Communication method, related device, and communication system

Abstract

The present application relates to the field of computing power networks, and specifically relates to a communication method, a related device, and a communication system. The method comprises: an access network device receiving first information from a UE or a computing node, wherein the first information is determined on the basis of information obtained by the UE and / or the computing node using a first model to segment a first computing task; and the access network device determining a first QoS file on the basis of the first information, and scheduling, on the basis of a QoS flow corresponding to the first QoS file, a data packet corresponding to the first computing task. Using the embodiments of the present application helps a UE and / or a computing node to meet end-to-end QoS requirements of a computing task when using a model to perform segmentation processing on the computing task.

Description

Communication methods, related equipment and communication systems

[0001] This application claims priority to Chinese Patent Application No. 202510127767.9, filed on January 27, 2025, entitled "Communication Method, Related Equipment and Communication System", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of computing power networks, and in particular to a communication method, related equipment and communication system. Background Technology

[0003] With the development of wireless communication networks and artificial intelligence technologies, the computational power requirements for inference in multimodal interactive artificial intelligence (AI) applications have surged. Due to limitations in computing power, power, and memory, terminal devices (user equipment, UE) face new challenges in developing computing power applications. Industry experts believe that cloud-edge-device collaboration can alleviate the computational pressure on the device side.

[0004] The existing network architecture is designed with routing methods tailored to communication requirements. For example, after uplink data packets are sent to the user plane function (UPF) network element, the UPF element forwards the packets to the corresponding server based on their destination IP address. The core network processes data packets, only responsible for packet routing and forwarding. A concrete analogy can be drawn from existing applications (APPs) providing generative AI services. Users experience this service through an APP on the UE (User Equipment). For example, the UE's APP sends generative AI tasks to the server: "Generate a picture of a golden shaded cat on the grass," or "Find an ancient poem describing spring scenery." When the remote server receives the task, it selects an appropriate large language model for inference based on the task, obtains a response, and then sends it back to the UE. For example, the response could be multiple images or the text of a relevant ancient poem.

[0005] To fully utilize the computing resources on the edge, network, or application server sides, a model segmentation approach is often used during computation tasks. Different parts of the large language model are executed by the edge, network, and / or application server sides. For example, the large language model might consist of two parts: sub-model A and sub-model B. For an image recognition task, the edge uses sub-model A to process the input image, obtaining intermediate results. The edge transmits these intermediate results to the network side, which then uses sub-model B to process the intermediate results, obtaining the final image recognition result. The network side then feeds this result back to the edge. For this approach, meeting the end-to-end quality of service (QoS) requirements of the computation task is a crucial issue that needs to be addressed. Summary of the Invention

[0006] This application provides a communication method, related equipment, and communication system. Using this application is beneficial for meeting the end-to-end QoS requirements of computing tasks when the UE and computing nodes use a model to segment and process computing tasks.

[0007] Firstly, embodiments of this application provide a communication method. This method can be applied to a first communication device (or, as can be expressed, the method can be executed by the first communication device), wherein the first device can be an access network device or a communication module / processing module within the access network device, or a circuit or chip within the access network device (such as a modem chip, also known as a baseband chip, or a system-on-a-chip (SoC) chip containing a modem core, or a system-in-package (SIP) chip, etc.), or a module or software capable of implementing all or part of the functions of the access network device; alternatively, the first communication device can be a network device, or a module within the network device (e.g., a module, circuit, chip, or chip system, etc.), or a logical node, logical module, or software capable of implementing all or part of the functions of the network device. For ease of description, the following explanation will use an access network device as an example of the first communication device. The method includes:

[0008] The system receives first information from the UE, compute node, or user plane function (UPF) network element. The first information is determined based on the information obtained by the UE and compute node or application server using the first model to segment the first computing task. Based on the first information, the system determines the first QoS file and schedules the data packets corresponding to the first computing task based on the QoS flow corresponding to the first QoS file.

[0009] The information used by the UE and / or computing node (or application server) to segment the first computing task using the first model includes one or more of the following: the segmentation point of the first model, the termination point of the first model on the UE or the starting point on the computing node, the segmentation level of the first model, the segmentation grade of the first model, the segmentation identifier of the first model, the model layer identifier of the first model, and end-to-cloud collaboration mode information.

