latency requirement, a second computing model that meets the computing latency requirement from the multiple computing models supported by the EAS for the terminal device, and sends the identity of the second computing model to the EAS. The terminal device can perform data interaction of a second computing task with the EAS via the RAN and the UPF.

[0263] The specific procedures of the first computing node in the case of core network element and application server will be described below in combination with FIG. 7 to FIG. 12. FIG. 7 to FIG. 9 and FIG. 10 to FIG. 12 are based on the following scenarios respectively: the CMF obtains the prediction model in advance, for example, by the procedure shown in FIG. 5. The prediction model can be used to predict the transmission delay of each service. Multiple computing nodes are deployed in the network, and two computing nodes are taken as an example in the description. The computing nodes in FIG. 7 to FIG. 9 are UPFs (i.e., an example of core network elements), and the two computing nodes are UPF#1 and UPF#2 respectively. The computing nodes in FIG. 10 to FIG. 12 are EASs (i.e., an example of ASs), and the two computing nodes are EAS#1 and EAS#2 respectively, wherein the computing node with the number #1 is an example of the first computing node.

[0264] As an example, FIG. 7 is a schematic flowchart of a communication method provided by another embodiment of the present application. The method 600 shown in FIG. 7 includes steps 601 to 617. Each step in FIG. 7 will be described in detail below.

[0265] In step 601, each UPF sends node information to the NRF. Accordingly, the NRF receives the node information from each UPF.

[0266] Each UPF can load one or more computing models locally, and obtain the identification of each computing model and the computing capability supported by each computing model. After each UPF goes online as a computing node, it can actively send its own node information to the NRF to open its capability information to the NRF. The NRF can store the node information of each computing node locally.

[0267] Two computing nodes, UPF#1 and UPF#2, are shown in this embodiment. Correspondingly, step 601 can include: step 6011, UPF#1 sends node information #1 to the NRF; and step 6012, UPF#2 sends node information #2 to the NRF.

[0268] The node information #i (i can be 1 or 2) is the node information of UPF#i, which can indicate the identification of each computing model in one or more computing models supported by the UPF#i, and the computing capability supported by each computing model in the multiple computing models.

[0269] Optionally, the node information #i further includes the identification of the UPF#i (i.e., the identification of the computing node). By providing the identification of each UPF, each computing node and its node information can be distinguished.

[0270] Optionally, the node information #i further comprises location information of the UPF #i (i.e., location information of the computing node), such as DNAI. By providing the location information of the UPF, the geographical location of the computing node in the network can be provided to the NRF.

[0271] Optionally, the node information #i further comprises address information of the UPF #i (i.e., address information of the computing node), such as IP address and port number. By providing the address information of each UPF, the CMF can send information to each computing node, such as sending the identification of the first computing model to the UPF #1.

[0272] Exemplarily, the node information of each UPF can be carried in a registration message. For example, the registration message can be specifically an NF management NF registration (Nnrf_NFManagement_NFResgister) message of the NRF interface.

[0273] In step 602, the AF sends a request message to the PCF. Accordingly, the PCF receives the request message from the AF.

[0274] Exemplarily, the AF can send the request message to the PCF through the NEF. The request message can carry one or more of the following: DNN, S-NSSAI, service model, or end-to-end delay threshold of the service. Among them, the DNN and / or S-NSSAI correspond to the service, which can be understood as the service supported by the AF, or in other words, the service that can be provided. The service model and the end-to-end delay threshold of the service are related to the computing task of the service, and can be used to determine the computing delay requirement of the service. For more detailed description of each parameter, please refer to the related description in step 420 above, which will not be repeated here.

[0275] In step 603, the AF performs the AF requests to influence traffic routing procedure.

[0276] It should be noted that in the agreement of the third generation partnership project (3 rd generation partnership project, 3GPP), it is stipulated that the AF can send a request to the NFs of the 5G core network (5G core network, 5GC) to influence the traffic routing decision of the PDU session. The essence of this procedure is that the AF influences the selection or reselection of the UPF by the SMF, and controls the user traffic to be routed to the target data network (data network, DN) identified by the DNAI.

[0277] If the AF invokes the Nnef_TrafficInfluence_Create service operation of the NEF interface or the Nnef_TrafficInfect_Update service operation of the NEF interface, the NEF can store the parameters carried in the request message from the AF in the UDR. The UDR notifies the PCF of the parameters of the information relates to AF traffic influence request information for traffic routing through the Nudr_DM_Notify service operation of the UDR interface data management (DM).

[0278] In this embodiment, through the foregoing procedure, the service model and / or the end-to-end delay threshold carried in the request message sent by the AF are sent to the PCF. The PCF can obtain the policy information of the service based on the parameters carried in the request message from the AF, and the policy information can be used to determine the end-to-end delay threshold and the transmission delay of the service. Exemplarily, the service model can be used as an input of a prediction model to predict the transmission delay.

[0279] In step 604, the terminal device sends a first session establishment request to the CMF. Correspondingly, the CMF receives the first session establishment request from the terminal device.

[0280] When the terminal device has a demand to access a service, the terminal device can initiate access to the service through the first session establishment request. Exemplarily, the first session establishment request is a computing task session establishment request, which can be used to request access to the service and perform a computing task.

[0281] Optionally, the first session establishment request carries a DNN and / or an S-NSSAI. As described above, the DNN and / or the S-NSSAI correspond to the service. Based on the DNN and / or the S-NSSAI, the CMF can determine the service requested to be accessed by the terminal device.

[0282] Optionally, the first session establishment request carries a PDU session identifier. The PDU session identifier can be used to identify a PDU session. In this embodiment, the session of the same service can correspond to one PDU session identifier, and the data (i.e., user plane data) interaction of the computing task between the terminal device and the computing node can be carried through the PDU session.

[0283] Exemplarily, the terminal device can send the first session establishment request to the CMF in one of the following multiple manners.

[0284] For example, the terminal device can send the first session establishment request to the AMF through the access network device; the AMF can select the SMF according to the DNN and / or S-NSSAI included in the first session establishment request after receiving the first session establishment request, and send the first session establishment request to the selected SMF; the SMF can select the CMF according to the DNN and / or S-NSSAI included in the first session establishment request after receiving the first session establishment request, and send the first session establishment request to the CMF.

[0285] For another example, the terminal device can send the first session establishment request to the AMF through the access network device; the AMF can select the CMF according to the DNN and / or S-NSSAI included in the first session establishment request after receiving the first session establishment request, and send the first session establishment request to the CMF.

[0286] For another example, the terminal device can also send the first session establishment request to the CMF directly through the access network device.

[0287] In addition, after receiving the first session establishment request sent by the terminal device, the access network device can send a message to the CMF, such as sending via the AMF and the SMF, such as sending via the AMF, and such as sending directly to the CMF. The message contains the location information of the terminal device and the first session establishment request. The location information may, for example, be a tracking area identity (TAI).

[0288] Embodiments of the present application do not limit the way in which the terminal device sends the first session establishment request to the CMF.

[0289] Corresponding to the various possible ways in which the terminal device sends the first session establishment request to the CMF, the CMF can also obtain the location information of the terminal device. For example, the CMF can obtain the location information of the terminal device from the access network device, or from the AMF, or from the SMF, without limitation.

[0290] In a possible design, the first session establishment request is a computing task session establishment request, which can be used to request the establishment of a transmission link between the terminal device and the computing node for transmitting user plane data.

[0291] In step 605, the CMF requests the PCF to obtain the policy information of the service according to the first session establishment request. Correspondingly, the PCF receives the request for the policy information of the service from the CMF.

[0292] The CMF can determine, according to the first session establishment request, that the terminal device accesses a service, and then request the PCF to obtain policy information of the service.

[0293] Exemplarily, the CMF can determine, according to the DNN and / or S-NSSAI carried in the first session establishment request, that the terminal device accesses a service, and then request the PCF to obtain policy information of the service.

[0294] In step 606, the PCF delivers the policy information of the service to the CMF. Correspondingly, the CMF receives the policy information of the service from the PCF.

[0295] In another implementation, the PCF can obtain, after receiving the request of the CMF for the policy information of the service, the service model of the service and the end-to-end delay threshold of the service from the AF, and then obtain the policy information of the service. In other words, steps 602-604 above can be executed before step 605, or can be executed after step 605 and before step 606. The present application does not limit this.

[0296] In step 607, the CMF determines the computing delay requirement of the service.

[0297] The CMF can determine the computing delay requirement of the service according to the policy information of the service. The determination of the computing delay requirement of the service can refer to the detailed description in step 410 of method 400 above, and will not be repeated here.

[0298] In step 608, the CMF sends a discovery request to the NRF, and the discovery request carries information of the computing delay requirement. Correspondingly, the NRF receives the discovery request from the CMF.