[0010] It can be seen that when determining the first QoS file, the information of the UE and / or computing node (or application server) using the first model to segment the first computing task is taken into account, so that when the access network device uses the QoS flow corresponding to the first QoS file to schedule the data packets corresponding to the first computing task, it can meet the end-to-end QoS requirements of the computing task.

[0011] In conjunction with the first aspect, in one possible implementation, the method of this embodiment further includes:

[0012] Receive multiple QoS files from management function network elements, and a first correspondence between the information of each QoS file and multiple model segmentation information; the first information includes the first segmentation information of the first model from the UE, computing node or application server;

[0013] Determining the first QoS file based on the first information includes: determining the information of the first QoS file based on the first segmentation information of the first model and the aforementioned first correspondence relationship, and determining the first QoS file from multiple QoS files based on the information of the first QoS file.

[0014] The QoS file information includes either the QoS file index or the QoS file identifier.

[0015] It can be seen that configuring multiple QoS files on the access network device, and the correspondence between the information of each QoS file and the multiple model segmentation information, enables the access network device to determine the first QoS file based on the first segmentation information of the first model from the UE. This allows the access network device to use the QoS flow corresponding to the first QoS file to schedule the data packets corresponding to the first computing task, so as to meet the end-to-end QoS requirements of the computing task.

[0016] In conjunction with the first aspect, in one possible implementation, the first information includes information about the first QoS file. The information about the first QoS file is determined by the UE based on the first segmentation information of the first model and the first correspondence between the information of each QoS file in the multiple QoS files and the segmentation information of the multiple models. The method of this embodiment further includes: receiving multiple QoS files from the management function network element.

[0017] Determining a first QoS file based on first information includes: determining the first QoS file from multiple QoS files based on information from the first QoS file.

[0018] As can be seen, configuring multiple QoS files on the first access network device enables the access network device to determine the first QoS file from among the multiple QoS files after receiving the information of the first QoS file. This allows the access network device to schedule the data packets corresponding to the first computing task using the QoS flow corresponding to the first QoS file, thereby meeting the end-to-end QoS requirements of the computing task. Furthermore, this method does not require the access network device to determine the information of the first QoS file, reducing the workload of the access network device.

[0019] In conjunction with the first aspect, in one possible implementation, the first segmentation information of the first model is determined by the UE based on the air interface state information between the UE and the access network equipment.

[0020] It can be seen that by selecting the first segmentation information of the first model through the air interface status information, the data packets generated by the UE and / or computing node using the first model to segment the first computing task can be transmitted between the UE and the access network equipment according to certain parameter requirements.

[0021] In conjunction with the first aspect, in one possible implementation, the first segmentation information of the first model is determined by the computing node based on one or more of the following: the computing node's load status, the interaction information between the computing node and the UE's application layer, and the air interface status information received by the computing node from the access network equipment side.

[0022] It can be seen that when determining the first segmentation information of the first model, the computing node considers its load status, ensuring that its computing resources are sufficient to execute the corresponding portion of the first computing task when the UE and / or the computing node subsequently use the first model to segment the first computing task. The computing node negotiates with the UE's application layer to determine the first segmentation information, ensuring that the UE and / or the computing node can successfully execute the first computing task when using the first model to segment it. The computing node also considers the air interface status information on the access network device side when determining the first segmentation information, ensuring that data packets generated by the UE and / or the computing node using the first model to segment the first computing task can be transmitted between the UE and the access network device. Furthermore, the first segmentation information of the first model is confirmed by the computing node, which helps reduce the workload of the UE.

[0023] In conjunction with the first aspect, in one possible implementation, the air interface state information includes one or more of the following: congestion state information, data rate, bandwidth, latency, and / or wireless channel quality.

[0024] In conjunction with the first aspect, in one possible implementation, the method of this embodiment further includes:

[0025] Receive data packets from the UE or computing node corresponding to the first computing task. The data packets corresponding to the first computing task are data packets generated by the UE or computing node using the first model to execute the first computing task based on the first segmentation information.

[0026] Secondly, embodiments of this application provide a communication method. This method should be applicable to a second communication device (or, as can be expressed, the method can be executed by the second communication device), wherein the second communication device can be a management function network element or a communication module / processing module within a management function network element, or a circuit or chip within a management function network element (such as a modem chip, also known as a baseband chip, or a SoC chip or SIP chip containing a modem core, etc.), or a module or software capable of implementing all or part of the functions of the management function network element; alternatively, the second communication device can be a network device, or a module within a network device (e.g., a module, circuit, chip, or chip system, etc.), or a logical node, logical module, or software capable of implementing all or part of the functions of the network device. The management function network element can be a session management function (SMF) network element or a compute management function (CMF) network element or other network elements capable of implementing the following functions. For ease of explanation, the following description will use the second communication device as a management function network element as an example. The method includes:

[0027] The system acquires multiple computational information from the UE, compute nodes, and / or application servers. Each piece of computational information corresponds to a model segmentation information. The model segmentation information includes a first segmentation information of a first model. The first segmentation information of the first model is information used by the UE and / or compute nodes to segment the first computational task using the first model, or information used by the UE and compute nodes to segment the first model to perform computation on the first computational task. The system determines multiple QoS files based on the multiple computational information. The system sends multiple QoS files to the access network device. The multiple QoS files include the first QoS file. The QoS stream corresponding to the first QoS file is used by the access network device to schedule data packets corresponding to the first computational task.

[0028] It can be seen that the management function network element sets up multiple QoS files corresponding to multiple model segmentation information and sends these multiple QoS files to the access network device, so that the access network device can select the first QoS file from the multiple QoS files based on the information fed back by the UE. In this way, when the access network device schedules the data packets corresponding to the first computing task using the QoS flow corresponding to the first QoS file, it can meet the end-to-end QoS requirements of the computing task.

[0029] In one possible implementation, each of the multiple QoS files includes one or more of the following: guaranteed flow bit rate (GFBR), maximum flow bit rate (MFBR), maximum data burst volume (MDBV), and delay budget. The calculation information includes the computational delay corresponding to the model or the transmission output data volume corresponding to the model. The delay budget includes the packet delay budget (PDB) or the computational task transmission delay budget.

[0030] One possible interpretation is that the packet delay budget is used to indicate the upper limit of the delay for any data packet corresponding to the computing task to be transmitted or completed between the UE and the UPF network element or computing node; the computing task transmission delay budget is used to indicate the upper limit of the delay for multiple data packets corresponding to the computing task to be transmitted or completed between the UE and the UPF network element or computing node; the computing delay is used to indicate one or more of the maximum delay, average delay, P95 delay, or service level agreement (SLA) delay corresponding to the model executing the computing task; and the transmission output data volume corresponding to the model is used to indicate the average or maximum data volume or burst data volume corresponding to the model's output results.

[0031] In conjunction with the second aspect, in one possible implementation, each of the multiple QoS files includes a latency budget, and the calculation information includes the computation latency corresponding to the model. Multiple QoS files are determined based on the computation information corresponding to the multiple model segmentation information, including:

[0032] The latency budget included in the QoS file is determined based on the UE's quality of experience (QoE) latency and the computational latency corresponding to the model.

[0033] One possible interpretation is that the UE's QoE latency indicates the upper limit of the latency from when the UE sends a computation task to when it receives a response. The UE's QoE latency includes the RTT / round-trip latency of the UE's computation task transmitted between the UE and the compute node or application server, and the computation latency between the UE and the compute node or application server.

[0034] It can be seen that by introducing a latency budget based on QoE latency and the computational latency corresponding to the model into the QoS file, the subsequent management function network elements can meet the UE's QoE latency when scheduling the data packets corresponding to the QoS flow computation task based on the QoS file.

[0035] In conjunction with the second aspect, in one possible implementation, the method of this embodiment further includes:

[0036] Send a mapping rule to the access network device. This mapping rule indicates the first correspondence between the information of each QoS file in multiple QoS files and the multiple model segmentation information.

[0037] It can be seen that by sending the first correspondence to the access network device, the access network device can determine the information of the first QoS file based on the first segmentation information of the first model sent by the UE, and then determine the first QoS file from multiple QoS files based on the information of the first QoS file.

[0038] In conjunction with the second aspect, in one possible implementation, the mapping rule is also used to instruct the access network device to schedule the data packets corresponding to the first computation task based on the QoS flow corresponding to the first QoS file determined according to the first segmentation information of the first model determined by the UE.