[0299] The information of the computing delay requirement can be used to determine the computing delay requirement. For example, the computing delay requirement of the service is [T E2E -T trans2 , T E2E -T trans1 ], and the information of the computing delay requirement can be two critical values of the range, i.e., the upper limit (T E2E -T trans2 ) and the lower limit (T E2E -T trans1 ), or can be the upper limit (T E2E -T trans1 ) of the range, or can be the upper limit (T E2E -T trans1 ) of the range and the difference (T trans1 -T trans2 ) or (T trans2 -T trans1 ) between the upper limit and the lower limit. And so on, which will not be listed here.

[0300] The information of the computing latency carried in the discovery request can also be understood as that the discovery request is used to request to discover the computing node satisfying the computing latency requirement.

[0301] In step 609, the NRF sends a discovery response to the CMF, and the discovery response carries the node information #1 of the UPF #1. Accordingly, the CMF receives the discovery response from the NRF.

[0302] The CMF can send the information of the computing latency requirement to the NRF through the discovery request, so that the NRF determines the computing node satisfying the computing latency requirement based on the information. It should be understood that the understanding of the computing node satisfying the computing latency requirement can refer to the related description in step 420 of the method 400, which will not be repeated here.

[0303] In this embodiment, the NRF can determine the UPF satisfying the computing latency requirement according to the node information received from each UPF in response to the discovery request, and the UPF #1 is an example of the first computing node. The NRF can send the node information #1 of the UPF #1 to the CMF. That is, as a response to the discovery request, the NRF implicitly informs the CMF that the UPF #1 satisfies the computing latency requirement through the node information #1.

[0304] Optionally, the discovery request further carries the location information of the terminal device.

[0305] Based on the location information of the terminal device, the NRF can select the computing node closest to the terminal device from the computing nodes satisfying the computing latency requirement according to the location information of the terminal device and the location information (such as DNAI) of each computing node. In other words, the priority of the computing latency requirement is higher than that of the geographical location. The NRF can preferentially select the computing node satisfying the computing latency requirement, and on this basis, the first computing node can be further selected in combination with the location information of the terminal device.

[0306] For example, the computing nodes satisfying the computing latency requirement include the UPF #1 deployed in city 1 and the UPF #2 deployed in city 2, and since the geographical location of the terminal device is also in city 1, or city 1 is closer to city 1 than city 2, the UPF #1 can be determined as the first computing node.

[0307] For example, the computing node includes UPF#1 deployed in city 1 and UPF#2 deployed in city 2. Among them, UPF#1 meets the computing latency requirement, and UPF#2 does not meet the computing node requirement. The geographical location of the terminal device is in city 2, but since the computing node deployed in city #2 does not meet the computing latency requirement, the UPF#1 far away is still selected as the first computing node.

[0308] It should be understood that the above two examples are only for the convenience of understanding, and different assumptions are made for UPF#1 and UPF#2 respectively, but this should not constitute any limitation on the present application.

[0309] Optionally, the discovery request also carries DNN and / or S-NSSAI.

[0310] Since the policy information of multiple services can be stored in the NRF, the CMF can carry the DNN and / or S-NSSAI in the discovery request to facilitate the NRF to determine which service policy information the CMF requests to obtain.

[0311] Exemplarily, the above discovery request is the NF discovery request (Nnrf_NFDiscovery_Request) of the NRF interface, and the above discovery response is the NF discovery response (Nnrf_NFDiscovery_Response) of the NRF interface.

[0312] In step 610, the CMF determines the first computing model meeting the computing latency requirement according to the node information #1 of the UPF#1.

[0313] After receiving the node information #1 of the UPF#1, the CMF can determine the first computing model meeting the computing latency requirement from the multiple computing models supported by the UPF#1 according to the computing capabilities of the computing models indicated in the node information #1 of the UPF#1.

[0314] For more detailed description of the CMF determining the first computing model meeting the computing latency requirement from the multiple computing models, please refer to the related description in step 420 of method 400, which will not be repeated here.

[0315] In step 611, the CMF sends a second session establishment request to the UPF#1, carrying the identification of the first computing model. Accordingly, the UPF#1 receives the identification of the first computing model from the CMF.

[0316] The CMF can send the identification of the first computing model to the UPF#1 according to the identification of the first computing model in the node information #1 of the UPF#1 after determining the first computing model.

[0317] In a possible design, the CMF can send the identification of the first computing model to the UPF#1 through a second session establishment request. In other words, the identification of the first computing model is carried in the second session establishment request. The second session establishment request can be regarded as a session establishment request sent by the CMF to the UPF#1 in response to the first session establishment request received from the terminal device. Exemplarily, the second session establishment request is a computing task session establishment request.

[0318] There can be different ways for the CMF to send the second session establishment request to the UPF#1. For example, the CMF can send the second session establishment request to the SMF, and the SMF sends the second session establishment request to the UPF#1. Alternatively, the CMF can send the second session establishment request to the UPF#1 directly without going through other network elements.

[0319] In step 612, the UPF#1 sends a second session establishment response to the CMF. Accordingly, the CMF receives the second session establishment response from the UPF#1.

[0320] The UPF#1 can return the second session establishment response after receiving the identification of the first computing model. For example, the UPF#1 can return the second session establishment response in a case where it is confirmed that the first computing model can perform the computing task.

[0321] Optionally, the UPF#1 also sends address information of the UPF#1, such as an IP address and a port number of the UPF#1, to the CMF. The address information is used to establish a user plane of an N3 tunnel between the RAN and the UPF#1.

[0322] Optionally, the address information of the first computing model can be carried in the ACK.

[0323] In a possible design, the UPF#1 can return an ACK to the CMF through the second session establishment response. Optionally, the UPF#1 returns address information to the CMF through the second session establishment response. Optionally, the UPF#1 returns N3 tunnel information through the second session establishment response. The N3 tunnel information can include tunnel identification information allocated by the UPF#1 and can be used to establish a user plane between the RAN and the UPF#1. It can be understood that the user plane can be used to transmit data of the computing task between the terminal device and the UPF#1.

[0324] Exemplarily, the second session establishment response is a computing task session establishment response as a response to the computing task session establishment request.

[0325] The UPF#1 can send the second session establishment response to the CMF in different ways. For example, the UPF#1 can send the second session establishment response to the SMF, and the SMF sends the second session establishment response to the CMF. Alternatively, the UPF#1 can send the second session establishment response to the CMF directly without the mediation of other network elements.

[0326] In step 613, the CMF sends the first session establishment response to the terminal device. Correspondingly, the terminal device receives the first session establishment response from the CMF.

[0327] As a response to the first session establishment request received from the terminal device, the CMF can send the first session establishment response to the terminal device after receiving the second session establishment response from the UPF#1. Optionally, the first session establishment response can carry the address information of the UPF#1. Optionally, the first session establishment response can carry the N3 tunnel information.

[0328] In step 614, the CMF establishes the user plane between the RAN and the UPF#1.

[0329] As described above, the user plane can be an N3 tunnel. The specific process of establishing the N3 tunnel can refer to the existing session establishment process, which will not be described in detail herein. The terminal device can establish the N3 tunnel with the UPF#1 based on the N3 tunnel information received in the first session establishment response. At this point, the end-to-end transmission link for transmitting user plane data between the terminal device and the UPF#1 is established.

[0330] It should be understood that the first session establishment request and the first session establishment response, and the second session establishment request and the second session establishment response exemplified above are named to distinguish the messages transmitted between different network elements, wherein the first session establishment request and the first session establishment response are transmitted between the terminal device and the CMF, and the second session establishment request and the second session establishment response are transmitted between the CMF and the UPF#1.

[0331] In step 615, the terminal device sends the data of the first computing task to the UPF#1.

[0332] The terminal device can send the data of the first computing task to the UPF#1 through the user plane. It can be understood that the data of the first computing task sent by the terminal device to the UPF#1 is uplink data.

[0333] In step 616, the UPF#1 executes the first computing task by using the first computing model.

[0334] The UPF#1 can input the data of the first computing task into the first computing model to execute the first computing task through the first computing model and obtain a computing result of the first computing task when receiving the data of the first computing task from the terminal device.

[0335] In step 617, the UPF#1 sends the computing result of the first computing task to the terminal device.

[0336] The UPF#1 can send the computing result of the first computing task to the terminal device through a user plane. It can be understood that the computing result sent by the UPF#1 to the terminal device is downlink data.

[0337] Based on the above process, the terminal device can interact with the UPF#1 for data of a computing task. Since the UPF#1 and the first computing model used to execute the first computing task are both determined by the CMF according to the computing delay requirement of the service, when the first computing model is used to execute the first computing task, the processing delay of the service meets the computing delay requirement. In this way, the processing delay of the service is controlled, thereby facilitating the control of the overall delay of the service to meet the delay requirement of the service.

[0338] Since the environment of the transmission link changes and the load of the computing model also changes, the first computing model may not always meet the aforementioned computing delay requirement. The CMF can continuously monitor during the execution of the first computing task, so as to determine a new computing model for the service of the terminal device when the first computing model no longer meets the computing delay requirement, and then instruct the UPF#1 to perform model switching.