[0039] It can be seen that by instructing the access network device to schedule the data packets corresponding to the first computing task based on the QoS flow corresponding to the first QoS file determined according to the first segmentation information of the first model, the access network device can meet the end-to-end QoS requirements of the computing task when scheduling the data packets corresponding to the first computing task using the QoS flow corresponding to the first QoS file. Furthermore, by implicitly indicating through mapping rules, it is not necessary to send additional indication information to the access network device, which helps to reduce the amount of data transmitted and thus reduce transmission resource overhead.

[0040] In conjunction with the second aspect, in one possible implementation, the method of this embodiment further includes:

[0041] Send a first indication message to the UE, the first indication message being used to instruct the UE to send the first segmentation information of the first model to the access network device.

[0042] It can be seen that by instructing the UE to send the first segmentation information of the first model to the access network device, the access network device can determine the first QoS file identifier or index based on the first segmentation information of the first model, and then select the first QoS file from multiple QoS files based on the information of the first QoS file. This enables the access network device to meet the end-to-end QoS requirements of the computing task when scheduling the data packets corresponding to the first computing task using the QoS flow corresponding to the first QoS file.

[0043] In conjunction with the second aspect, in one possible implementation, the method of this embodiment further includes:

[0044] A second correspondence is established between multiple QoS flow information and multiple QoS file information sent to the access network device.

[0045] The first QoS flow information corresponds to the first segmentation information of the first model, and the second correspondence is used by the access network device to determine the first QoS file based on the first QoS flow information. There is a correspondence between the first QoS flow information and the first QoS file.

[0046] In one example, the QoS flow information is the QoS flow identifier (QFI), but it can also be other information that can be used to identify QoS flows, which is not limited here.

[0047] In conjunction with the second aspect, in one possible implementation, the method of this embodiment further includes:

[0048] Send QoS rules to the UE. The QoS rules are used to indicate the third correspondence between multiple QoS flow information and multiple model segmentation information and / or to instruct the UE to determine the first QoS flow information based on the first segmentation information and the third correspondence of the first model, and to map the data packets corresponding to the first computing task to the first QoS flow for transmission based on the first QoS flow information.

[0049] It can be seen that by sending QoS rules to the UE, the UE can determine the first QoS information corresponding to the first segmentation information of the first model based on the first segmentation information of the first model and the third correspondence indicated by the QoS rules. Then, the UE maps the data packets corresponding to the first computing task to the first QoS flow for transmission based on the first QoS flow information. By sending the second correspondence to the access network device, the access network device can determine the first QoS file based on the first QoS flow information and the second correspondence. This enables the access network device to meet the end-to-end QoS requirements of the computing task when scheduling the data packets corresponding to the first computing task using the QoS flow corresponding to the first QoS file.

[0050] In conjunction with the second aspect, in one possible implementation, the method of this embodiment further includes:

[0051] The UE is sent a mapping rule and a second indication information. The mapping rule is used to indicate the first correspondence between the information of each QoS file of the multiple QoS files and the multiple model segmentation information. The second indication information is used to instruct the UE to send the information of the first QoS file to the access network device. The information of the first QoS file is determined by the UE based on the first segmentation information of the first model and the first correspondence.

[0052] It can be seen that by sending mapping rules and second indication information to the UE, the UE can determine the information of the first QoS file based on the first segmentation information of the first model and send the information of the first QoS file to the access network device. This enables the access network device to select the first QoS file from multiple QoS files based on the information of the first QoS file. When the access network device schedules the data packets corresponding to the first computing task using the QoS flow corresponding to the first QoS file, it can meet the end-to-end QoS requirements of the computing task.

[0053] In conjunction with the second aspect, in one possible implementation, the method of this embodiment further includes:

[0054] Send a third instruction message to the device. The third instruction message is used to instruct the access network device to schedule the data packets corresponding to the first computing task based on the QoS flow corresponding to the first QoS file determined by the first segmentation information of the first model from the UE or the computing node.

[0055] In conjunction with the second aspect, in one possible implementation, the method of this embodiment further includes:

[0056] A fourth instruction message is sent to the computing node, which instructs the computing node to send the first segmentation information of the first model to the access network device or UE.

[0057] It can be seen that by instructing the computing node to send the first segmentation information of the first model to the access network device, the UE does not need to report, which reduces the workload of the UE and reduces the transmission overhead.