[0339] For the convenience of understanding, the process of model switching will be described below in combination with the processes shown in FIG. 8 and FIG. 9. It should be understood that FIG. 8 and FIG. 9 respectively show two different cases in which the first computing model no longer meets the computing delay requirement, and for the convenience of distinguishing and describing, the process shown in FIG. 8 is denoted as 600A, and the process shown in FIG. 9 is denoted as 600B. The process shown in FIG. 7 can be executed alone, or can be combined with the processes shown in FIG. 8 and FIG. 9. For the sake of brevity, the same steps in the processes shown in FIG. 8 and FIG. 9 as in FIG. 7 will not be described again, and the description of the related steps in the foregoing can be referred to.

[0340] Referring to FIG. 8, the process 600A shown in FIG. 8 can be executed after step 615 of the foregoing method 600, or can be executed after step 617, which is not limited in the present application. In the figure, only the case after step 617 is shown, but this should not constitute any limitation on the present application. The execution order of each step should be determined according to the logical order between steps.

[0341] The steps in FIG. 8 will be described in detail below.

[0342] In step 618, the UPF#1 reports the uplink transmission latency to the CMF. Accordingly, the CMF receives the uplink transmission latency from the UPF#1.

[0343] The UPF#1 can count the uplink transmission latency and send the uplink transmission latency to the CMF. For example, the UPF#1 can perform quality of service (QoS) monitoring to obtain the uplink latency between the terminal device and the UPF#1. Alternatively, the UPF#1 can periodically report the uplink transmission latency to the CMF.

[0344] It should be understood that step 618 can be periodically performed, and although not shown in the figure, it can be understood that step 618 can be performed multiple times, such as continuously in the flow 600A shown in FIG. 8.

[0345] In step 619, the CMF determines that the computing latency requirement of the service has changed.

[0346] In step 607, it is mentioned that the CMF can determine the computing latency requirement of the service according to the end-to-end latency threshold of the service and the transmission latency, which can be predicted by the prediction model. This is because the UPF#1 cannot count the uplink transmission latency before the terminal device and the UPF#1 start data interaction for the computing task, so the transmission latency can be obtained by prediction. However, after the terminal device and the UPF#1 start data interaction for the computing task, the UPF#1 can count the uplink transmission latency, and the counted uplink transmission latency is more accurate than the predicted uplink transmission latency, so the CMF can determine the transmission latency of the service according to the predicted downlink transmission latency and the uplink transmission latency received from the UPF#1.

[0347] The CMF can update the transmission latency of the service according to the uplink transmission latency reported by the UPF#1, and further determine whether the computing latency requirement of the service has changed. For example, the CMF can determine the computing latency requirement of the service according to the uplink transmission latency received most recently after each time the CMF receives the uplink transmission latency reported by the UPF#1. In other words, the CMF can periodically determine the computing latency requirement of the service. For another example, the CMF can compare the uplink transmission latency received most recently with the uplink transmission latency used last time to determine the computing latency requirement after each time the CMF receives the uplink transmission latency reported by the UPF#1, and in the case of large change, the CMF can determine the computing latency requirement of the service again.

[0348] It can be understood that the computing latency requirement of the CMF re-determination may have changed from the previous one, and for the convenience of distinguishing and description, the computing latency requirement of the CMF re-determination is referred to as the changed computing latency requirement.

[0349] In step 620, the CMF determines that the first computing model no longer meets the changed computing latency requirement.

[0350] For the convenience of understanding, the embodiment assumes that the computing capability of the first computing model has not changed. However, due to the change of the computing latency requirement, the computing capability of the first computing model may no longer meet the changed computing latency requirement.

[0351] For example, in step 607, the CMF determines that the computing latency requirement of the service is less than or equal to 40 ms; the node information #1 of the UPF #1 contains the correspondence shown in Table 2, and in step 610, the CMF determines the first computing model as model 4 in Table 2 (i.e., an example of the first computing model), which can meet the computing latency requirement of less than or equal to 40 ms. However, in step 619, the CMF re-determines the changed computing latency requirement as less than or equal to 30 ms, at this time, model 4 no longer meets the computing latency requirement.

[0352] In step 621, the CMF determines a second computing model that meets the computing latency requirement from a plurality of computing models supported by the UPF #1.

[0353] It can be understood that the computing latency requirement mentioned in step 620 is the changed computing latency requirement, for example, less than or equal to 30 ms in the above example. The CMF can determine a second computing model that meets the computing latency requirement from a plurality of computing models supported by the UPF #1 according to the node information #1 of the UPF #1.

[0354] For example, in step 619, the CMF re-determines the changed computing latency requirement as less than or equal to 30 ms, and the node information #1 of the UPF #1 contains the correspondence shown in Table 2, then the CMF can determine the second computing model as model 1 in Table 2 (i.e., an example of the second computing model).

[0355] In step 622, the CMF sends an identifier of the second computing model to the UPF #1. Correspondingly, the UPF #1 receives the identifier of the second computing model from the CMF.

[0356] The CMF sends the identifier of the second computing model to the UPF #1, which can be used to instruct the UPF #1 to perform model switching. For more detailed description of step 622, please refer to the relevant description in step 470 of method 400 above, except that the first computing node in step 470 should be replaced by the UPF #1 in the embodiment, which will not be repeated here.

[0357] Exemplarily, the identification of the second computing model is carried in the computing task session modification message.

[0358] In step 623, the UPF#1 performs model switching.

[0359] In this embodiment, the first computing model can continue to be used for the first computing task, and send the result of the first computing task to the terminal device; the first computing model can also stop executing the first computing task. The present application does not limit this.

[0360] In step 624, the terminal device sends data of the second computing task to the UPF#1. Correspondingly, the UPF#1 receives the data of the second computing task from the terminal device. It can be understood that the data of the first computing task sent by the terminal device to the UPF#1 is uplink data.

[0361] In step 625, the UPF#1 executes the second computing task by using the second computing model.

[0362] The UPF#1 can input the data of the second computing task from the terminal device into the second computing model to execute the second computing task by using the second computing model, and obtain the computing result of the second computing task.

[0363] It should be noted that the UPF#1 executing the second computing task by using the second computing model can be regarded as one possible operation performed by the UPF#1 in model switching. In specific implementation, the UPF#1 can also perform other operations, such as loading the second computing model; for example, stopping executing the first computing task by using the first computing model; for example, after stopping executing the first computing task, releasing the resources occupied by executing the first computing task, and the like, which are not limited by the present application.

[0364] In some implementations, steps 623 and 625 can also be executed as one step.

[0365] In step 626, the UPF#1 sends the computing result of the second computing task to the terminal device. Correspondingly, the terminal device receives the computing result of the second computing task from the UPF#1. It can be understood that the computing result of the second computing task sent by the UPF#1 to the terminal device is downlink data.

[0366] Based on the above process, the CMF can dynamically adjust the computing model in the case that the transmission delay of the service changes and further causes the computing delay requirement of the service to change, so as to adapt to the change of the transmission link environment, thereby more effectively controlling the processing delay of the service and further controlling the overall delay of the service. And in the model switching process, the terminal device can always interact with the UPF#1 based on the same address information for data interaction of the computing task, thereby avoiding service migration, avoiding service interruption, and improving user experience. Especially for those services with strict requirements on delay stability, better service experience can be obtained.

[0367] Referring to FIG. 9, in the process 600B shown in FIG. 9, the step 627 can be executed after the step 609, or can also be executed synchronously with any one or more steps in the steps 610 to 617. The step 628 can be executed after the step 611, or can also be executed synchronously with any one or more steps in the steps 611 to 617. In the figure, only the case that the step 627 is executed after the step 609 and the step 628 is executed after the step 617 is shown, but this should not constitute any limitation on the present application. The execution order of each step should be determined according to the logical order between steps.

[0368] The details of each step in FIG. 9 are described below.

[0369] In the step 627, the CMF sends a first subscription request to the NRF, and the first subscription request is used to subscribe to the change of the node information #1 of the UPF#1. Accordingly, the NRF receives the first subscription request from the CMF.

[0370] Exemplarily, the first subscription request carries the identifier of the UPF#1 to request to subscribe to the change of the node information #1 of the UPF#1.

[0371] In the step 628, the UPF#1 sends the changed node information #1 to the NRF. Accordingly, the NRF receives the changed node information #1 from the UPF#1.

[0372] Each computing node can actively send the changed node information to the NRF in the case that the node information of the computing node changes. In the present embodiment, the node information #1 of the UPF#1 changes, and therefore the UPF#1 can send the changed node information #1 of the UPF#1 to the NRF.

[0373] In the step 629, the NRF sends the changed node information #1 of the UPF#1 to the CMF. Accordingly, the CMF receives the changed node information #1 of the UPF#1 from the NRF.

[0374] Based on the first subscription request, the NRF can send the changed node information of the UPF#1 to the CMF upon receiving the changed node information of the UPF#1.

[0375] Since the NRF can locally store the node information of at least one computing node, the NRF can send the changed node information #1 of the UPF#1 to the CMF according to the identity of the UPF#1.

[0376] In step 630, the CMF determines that the first computing model no longer meets the computing latency requirement.

[0377] Unlike step 620 in 600A, the computing latency requirement in step 630 has not changed, i.e., the computing latency requirement can be the computing latency requirement determined through step 607 of FIG. 7. However, since the node information #1 of the UPF#1 has changed, the CMF needs to rejudge whether the computing capability of the first computing model meets the computing latency requirement according to the changed node information #1 of the UPF#1.

[0378] For example, in step 607, the CMF determines that the computing latency requirement of the service is less than or equal to 40 ms; the node information #1 of the UPF#1 contains the corresponding relationship shown in Table 2, and in step 610, the CMF determines that the first computing model is model 4 (i.e., an example of the first computing model) in Table 2, which can meet the computing latency requirement of less than or equal to 40 ms. However, since the node information #1 of the UPF#1 has changed, the computing capability of the model 4 no longer supports the computing latency of 40 ms, but supports the computing latency of 45 ms, so the CMF can determine that the first computing model no longer meets the computing latency requirement.

[0379] The CMF can determine a second computing model that meets the computing latency requirement from a plurality of computing models supported by the UPF#1 through step 621, and send the identity of the second computing model to the UPF#1 through step 622, so that the UPF#1 switches the computing model through step 623.

[0380] Thereafter, steps 624 to 626 can be performed between the terminal device and the UPF#1.

[0381] For more detailed descriptions of steps 621 to 626, please refer to the relevant descriptions of steps 621 to 626 in the above 600A.

[0382] Based on the above process, the CMF can dynamically adjust the computing model when the computing capability of the first computing model changes and the computing latency requirement of the service is no longer met, so as to adapt to the change of the transmission link environment, thereby more effectively controlling the processing latency of the service and the overall latency of the service. During the model switching process, the terminal device can always perform data interaction with the UPF#1 based on the same address information, thereby avoiding service migration, avoiding service interruption, and improving user experience. Especially for services with strict latency stability requirements, better service experience can be obtained.

[0383] As an example, FIG. 10 is a schematic flowchart of a communication method according to another embodiment of the present application. The method 1000 shown in FIG. 10 can include steps 1001 to 1017. Each step in FIG. 10 will be described in detail below.

[0384] It should be noted that in one design, the EAS and the AF can be independent physical nodes, in which case the EAS can send information to the NEF through the AF and can also receive information from the NEF through the AF. In another design, the EAS and the AF can be located in the same physical node as modules or units for implementing different functions in the same physical node, in which case the EAS sending information to the NEF or receiving information from the NEF through the AF can be manifested as the above-mentioned physical node sending information to the NEF or receiving information from the NEF. In FIG. 10, the EAS and the AF are shown as different functional modules for example only, and the specific form of the EAS and the AF is not limited.

[0385] In step 1001, each EAS sends node information to the NEF. Correspondingly, the NEF receives the node information from each EAS.

[0386] Each EAS can load one or more computing models locally, and obtain the identifier of each computing model and the computing capability supported by each computing model. Since the EAS does not belong to the core network element, it needs to interact with the core network element through the NEF. Therefore, each EAS can actively send its own node information to the NEF after its function as a computing node is online, so as to expose its capability information to the NEF.

[0387] Two computing nodes, EAS#1 and EAS#2, are shown in this embodiment. Correspondingly, step 1001 can include: step 10011, EAS#1 sending node information #1 to the NEF; and step 10012, EAS#2 sending node information #2 to the NEF.

[0388] The node information #i (i can be 1 or 2) is the node information of the UPF #i, which can indicate the identification of each of one or more computing models supported by the UPF #i, and the computing capability supported by each of the computing models.

[0389] Optionally, the node information #i further includes the identification of the EAS #i (i.e., the identification of the computing node). By providing the identification of each EAS, each computing node and its node information can be distinguished.

[0390] Optionally, the node information #i further includes the location information of the EAS #i (i.e., the location information of the computing node), such as the DNAI. By providing the location information of each EAS, the geographical location of each computing node in the network can be provided to the NEF.

[0391] Optionally, the node information #i further includes the address information of the EAS #i (i.e., the address information of the computing node), such as the IP address and the port number. By providing the address information of each EAS, the CMF can send information to each computing node, such as sending the identification of the first computing model to the EAS #1.

[0392] Since the EAS itself is an AS, it can support one or more services. Therefore, each EAS can also expose the services it supports and the parameters related to the services to the NEF.

[0393] Optionally, the node information #i further includes one or more of the following: FQDN, DNN, S-NSSAI, service model, or end-to-end latency threshold of the service. The FQDN can be used to DNS resolve the address information of the EAS #i. The DNN and / or S-NSSAI correspond to the service, which can be understood as the service supported by the EAS #i, or in other words, the service that can be provided. The service model and the end-to-end latency threshold of the service are related to the computing task of the service, and can be used to determine the computing latency requirement of the service. For more detailed descriptions of each parameter, please refer to the related descriptions in step 420 above, except that the computing node in step 420 should be replaced by the EAS in this embodiment, which will not be described again.

[0394] Exemplarily, the node information of each EAS can be EAS deployment information, and the AF can send the EAS deployment information of each EAS to the NEF through an EAS deployment creation / update / deletion request (Nnef_EASDeployment_Create / Update / Delete Request) of a NEF interface. The present application does not limit this. Further, the NEF can also send the received EAS deployment information of each EAS to the UDR through a DM creation / update / deletion request (Nudr_DM_Create / Update / Delete Request) of a UDR interface.

[0395] In step 1002, the CMF sends a second subscription request to the NEF, and the second subscription request is used to subscribe to the node information of each EAS. Accordingly, the NEF receives the second subscription request from the CMF.

[0396] The CMF can subscribe to the node information of each EAS through the second subscription request.

[0397] In step 1003, the NEF sends the node information of each EAS to the CMF.

[0398] In response to the second subscription request, the NEF sends the received node information of each EAS to the NEF.

[0399] Exemplarily, the second subscription request described above can be an EAS deployment subscription request (Nnef_EASDeployment_Subscribe request) of a NEF interface. In response, the node information of each EAS can be carried in an EAS deployment subscription response (Nnef_EASDeployment_Subscribe response) of a NEF interface.

[0400] In step 1004, the terminal device sends a session establishment request to the CMF. Accordingly, the CMF receives the session establishment request from the terminal device.

[0401] For more detailed description of step 1004, please refer to the related description of step 604 in method 600 described above, which will not be repeated here.

[0402] In step 1005, the CMF performs UPF selection and establishes an N3 tunnel between the RAN and the UPF.

[0403] The CMF can select a UPF for the terminal device in response to the session establishment response, and further establish an N3 tunnel between the RAN and the UPF. Illustratively, the CMF can select the UPF for the terminal device according to the location information of the terminal device.

[0404] It should be noted that, since the computing node is the EAS in this embodiment, the terminal device needs to interact with the EAS for data of a computing task through the UPF, and therefore the N3 tunnel between the RAN and the UPF still needs to be established. In addition, the establishment process of the N3 tunnel can refer to the existing session establishment process, and will not be described in detail herein.

[0405] In step 1006, the terminal device sends a DNS query to an EAS discovery function (EASDF), and the DNS query carries the FQDN. Correspondingly, the EASDF receives the DNS query from the terminal device.

[0406] In this embodiment, the FQDN can be used to indicate a service to be accessed by the terminal device, and therefore the terminal device can carry the FQDN corresponding to the service in the DNS query.

[0407] In step 1007, the EASDF sends a DNS context notification request to the CMF, and the DNS context notification request carries the FQDN. Correspondingly, the terminal device receives the DNS context notification request from the EASDF.

[0408] The EASDF can carry the FQDN received from the terminal device in the DNS context notification request. Alternatively, the FQDN is contained in a DNS information report, and the DNS information report is carried in the DNS context notification request.

[0409] In step 1008, the CMF determines the computing latency requirement of the service.

[0410] In response to the session establishment request from the terminal device, the CMF can determine the computing latency requirement of the service. For more detailed description of step 1008, refer to the detailed description of step 410 of method 400 described above, which will not be repeated.

[0411] In step 1009, the CMF determines a first computing node and a first computing model that meet the computing latency requirement according to the node information of each EAS.

[0412] The CMF can determine the first computing node and the first computing model satisfying the computing latency requirement according to the computing capability supported by each computing model in the node information of each EAS. As described above, the computing node satisfying the computing latency requirement is the computing node in which there is at least one computing model satisfying the computing latency requirement. Therefore, in an implementation, the CMF can directly determine the first computing model satisfying the computing latency requirement from the plurality of computing models supported by the EAS#1 and the EAS#2, and then determine the computing node to which the first computing model belongs as the first computing node, for example, the EAS#1.