[0058] In conjunction with the second aspect, in one possible implementation, the method of this embodiment further includes:

[0059] A fifth indication message is sent to the user plane function network element. This fifth indication message is used by the user plane function network element to send the first segmentation information of the first model to the access network device or UE. In one possible implementation, the user plane function network element identifies the first segmentation information of the first model sent by the application server or computing node, and adds it to the General Packet Radio Service (GPRS) tunneling protocol for the user plane (GTP-U) header of the data packet before sending it to the access network device.

[0060] It can be seen that by instructing the user plane function network element to send the first segmentation information of the first model to the access network device, the UE does not need to report, which reduces the workload of the UE and reduces the transmission overhead.

[0061] In conjunction with the second aspect, in one possible implementation, the first segmentation information of the first model is determined by the computing node based on one or more of the following: the load status of the computing node, the interaction information between the computing node and the UE's application layer, and the air interface status information received by the computing node from the access network device. Alternatively, the first segmentation information of the first model is determined by the UE based on the air interface status information between the UE and the access network device or the indication information of the computing node.

[0062] It can be seen that when determining the first segmentation information of the first model, the computing node considers its load status, ensuring that its computing resources are sufficient to execute the corresponding portion of the first computing task when the UE and / or computing node subsequently use the first model to segment the first computing task. The computing node negotiates with the UE's application layer to determine the first segmentation information, ensuring that the UE and / or computing node can successfully execute the first computing task when using the first model to segment it. When determining the first segmentation information of the first model, air interface status information is considered, ensuring that data packets generated by the UE and / or computing node in segmenting the first computing task using the first model can be transmitted between the UE and the access network equipment. Furthermore, when the first segmentation information of the first model is determined by the computing node, it reduces the workload of the UE. Conversely, when the first segmentation information of the first model is determined by the UE, it reduces the workload of the computing node.

[0063] In conjunction with the second aspect, in one possible implementation, the air interface state information includes one or more of the following: congestion state information, data rate, bandwidth, latency, and / or wireless channel quality.

[0064] Thirdly, embodiments of this application provide a communication method. This method is applied to a third communication device (or, as can be expressed, the method can be executed by a third communication device), wherein the third device can be a UE or a communication module / processing module in the UE, or a circuit or chip in the UE (such as a modem chip, also known as a baseband chip, or a SoC chip or SIP chip containing a modem core, etc.), or a module or software capable of implementing all or part of the UE's functions; alternatively, the third communication device can be a network device, or a module in a network device (e.g., a module, circuit, chip, or chip system, etc.), or a logical node, logical module, or software capable of implementing all or part of the network device's functions. For ease of explanation, the following description uses a UE as an example of the third communication device. The method includes:

[0065] The first information is determined based on information obtained by the UE and / or computing node or application server using the first model to segment the first computing task; the first information is sent to the access network device to determine the QoS file of the QoS stream for transmitting the data packets corresponding to the first computing task.

[0066] It can be seen that by having the UE feed back the first information to the access network device, the access network device can determine the first QoS file based on the first information, and when the access network device schedules the data packets corresponding to the first computing task using the QoS flow corresponding to the first QoS file, it can meet the end-to-end QoS requirements of the computing task.

[0067] In conjunction with the third aspect, in one possible implementation, the first information includes the first segmentation information of the first model.

[0068] In conjunction with the third aspect, in one possible implementation, the first information includes information of the first QoS file. The method of this embodiment includes: the UE receiving a mapping rule from the management function network element; the mapping rule is used to indicate a first correspondence between the information of each QoS file in the multiple QoS files and the multiple model segmentation information.

[0069] Determining the first information includes: determining the first segmentation information of the first model; determining the information of the first QoS file based on the first segmentation information of the first model and the first correspondence relationship, wherein the information of the first QoS file has a correspondence relationship with the first segmentation information of the first model, and the first QoS file is used to schedule the data packets corresponding to the first computing task.

[0070] It can be seen that by having the UE determine the information of the first QoS file based on the first segmentation information of the first model and send the information of the first QoS file to the access network device, the access network device can select the first QoS file from multiple QoS files based on the information of the first QoS file. When the access network device schedules the data packets corresponding to the first computing task using the QoS flow corresponding to the first QoS file, it can meet the end-to-end QoS requirements of the computing task. Furthermore, the access network device can determine the first QoS file based on the information of the first QoS file without performing other operations, which helps to reduce the workload of the access network device.