[0413] In another implementation, the CMF can also determine the first computing node satisfying the computing latency requirement, for example, the EAS#1, according to the computing capability supported by each computing model in the node information of each EAS, and then determine the first computing model satisfying the computing latency requirement from the plurality of computing models supported by the EAS#1.

[0414] For more detailed description of how the CMF determines the first computing model satisfying the computing latency requirement from the plurality of computing models, reference can be made to the description of step 420 of the method 400, which will not be repeated here.

[0415] It should be understood that the step 1005 can be performed after the step 1004, or after the step 1009, or after any one of the steps 1004 to 1009, or simultaneously, and the present application does not limit this. When the step 1005 is performed after the step 1009, the CMF can select the UPF in combination with the geographic location of the terminal device and the geographic location of the EAS#1, and then establish the N3 tunnel between the RAN and the UPF.

[0416] In step 1010, the CMF sends the identifier of the EAS#1 and the identifier of the first computing model to the AF.

[0417] The CMF can send the identifier of the first computing model to the EAS#1 through the AF. Since the AF can interact with multiple EASs, the CMF can send the identifier of the EAS#1 and the identifier of the first computing model to the AF.

[0418] For example, the identifier of the EAS#1 and the identifier of the first computing model sent by the CMF to the AF can be carried in the time exposure notification (N nef _EventExposure_Notify) of the NEF interface.

[0419] In step 1011, the AF sends the identifier of the first computing model to the EAS#1.

[0420] The AF can send the identifier of the first computing model to the EAS#1 according to the identifier of the EAS#1.

[0421] In step 1012, the CMF sends a DNS Context Notify Response to the EASDF, which carries the address information of the EAS#1. Accordingly, the EASDF receives the DNS Context Notify Response from the CMF.

[0422] On the other hand, in response to the DNS Context Notify Request sent by the EASDF to the CMF in step 1007, the CMF can send the address information of the EAS#1, such as the IP address and port number of the EAS#1, to the EASDF.

[0423] Exemplarily, the DNS information report in step 1007 is carried in a DNS Context Notify Request (DNSContext_Notify Request) of the EASDF interface. In response, the address information of the EAS#1 in step 1012 is carried in a DNS Context Notify Response (DNSContext_Notify Response) of the EASDF interface. easdf easdf

[0424] In step 1013, the EASDF sends a DNS response to the terminal device, which carries the address information of the EAS#1. Accordingly, the terminal device receives the DNS response from the EASDF.

[0425] In response to the DNS query received in step 1006, the EASDF sends a DNS response to the terminal device after obtaining the address information of the EAS#1, and carries the address information of the EAS#1 in the DNS response.

[0426] It can be seen that by performing the aforementioned steps 1006, 1007, 1012 and 1013, the terminal device obtains the address information of the EAS#1 from the EASDF. However, this is only a possible implementation manner, and should not constitute any limitation to the present application. In another implementation manner, the above steps can be replaced by the step 1014 shown below.

[0427] In step 1014, the CMF sends a session establishment response to the terminal device, which carries the address information of the EAS#1. Accordingly, the terminal device receives the session establishment response from the CMF.

[0428] The CMF can send the address information of the EAS#1 to the terminal device through the session establishment response. Thus, the terminal device can obtain the address information of the EAS#1, and further can perform data interaction of the computing task with the EAS#1.

[0429] ​​Therefore, step 1014 can be used to replace the foregoing steps 1006, 1007, 1012 and 1013. In other words, step 1014 and the foregoing steps 1006, 1007, 1012 and 1013 provide two possible implementation manners for the terminal device to obtain the address information of the EAS#1 (i.e., the address information of the first computing node). The two implementation manners can be executed alternatively, or in other words, step 1014 and the foregoing steps 1006, 1007, 1012 and 1013 can be executed alternatively, and do not necessarily have to be executed all.

[0430] It should be understood that when step 1014 is used to replace the foregoing steps 1006, 1007, 1012 and 1013, step 1014 can be executed after step 1009 and before step 1015.

[0431] In step 1015, the terminal device sends data of the first computing task to the EAS#1.

[0432] In step 1016, the EAS#1 executes the first computing task by using the first computing model.

[0433] In step 1017, the EAS#1 sends a computing result of the first computing task to the terminal device.

[0434] The procedures of steps 1015 to 1017 can be understood with reference to the related description in steps 615 to 617 of the foregoing method 600, except that the UPF#1 in steps 615 to 617 should be replaced by the EAS#1 in the present embodiment, and thus no longer be described herein.

[0435] Based on the foregoing procedures, the terminal device can interact with the EAS#1 for data of a computing task. Since the EAS#1 and the first computing model used to execute the first computing task are both determined by the CMF according to the computing delay requirement of the service, when the first computing task is executed by using the first computing model, the processing delay of the service meets the computing delay requirement. In this way, the processing delay of the service is controlled, thereby facilitating the control of the overall delay of the service to meet the delay requirement of the service.

[0436] As mentioned above, since the environment of the transmission link changes and the load of the computing model also changes. Therefore, the first computing model does not necessarily always meet the foregoing computing delay requirement. The CMF can continuously monitor during the execution of the first computing task, so as to determine a new computing model for the service of the terminal device in the case that the first computing model no longer meets the computing delay requirement, and then instruct the EAS#1 to perform model switching.

[0437] For the convenience of understanding, the process of model switching will be described below in combination with the flows shown in FIG. 11 and FIG. 12. It should be understood that FIG. 11 and FIG. 12 respectively show two different cases in which the first computing model no longer meets the computing latency requirement, and for the convenience of distinguishing and description, the flow shown in FIG. 8 is denoted as 1000A, and the flow shown in FIG. 12 is denoted as 1000B. The flows shown in FIG. 11 and FIG. 12 can be combined with the flow shown in FIG. 10. For brevity, the same steps as in FIG. 10 will not be described again. The flow shown in FIG. 10 can be executed alone or in combination with the flows shown in FIG. 11 and FIG. 12. For brevity, the same steps as in FIG. 10 in the flows shown in FIG. 11 and FIG. 12 will not be described again, and reference can be made to the description of the relevant steps above.

[0438] Referring to FIG. 11, the flow 1000A shown in FIG. 11 can be executed after step 1015 of the foregoing method 1000, or after step 1017, which is not limited in the present application. The flow is shown in the figure after step 1017 only as an example, but this should not constitute any limitation on the present application. The execution order of each step should be determined according to the logical order between steps.

[0439] In step 1018, EAS#1 reports the uplink transmission latency to AF. Correspondingly, AF receives the uplink transmission latency from EAS#1.

[0440] In step 1019, AF reports the uplink transmission latency to CMF. Correspondingly, CMF receives the uplink transmission latency from AF.

[0441] Steps 1018 and 1019 can be understood with reference to step 618 in flow 600A, except that UPF#1 in step 618 should be replaced by EAS#1 in the present embodiment, and EAS#1 sends the uplink transmission latency to CMF through AF, which will not be described again.

[0442] In step 1020, CMF determines that the computing latency requirement of the service has changed.

[0443] In step 1021, CMF determines that the first computing model no longer meets the changed computing latency requirement.

[0444] In step 1022, CMF determines a second computing model that meets the computing latency requirement from a plurality of computing models supported by EAS#1.

[0445] Steps 1020 to 1022 can be understood with reference to steps 619 to 621 in flow 600A, except that UPF#1 in steps 619 to 621 should be replaced by EAS#1 in the present embodiment, which will not be described again.

[0446] In step 1023, the CMF sends the identifier of the EAS#1 and the identifier of the second computing model to the AF. Accordingly, the AF receives the identifier of the EAS#1 and the identifier of the second computing model from the CMF.

[0447] Exemplarily, the identifier of the EAS#1 and the identifier of the second computing model sent by the CMF to the AF in step 1023 can be carried in the NEF interface event exposure notification (Nnef_EventExposure_Notify).

[0448] In step 1024, the AF sends the identifier of the second computing model to the EAS#1. Accordingly, the EAS#1 receives the identifier of the second computing model from the AF.

[0449] Steps 1023 and 1024 can be understood with reference to step 622 in flow 600A, except that the UPF#1 in step 622 should be replaced by the EAS#1 in the present embodiment, and the CMF can send the identifier of the second computing model to the EAS#1 through the AF, at this time, the CMF sends not only the identifier of the second computing model to the AF, but also the identifier of the EAS#1, which will not be repeated here.

[0450] In step 1025, the EAS#1 performs model switching.

[0451] Step 1025 can be understood with reference to step 623 in flow 600A, except that the UPF#1 in step 623 should be replaced by the EAS#1 in the present embodiment, which will not be repeated here.

[0452] Thereafter, steps 1026 to 1028 can be performed between the terminal device and the EAS#1.