[0071] In conjunction with the third aspect, in one possible implementation, the first segmentation information of the first model is determined by the computing node based on one or more of the following: the computing node's load status, the interaction information between the computing node and the UE's application layer, and the air interface status information received by the computing node from the access network device. Alternatively, the first segmentation information of the first model is determined by the UE based on the air interface status information between the UE and the access network device or the indication information sent by the computing node.

[0072] It can be seen that when determining the first segmentation information of the first model, the computing node considers its load status, ensuring that its computing resources are sufficient to execute the corresponding portion of the first computing task when the UE and / or computing node subsequently use the first model to segment the first computing task. The computing node negotiates with the UE's application layer to determine the first segmentation information, ensuring that the UE and / or computing node can successfully execute the first computing task when using the first model to segment it. When determining the first segmentation information of the first model, air interface status information is considered, ensuring that data packets generated by the UE and / or computing node in segmenting the first computing task using the first model can be transmitted between the UE and the access network equipment. Furthermore, when the first segmentation information of the first model is determined by the computing node, it reduces the workload of the UE. Conversely, when the first segmentation information of the first model is determined by the UE, it reduces the workload of the computing node.

[0073] In conjunction with the third aspect, in one possible implementation, the air interface state information includes one or more of the following: congestion state information, data rate, bandwidth, latency, and / or wireless channel quality.

[0074] In conjunction with the third aspect, in one possible implementation, the method of this embodiment further includes:

[0075] The system receives a first indication message from a management function network element, the first indication message being used to instruct the UE to send the first segmentation information of the first model to the access network device.

[0076] It can be seen that by instructing the UE to report the first segmentation information of the first model, there is no need for the computing node to report, which helps to reduce the workload of the UE.

[0077] In conjunction with the third aspect, in one possible implementation, the method of this embodiment further includes:

[0078] The system receives a second instruction message from a management function network element. The second instruction message is used to instruct the UE to send information about a first QoS file to the access network device.

[0079] As can be seen, by instructing the UE to send the information of the first QoS file to the access network device, the access network device can determine the first QoS file based on the information of the first QoS file without performing other operations, which helps to reduce the workload of the access network device.

[0080] In conjunction with the third aspect, in one possible implementation, the method of this embodiment further includes:

[0081] Send the data packet corresponding to the first computing task to the access network device. The data packet corresponding to the first computing task is the data packet generated by the UE using the first model to execute the first computing task based on the first segmentation information.

[0082] Fourthly, embodiments of this application provide a communication device, characterized in that the communication device includes at least one processor, the at least one processor being coupled to at least one memory, the at least one processor being configured to execute a computer program or instructions stored in the at least one memory, so that the communication device performs the method provided by the first aspect or any possible implementation of the first aspect, or the method provided by the second aspect or any possible implementation of the second aspect, or the method provided by the third aspect or any possible implementation of the third aspect.

[0083] Fifthly, embodiments of this application provide a communication device, including a processor and a memory. The memory is used to store program code. The processor is used to invoke the program code stored in the memory to execute the method provided by the first aspect or any possible implementation of the first aspect, or the method provided by the second aspect or any possible implementation of the second aspect, or the method provided by the third aspect or any possible implementation of the third aspect.

[0084] In a sixth aspect, embodiments of this application provide a computer storage medium including computer instructions that, when executed on an electronic device, cause the electronic device to perform the method provided by the first aspect or any possible implementation of the first aspect, or the method provided by the second aspect or any possible implementation of the second aspect, or the method provided by the third aspect or any possible implementation of the third aspect.

[0085] In a seventh aspect, embodiments of this application provide a computer program product that, when run on a computer, causes the computer to perform a method as provided in the first aspect or any possible implementation of the first aspect, or a method as provided in the second aspect or any possible implementation of the second aspect, or a method as provided in the third aspect or any possible implementation of the third aspect.

[0086] Eighthly, embodiments of this application also provide a communication method applied to a communication system, the communication system including an access network device, a management function network element, a computing node, and a UE. The method includes:

[0087] The management function network element obtains multiple computing information of the UE and / or computing node. Each computing information corresponds to a model segmentation information. The model segmentation information corresponding to the multiple computing information includes the first segmentation information of the first model. The first segmentation information of the first model is the information of the UE and computing node to segment the first computing task using the first model.