[0453] Steps 1026 to 1028 can be understood with reference to steps 624 to 626 in flow 600A, except that the UPF#1 in steps 624 to 626 should be replaced by the EAS#1 in the present embodiment, which will not be repeated here.

[0454] Based on the above flow, the CMF can dynamically adjust the computing model in the case that the transmission delay of the service changes and further causes the computing delay requirement of the service to change, so as to adapt to the change of the transmission link environment, thereby more effectively controlling the processing delay of the service, and further controlling the overall delay of the service. And in the model switching process, the terminal device can always perform data interaction with the UPF#1 based on the same address information for the computing task, thereby avoiding service migration, avoiding service interruption, and improving user experience. Especially for those services with strict requirements on delay stability, better service experience can be obtained.

[0455] Referring to FIG. 12, in the flow 1000B shown in FIG. 12, the step 1029 can be executed after the step 1017, or can also be executed after the step 1011, or can also be executed after the step 1003, or can also be executed after or synchronously with any one or more of the steps 1003 to 1017. The step 1029 is shown as an example as being executed after the step 1017, but this should not constitute any limitation on the present application. The execution order of the steps should be determined according to the logical order between the steps.

[0456] The steps in FIG. 12 will be described in detail below.

[0457] In the step 1029, the EAS#1 sends the changed node information#1 to the NEF. Correspondingly, the NEF receives the changed node information#1 from the EAS#1.

[0458] In the step 1030, the NEF sends the changed node information#1 of the EAS#1 to the CMF. Correspondingly, the CMF receives the changed node information#1 of the EAS#1 from the AF.

[0459] As mentioned in the steps 1002 and 1003 above, the CMF can subscribe to the node information of each EAS. Therefore, the EAS#1 can actively expose the changed information to the NEF in the case that the node information#1 of the EAS#1 is changed. In the present embodiment, the EAS#1 sends the changed node information#1 to the NEF through the AF, and the NEF sends the received changed node information#1 of the EAS#1 to the CMF.

[0460] Exemplarily, the changed node information#1 sent by the EAS#1 to the NEF in the step 1029 can be carried in the EASDeployment Information, or the node information of each EAS can be the EASDeployment Information, which is not limited in the present application.

[0461] In the step 1031, the CMF determines that the first computing model no longer meets the computing latency requirement.

[0462] Unlike the step 1021 in 1000A, the computing latency requirement in the step 1032 is not changed, i.e., the computing latency requirement can be the computing latency requirement determined through the step 1008 in FIG. 10. However, due to the change of the node information#1 of the EAS#1, the CMF needs to re-judge whether the computing capability of the first computing model meets the computing latency requirement according to the changed node information#1 of the EAS#1.

[0463] Step 1032 can be understood with reference to step 630 in flow 600B, except that UPF#1 in step 630 should be replaced by EAS#1 in the present embodiment, and thus is not described herein again.

[0464] In step 1022, the CMF determines a second computing model satisfying the computing latency requirement from a plurality of computing models supported by the EAS#1.

[0465] Step 1022 can be understood with reference to step 621 in flow 600A, except that UPF#1 in step 621 should be replaced by EAS#1 in the present embodiment, and thus is not described herein again.

[0466] In step 1023, the CMF sends, to the AF, an identification of the EAS#1 and an identification of the second computing model. Accordingly, the AF receives, from the CMF, the identification of the EAS#1 and the identification of the second computing model.

[0467] Exemplarily, the identification of the EAS#1 and the identification of the second computing model sent by the CMF to the AF in step 1023 can be carried in a NEF interface event exposure notification (N nef _EventExposure_Notify).

[0468] In step 1024, the AF sends, to the EAS#1, the identification of the second computing model. Accordingly, the EAS#1 receives, from the AF, the identification of the second computing model.

[0469] Steps 1023 and 1024 can be understood with reference to step 622 in flow 600A, except that UPF#1 in step 622 should be replaced by EAS#1 in the present embodiment, and the CMF can send, to the EAS#1, the identification of the second computing model through the AF, in which case the CMF sends, to the AF, not only the identification of the second computing model but also the identification of the EAS#1, and thus is not described herein again.

[0470] In step 1025, the EAS#1 performs model switching.

[0471] Step 1025 can be understood with reference to step 623 in flow 600A, except that UPF#1 in step 623 should be replaced by EAS#1 in the present embodiment, and thus is not described herein again.

[0472] In step 1026, the terminal device sends, to the EAS#1, data of a second computing task.

[0473] In step 1027, the EAS#1 executes the second computing task by using the second computing model.

[0474] In step 1028, the EAS#1 sends, to the terminal device, a computing result of the second computing task.

[0475] Steps 1026 to 1028 can be understood with reference to steps 624 to 626 in flow 600A, except that UPF#1 in steps 624 to 626 should be replaced by EAS#1 in the present embodiment, and will not be repeated.

[0476] Based on the above flow, the CMF can dynamically adjust the computing model when the computing capability of the first computing model changes and the computing latency requirement of the service is no longer met, so as to adapt to the change of the transmission link environment, thereby more effectively controlling the processing latency of the service, and further controlling the overall latency of the service. And in the model switching process, the terminal device can always interact with the UPF#1 based on the same address information for data interaction of the computing task, thereby avoiding service migration, avoiding service interruption, and improving user experience. Especially for those services with strict requirements on latency stability, better service experience can be obtained.

[0477] The above describes in detail the method provided by the embodiments of the present application in combination with the drawings. In the following, the apparatus provided by the present application will be described in combination with the drawings.

[0478] As an example, a communication apparatus provided by the present application is shown in FIG. 13, and the communication apparatus 1300 includes a transceiver unit 1310 and a processing unit 1320.

[0479] A possible design is that the communication apparatus 1300 is configured to implement the functions of the CMF in the method embodiments shown in FIGS. 4 to 12.

[0480] Exemplarily, the processing unit 1320 can be configured to: obtain a computing latency requirement of a service; determine a first computing model satisfying the computing latency requirement from a plurality of computing models supported by a first computing node according to node information of the first computing node, the node information indicating an identifier of each of the plurality of computing models and a computing capability supported by each of the plurality of computing models, the first computing model being used to perform a first computing task of the service; and the transceiver unit 1310 is further configured to send the identifier of the first computing model to the first computing node.

[0481] Optionally, the processing unit 1320 is further configured to determine a second computing model satisfying the computing latency requirement from the plurality of computing models; and the transceiver unit 1310 is further configured to send the identifier of the second computing model to the first computing node.

[0482] Optionally, the processing unit 1320 can further be configured to obtain a transmission delay of the service, the transmission delay being determined according to one or more of: an uplink maximum data volume of the service, a downlink maximum data volume of the service, an execution period of the service, or an uplink transmission delay reported by the first computing node; and determine that the first computing model no longer meets the computing delay requirement according to the transmission delay.

[0483] Optionally, the transceiver 1310 can further be configured to receive changed node information of the first computing node; and the processing unit 1320 can further be configured to determine that the first computing model no longer meets the computing delay requirement according to the changed node information of the first computing node.

[0484] Optionally, the processing unit 1320 can further be configured to obtain an end-to-end delay threshold of the service and a transmission delay, the transmission delay being determined according to one or more of: an uplink maximum data volume of the service, a downlink maximum data volume of the service, or an execution period of the service; and determine the computing delay requirement according to the end-to-end delay of the service and the transmission delay.

[0485] In an implementation manner, the first computing node is a core network element.

[0486] Optionally, the transceiver 1310 can further be configured to send a discovery request to an NRF, the discovery request carrying information of the computing delay requirement; and receive node information of the first computing node meeting the computing delay requirement from the NRF.

[0487] Optionally, the transceiver 1310 can further be configured to send a first subscription request to the NRF, the first subscription request being used to subscribe to a change of the node information of the first computing node.

[0488] In another implementation manner, the first computing node is an AS.

[0489] Optionally, the processing unit 1320 can further be configured to determine that the first computing node meets the computing delay requirement according to the node information of the first computing node.

[0490] Optionally, the transceiver 1310 can further be configured to: send a second subscription request to an NEF, the second subscription request being used to subscribe to node information of the AS; and receive the node information of the AS from the NEF.

[0491] Optionally, the transceiver 1310 is further configured to receive a session establishment request from a terminal device, and send a session establishment response to the terminal device, where the session establishment response carries address information of the first computing node. More details of the transceiver 1310 and the processing unit 1320 can be referred to the descriptions of the CMF in the method embodiments of FIG. 4 to FIG. 12.

[0492] In another possible design, the communication apparatus 1300 is configured to implement the functions of the computing nodes (including the first computing node) in the method embodiments of FIG. 4 and FIG. 5, the functions of the UPF#1 and the UPF#2 in the method embodiments of FIG. 7 to FIG. 9, and the functions of the EAS#1 and the EAS#2 in the method embodiments of FIG. 10 to FIG. 12.