[0088] The management function network element determines multiple QoS files based on multiple calculation information;

[0089] Management function network elements send multiple QoS files to access network devices;

[0090] The UE or computing node sends the first information to the access network device. The first information is determined by the first segmentation information of the first model.

[0091] The access network device determines the QoS file based on the first information and multiple QoS files; and schedules the data packets corresponding to the first computing task based on the QoS flow corresponding to the first QoS file.

[0092] In a ninth aspect, embodiments of this application also provide a communication system, which includes an access network device, a management function network element, a computing node, and a UE;

[0093] The management function network element is used to obtain multiple computing information of UE and / or computing node. Each computing information corresponds to a model segmentation information. The model segmentation information corresponding to the multiple computing information includes the first segmentation information of the first model. The first segmentation information of the first model is the information of UE and / or computing node segmenting the first computing task using the first model.

[0094] The management function network element is also used to determine multiple QoS files based on multiple calculation information;

[0095] The management function network element is also used to send multiple QoS files to access network devices;

[0096] The UE or computing node is used to send first information to the access network device. The first information is determined based on the first segmentation information of the first model.

[0097] The access network device is used to determine a QoS file based on first information and multiple QoS files; and to schedule data packets corresponding to the first computing task based on the QoS flow corresponding to the first QoS file.

[0098] It is understood that the beneficial effects of the embodiments described in aspects four through nine can be referred to the beneficial effects of the foregoing methods, and will not be repeated here. Attached Figure Description

[0099] Figure 1 is a schematic diagram of a communication system architecture provided in an embodiment of this application;

[0100] Figure 2 is a flowchart illustrating a communication method provided in an embodiment of this application;

[0101] Figure 3 is a flowchart illustrating another communication method provided in an embodiment of this application;

[0102] Figure 4 is a flowchart illustrating another communication method provided in an embodiment of this application;

[0103] Figure 5 is an interactive flowchart of a communication method provided in an embodiment of this application;

[0104] Figure 6 is an interactive flowchart of another communication method provided in an embodiment of this application;

[0105] Figure 7 is a schematic diagram of the structure of an access network device provided in an embodiment of this application;

[0106] Figure 8 is a schematic diagram of the structure of a management function network element provided in an embodiment of this application;

[0107] Figure 9 is a schematic diagram of the structure of a UE provided in an embodiment of this application;

[0108] Figure 10 is a schematic diagram of the structure of a communication device provided in an embodiment of this application. Detailed Implementation

[0109] The terms “first,” “second,” “third,” and “fourth,” etc., used in the specification, claims, and drawings of this application are used to distinguish different objects, not to describe a specific order.

[0110] "Multiple" refers to two or more. "And / or" describes the relationship between related objects, indicating three possible relationships. For example, A and / or B means: A exists alone, A and B exist simultaneously, or B exists alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0111] The embodiments of this application will now be described with reference to the accompanying drawings.

[0112] Referring to Figure 1, Figure 1 is a schematic diagram of a communication system architecture provided in an embodiment of this application. This communication system is a schematic diagram of the network architecture of 5th Generation Mobile Networks (5G). It includes a radio access network (RAN), which can be represented as two parts: RAN equipment and a core network (CN). The RAN equipment is used to provide network access functionality for authorized user equipment (UE) in a specific area and can use transmission tunnels of different qualities according to the UE's level and service requirements. For example, the RAN equipment can manage radio resources, provide access services to the UE, and thus complete the forwarding of control information and / or data information between the UE and the CN.

[0113] To facilitate understanding of the embodiments of this application, an application scenario of the embodiments of this application will be described in detail first with reference to FIG1.

[0114] 1. User equipment (UE): This can be referred to as terminal equipment, terminal, access terminal, user unit, user station, mobile station, mobile station, remote station, remote terminal, mobile device, user terminal, wireless communication equipment, user agent, or user device. Terminal equipment can also be a cellular phone, cordless phone, session initiation protocol (SIP) phone, wireless local loop (WLL) station, personal digital assistant (PDA), handheld device with wireless communication capabilities, computing device or other processing device connected to a wireless modem, vehicle-mounted device, drone, wearable device, terminal equipment in a 5G network, or terminal equipment in an evolved public land mobile network (PLMN), etc., and this application embodiment does not limit this.