[0493] For example, the processing unit 1320 is configured to determine an identity of each of one or more computing models supported by the computing node, and a computing capability supported by each of the one or more computing models, and the transceiver 1310 is configured to send node information, where the node information indicates the identity of each of the one or more computing models, and the computing capability supported by each of the one or more computing models.

[0494] Optionally, the transceiver 1310 is further configured to receive an identity of a first computing model, where the first computing model is one of the plurality of computing models, and the processing unit 1320 is further configured to perform a first computing task of the service by using the first computing model.

[0495] Optionally, the transceiver 1310 is further configured to receive an identity of a second computing model, where the first computing model is one of the plurality of computing models, and the second computing model supports a different computing capability from the computing capability supported by the first computing model, and the processing unit 1320 is further configured to perform a second computing task of the service by using the second computing model.

[0496] More details of the transceiver 1310 and the processing unit 1320 can be referred to the descriptions of the computing nodes (including the first computing node) in the method embodiments of FIG. 4 to FIG. 5, the descriptions of the UPF#1 and the UPF#2 in the method embodiments of FIG. 7 to FIG. 9, and the descriptions of the EAS#1 and the EAS#2 in the method embodiments of FIG. 10 to FIG. 12.

[0497] In yet another possible design, the communication apparatus 1300 is configured to implement the functions of the NRF in the method embodiments of FIG. 7 to FIG. 9.

[0498] Exemplarily, the transceiver 1310 can be configured to receive a discovery request from a CMF, the discovery request carrying information of a computing latency requirement; the processor 1320 can be configured to determine that a first computing node satisfies the computing latency requirement, the first computing node supporting a plurality of computing models, and computing capability of one or more computing models in the plurality of computing models satisfying the computing latency requirement; the transceiver 1310 can be further configured to send node information of the first computing node to the CMF, the node information indicating an identity of each computing model in the plurality of computing models and computing capability supported by each computing model in the plurality of computing models.

[0499] Optionally, the transceiver 1310 can be further configured to receive a subscription request from the CMF, the subscription request being used to subscribe to a change of the node information of the first computing node.

[0500] Optionally, the transceiver 1310 can be further configured to send the changed node information of the first computing node to the CMF upon receiving the changed node information from the first computing node.

[0501] For more details of the transceiver 1310 and the processor 1320, refer to the description of the NRF in the method embodiments of FIGS. 7-9.

[0502] In yet another possible design, the communication apparatus 1300 is configured to implement the functions of the NEF in the method embodiments of FIGS. 10-12.

[0503] Exemplarily, the transceiver 1310 can be configured to receive node information of a computing node, the node information including an identity of each computing model in one or more computing models supported by the computing node and computing capability supported by each computing model in the one or more computing models; receive a subscription request from a CMF, the subscription request being used to request the node information of the computing node; and send the node information of the computing node to the CMF.

[0504] For more details of the transceiver 1310 and the processor 1320, refer to the description of the NEF in the method embodiments of FIGS. 10-12.

[0505] It should be noted that the transceiver unit can also be referred to as a transceiver module, a transceiver, a transceiver device, or the like. The processing unit can also be referred to as a processor, a processing board, a processing module, or a processing device, or the like. Alternatively, the transceiver unit is used to perform the sending operation and the receiving operation of the intermediate relay or the U2N relay in the above method, the device in the communication module for realizing the receiving function can be regarded as a receiving unit, and the device in the communication module for realizing the sending function can be regarded as a sending unit, that is, the transceiver unit includes the receiving unit and the sending unit.

[0506] It should also be noted that in a possible design, the foregoing transceiver unit and / or processing unit can be implemented through a virtual module, for example, the processing unit can be implemented through a software function unit or a virtual device, and the transceiver unit can be implemented through a software function or a virtual device. In another possible design, the processing unit or the transceiver unit can also be implemented through an entity device, for example, if the device is implemented by using a chip / chip circuit, the transceiver unit can be an input / output circuit and / or a communication interface, and performs an input operation (corresponding to the foregoing receiving operation) and an output operation (corresponding to the foregoing sending operation); the processing unit is an integrated processor or a microprocessor or an integrated circuit.

[0507] The division of units in the embodiments of the present application is illustrative, and is merely a logical function division. In actual implementation, another division manner can be used. In addition, each function module in each example in the embodiments of the present application can be integrated in one processor, or can be physically separated, or two or more modules can be integrated in one module. The integrated module can be implemented in the form of hardware or in the form of a software function module.

[0508] As an example, another communication device provided by the present application is shown in FIG. 14. The communication device 1400 includes at least one processor 1410. The at least one processor 1410 can be used to execute computer programs or instructions in the memory to implement the steps performed by each CMF or the steps performed by each computing node in the embodiments shown in FIGS. 4 to 12, or to cause the computer to perform the method performed by the NRF or the steps performed by the AF in the embodiments shown in FIGS. 7 to 9, or to cause the computer to perform the steps performed by the NEF or the steps performed by the AF in the embodiments shown in FIGS. 10 to 12.

[0509] Optionally, the communication apparatus 1400 further includes at least one memory 1420 for storing instructions executed by the processor 1410 or storing input data required by the processor 1410 to run the instructions or storing data generated by the processor 1410 after running the instructions. The at least one processor 1410 and the at least one memory 1420 can be separately arranged. For example, each memory can be connected with one or more processors, so that the connected processor can read information from the memory, store and / or write information in the memory. Alternatively, the at least one processor 1410 and the at least one memory 1420 can be integrated together, for example, one or more memories can be integrated in one processor.

[0510] Optionally, the communication apparatus 1400 further includes an interface circuit 1430, which can be used for transmitting data and / or signaling. The at least one processor 1410 and the interface circuit 1430 are coupled with each other. It can be understood that the interface circuit 1430 can be a transceiver, an input / output circuit, a bus, a module, a pin or other types of communication interfaces, wherein the input circuit in the input / output circuit can be used for receiving, and the output interface can be used for transmitting.

[0511] When the communication apparatus 1400 is used to implement the methods shown in FIGS. 4 to 12, the processor 1410 is used to execute the functions of the processing units described above, and the interface circuit 1430 is used to execute the functions of the transceiving units described above. Whether the interface circuit 1430 is used for transmitting or receiving can be determined according to whether the communication apparatus 1400 is used to execute a transmitting action or a receiving action in the scheme.

[0512] It can be understood that when the communication apparatus 1400 is a communication device (such as an intermediate relay or a U2N relay), the interface circuit 1430 can be a transceiver, which can specifically include a transmitter and a receiver, the transmitter is used for transmitting signals, and the receiver is used for receiving signals. When the communication apparatus 1400 is a chip applied to a communication device, the interface circuit 1430 can be an input / output circuit, a bus, a module, a pin or other types of communication interfaces, wherein the input circuit in the input / output circuit can be used for receiving, and the output interface can be used for transmitting.

[0513] It should be understood that in the communication apparatus 1400 shown in FIG. 14, the processor 1410 can correspond to the processing unit 1320 in the communication apparatus 1300 described above, and the interface circuit 1430 can correspond to the transceiving unit 1310 in the communication apparatus 1300 described above.

[0514] It should also be understood that the coupling or the communication connection between the apparatuses, units or modules in the embodiments of the present application can be indirect coupling or communication connection between the apparatuses, units or modules, which can be electrical, mechanical or other forms, for information interaction between the apparatuses, units or modules. The specific connection medium between the at least one processor 1410, the at least one memory 1420, the interface circuit 1430 and the power supply circuit 1440 in the embodiments of the present application is not limited. In FIG. 14, the processor 1410, the memory 1420, the interface circuit 1430 and the power supply circuit 1440 are connected through the bus 1450. The bus 1450 is represented by a thick line in FIG. 14, and the connection mode between other components is only schematically illustrated and is not limited. 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 an address bus, a data bus, a control bus, etc. For convenience of representation, only one thick line is used to represent the bus in FIG. 14, but it does not mean that there is only one bus or only one type of bus.

[0515] It can be understood that the processor in the embodiments of the present application can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. The general-purpose processor can be a microprocessor or any conventional processor.

[0516] The memory in the embodiments of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically EPROM (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM) used as an external cache. By way of example, but not limitation, many forms of RAM can be used, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchlink DRAM (SLDRAM), and direct rambus RAM (DR RAM). It should be noted that the memory of the system and method described herein is intended to include, but not be limited to, these and any other suitable types of memory.

[0517] The present application also provides a communication system, comprising one or more of the following: a CMF, a computing node, a NEF, a NRF, an AF, or a terminal device. Optionally, the computing node is a UPF. Optionally, the computing node is an AS.

[0518] The present application also provides a computer program product, comprising a computer program (also called code or instructions) which, when executed by a computer, causes the computer to perform the method performed by each CMF or the method performed by the computing node in the embodiments shown in FIGS. 4-12, or causes the computer to perform the method performed by the NRF or the method performed by the AF in the embodiments shown in FIGS. 7-9, or causes the computer to perform the method performed by the NEF or the method performed by the AF in the embodiments shown in FIGS. 10-12.

[0519] The application further provides a computer readable storage medium storing a computer program (also referred to as code or instructions). When the computer program is run, the computer is caused to perform the method performed by each CMF in the embodiments shown in FIGS. 4 to 12 or the method performed by the computing node, or the computer is caused to perform the method performed by the NRF in the embodiments shown in FIGS. 7 to 9 or the method performed by the AF, or the computer is caused to perform the method performed by the NEF in the embodiments shown in FIGS. 10 to 12 or the method performed by the AF.

[0520] The terms "unit", "module", and the like used in the specification can be used to represent a computer-related entity, hardware, firmware, a combination of hardware and software, software, or software in execution.

[0521] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be realized in electronic hardware, or a combination of computer software and electronic hardware. Whether the functions are performed 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 the present application.

[0522] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the above-described system, device and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0523] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0524] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, i.e., they can be located in one place or distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the embodiments.

[0525] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit.

[0526] In the above embodiments, the functions of each functional unit can be implemented by software, hardware, firmware, or any combination thereof, in whole or in part. When implemented by software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions (programs). When loaded and executed on a computer, the computer program instructions (programs) produce the processes or functions described in the embodiments of the present application in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another, for example, the computer instructions can be transferred from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) mode. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. integrated with one or more available media. The available medium can be a magnetic medium (for example, floppy disk, hard disk, magnetic tape), optical medium (for example, digital video disc (DVD)), or semiconductor medium (for example, solid state disk (SSD)) and the like.

[0527] When the functions are implemented in the form of 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 solutions of the present application essentially or the parts that make contributions to the prior art or parts of the technical solutions can be embodied in the form of software products. The computer software product is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0528] The above merely provides the specific implementation of the present application, but the protection scope of the present application is not limited to this. Any person skilled in the art can easily think of the changes or replacements within the technical range disclosed by the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A communication method characterized by comprising: The method comprises: obtaining a computing latency requirement of a service; determining, according to node information of a first computing node, a first computing model satisfying the computing latency requirement from a plurality of computing models supported by the first computing node, the node information indicating an identity of each of the plurality of computing models and a computing capability supported by each of the plurality of computing models, the first computing model being used to perform a first computing task of the service; sending, to the first computing node, an identity of the first computing model.

2. The method of claim 1, wherein, In a case where the first computing model no longer satisfies the computing latency requirement, the method further comprises: determining, from the plurality of computing models, a second computing model satisfying the computing latency requirement, the second computing model corresponding to a computing latency different from a computing latency corresponding to the first computing model, the second computing model being used to perform a second computing task of the service; sending, to the first computing node, an identity of the second computing model.

3. The method of claim 2, wherein, Address information of the first computing node for the first computing task is same as address information of the first computing node for the second computing task.

4. The method of claim 2 or 3, wherein, The method further comprises: obtaining a transmission latency of the service, the transmission latency being determined according to one or more of: an uplink maximum data volume of the service, a downlink maximum data volume of the service, an execution period of the service, or an uplink transmission latency reported by the first computing node; determining, according to the transmission latency, that the first computing model no longer satisfies the computing latency requirement.

5. The method of claim 2 or 3, wherein, The method further comprises: receiving changed node information of the first computing node; determining, according to the changed node information of the first computing node, that the first computing model no longer satisfies the computing latency requirement.

6. The method of any one of claims 1 to 5, wherein, The first computing node is a core network element or an application server (AS).

7. The method of claim 6, wherein, The first computing node is the core network element, and before the determining, according to the node information of the first computing node, of the first computing model satisfying the computing latency requirement from the plurality of computing models supported by the first computing node, the method further comprises: sending, to a network storage function (NRF), a discovery request carrying information of the computing latency requirement; receiving, from the NRF, node information of the first computing node satisfying the computing latency requirement.

8. The method of claim 7, wherein, After the obtaining of the node information of the first computing node satisfying the computing latency requirement, the method further comprises: sending, to the NRF, a first subscription request for subscribing to a change of the node information of the first computing node.

9. The method of claim 6, wherein, The first computing node is the AS, and before the determining, according to the node information of the first computing node, of the first computing model satisfying the computing latency requirement from the plurality of computing models supported by the first computing node, the method further comprises: determining, according to the node information of the first computing node, that the first computing node satisfies the computing latency requirement.

10. The method of claim 9, wherein, Before the determining, according to the node information of the first computing node, that the first computing node satisfies the computing latency requirement, the method further comprises: sending a second subscription request to a network exposure function (NEF), the second subscription request being used to subscribe to node information of the AS; receiving the node information of the AS from the NEF.

11. The method of claim 9 or 10, wherein, The method further includes: receiving a session establishment request from a terminal device; sending a session establishment response to the terminal device, the session establishment response carrying address information of the first computing node.

12. The method of any one of claims 1 to 11, wherein, The obtaining of the computing latency requirement of the service includes: obtaining an end-to-end latency threshold of the service and a transmission latency, the transmission latency being determined according to one or more of: an uplink maximum data volume of the service, a downlink maximum data volume of the service, or an execution period of the service; determining the computing latency requirement according to the end-to-end latency of the service and the transmission latency.

13. A method of communication, comprising: including: determining an identity of each of one or more computing models supported by a computing node, and a computing capability supported by each of the one or more computing models; sending node information, the node information indicating the identity of each of the one or more computing models, and the computing capability supported by each of the one or more computing models.

14. The method of claim 13, wherein, The computing node is a core network element or an application server (AS).

15. The method of claim 13 or 14, wherein, The computing node supports multiple computing models, and the method further includes: receiving an identity of a first computing model, the first computing model being determined from the multiple computing models according to the node information; performing a first computing task of the service using the first computing model.

16. The method of claim 15, wherein, The method further includes: receiving an identity of a second computing model, the second computing model being determined from the multiple computing models according to the node information, and the second computing model supporting a different computing capability from the computing capability supported by the first computing model; performing a second computing task of the service using the second computing model.

17. The method of claim 16, wherein, The address information of the first computing node for the first computing task is the same as the address information of the first computing node for the second computing task.

18. A method of communication, comprising: including: receiving a discovery request from a computing management function (CMF), the discovery request carrying information of a computing latency requirement; determining, according to node information of a first computing node, that the first computing node satisfies the computing latency requirement, the first computing node supporting multiple computing models, there being one or more computing models in the multiple computing models whose computing capability satisfies the computing latency requirement, the node information indicating an identity of each of the multiple computing models, and a computing capability supported by each of the multiple computing models; sending the node information of the first computing node to the CMF.

19. The method of claim 18, wherein, The node information is used to determine a computing node and / or a computing model that satisfies the computing latency requirement.

20. The method of claim 18 or 19, wherein, The first computing node is a core network element.

21. The method of any one of claims 18 to 20, wherein, The method further includes: receiving a subscription request from the CMF, the subscription request being used to subscribe to a change in the node information of the first computing node.

22. The method of claim 21, wherein, The method further includes: In a case that the changed node information from the first computing node is received, send the changed node information of the first computing node to the CMF.

23. A method of communication, comprising: Comprise: receive node information of a computing node, the node information comprising: an identification of each of one or more computing models supported by the computing node, and a computing capability supported by each of the one or more computing models; receive a subscription request from a computing management function (CMF), the subscription request being used to request the node information of the computing node; send the node information of the computing node to the CMF.

24. The method of claim 23, wherein, The node information is used to determine a computing node and / or a computing model that meets a computing latency requirement.

25. A communication system, characterized by Comprise one or more of: a computing management function (CMF), a computing node, a network repository function (NRF), a network exposure function (NEF), or an application function (AF); wherein the CMF is configured to perform the method of any one of claims 1-12, the computing node is configured to perform the method of any one of claims 13-17, the NRF is configured to perform the method of any one of claims 18-22, and the NEF is configured to perform the method of claim 23 or 24.

26. A communications device, characterized by Comprise one or more functional units or modules configured to perform the method of any one of claims 1-12, or to perform the method of any one of claims 13-17, or to perform the method of any one of claims 18-22, or to perform the method of claim 23 or 24.

27. A communications device, characterized by Comprise a processor configured to execute program code to cause the communication apparatus to implement the method of any one of claims 1-12, or to implement the method of any one of claims 13-17, or to implement the method of any one of claims 18-22, or to implement the method of claim 23 or 24.

28. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by a processor, causes the method of any one of claims 1-12 to be performed, or causes the method of any one of claims 13-17 to be performed, or causes the method of any one of claims 18-22 to be performed, or causes the method of claim 23 or 24 to be performed.

29. A computer program product, characterised in that, The computer program, when executed by a processor, causes the method of any one of claims 1-12 to be performed, or causes the method of any one of claims 13-17 to be performed, or causes the method of any one of claims 18-22 to be performed, or causes the method of claim 23 or 24 to be performed.

Images