[0115] 2. Access Network (AN): Provides network access for authorized users in a specific area and can use transmission tunnels of different qualities depending on the user's level and service requirements. Access networks can employ different access technologies. Current access network technologies include: radio access network technologies used in 3G systems, radio access network technologies used in 4G systems, or next-generation radio access network (NG-RAN) technologies (such as those used in 5G systems).

[0116] An access network that uses wireless communication technology to implement access network functions can be called a radio access network (RAN). A RAN manages radio resources, provides access services to terminals, and facilitates the forwarding of control signals and user data between terminals and the core network.

[0117] Wireless access network equipment can be, for example, a base station (NodeB), an evolved NodeB (eNB or eNodeB), a next-generation Node base station (gNB) in a 5G mobile communication system, a base station in a future mobile communication system, or an access point (AP) in a Wi-Fi hotspot system. It can also be a wireless controller in a cloud radio access network (CRAN) scenario, or it can be a relay station, access point, vehicle-mounted equipment, drone, wearable device, or network equipment in a 5G network or an evolved PLMN. This application does not limit the specific technology or equipment form used in the wireless access network equipment.

[0118] As shown in Figure 1, the RAN equipment includes CUF network elements and CCF network elements.

[0119] 3. Access Management Network Element: Primarily used for mobility management and access management, responsible for transmitting user policies between user equipment and PCF network elements, etc. It can be used to implement other functions of the Mobile Management Entity (MME) besides session management. For example, access authorization (authentication) function.

[0120] In 5G communication systems, the access management network element can be an access and mobility management function (AMF) network element. In future communication systems, the access management network element can still be an AMF network element, or it can have other names; this application does not limit this.

[0121] 4. Management Function Network Elements: These are mainly used for session management, allocation and management of Internet Protocol (IP) addresses for user equipment, selection of manageable user plane functions, policy control and billing function interface endpoints, and downlink data communication. Management function network elements include SMF (Software-Defined Function) network elements or CMF (Computational Function) network elements. SMF network elements manage regular sessions, while CMF network elements manage computing sessions.

[0122] In 5G communication systems, management function network elements can be either SMF (Supervisory Management Function) or CMF (Computational Management Function) network elements. In future communication systems, management function network elements may still be SMF or CMF network elements, or they may have other names; this application does not limit this. Management function network elements include SMF or CMF network elements. SMF network elements manage regular sessions, while CMF network elements manage computational sessions.

[0123] 5. User plane network elements: used for packet routing and forwarding, QoS processing of user plane data, completion of user plane data forwarding, session / flow-level billing statistics, bandwidth limiting functions, etc.

[0124] In the system shown in Figure 1, the user plane network elements include a first UPF network element and a second UPF. The first UPF network element supports the UE's regular sessions, which are used for data transmission between the UE and the DN (e.g., EAS). That is, the first UPF network element forwards regular service flows between the UE and the DN. This session can be a PDU session. The first UPF network element can be used to implement user plane-related functions such as packet routing and transmission, packet inspection, service usage reporting, quality of service (QoS) processing, legality monitoring, uplink packet inspection, and downlink packet storage. For example, during the establishment of a PDU session, the SMF network element selects the first UPF network element for the session. This first UPF network element can be called the session anchor, such as the PDU session anchor (PSA). It can also be called the anchor UPF network element or the remote session anchor.

[0125] The second UPF network element supports the UE's computing session, which is used for data transmission between the UE and computing nodes. Specifically, the second UPF network element forwards computing service flows between the UE and the computing nodes. The second UPF network element can be used to implement user plane related functions such as packet routing and transmission, packet inspection, service usage reporting, QoS processing, legality monitoring, uplink packet inspection, and downlink packet storage.

[0126] In 5G communication systems, user plane network elements can be UPF network elements. In future communication systems, user plane network elements can still be UPF network elements, or they can have other names; this application does not limit this.

[0127] 6. Data network element: A network used to provide data transmission.

[0128] In 5G communication systems, data network elements can be data network (DN) elements. In future communication systems, data network elements may still be DN elements, or they may have other names; this application does not limit this. DN elements, also known as packet data networks (PDNs), are typically networks located outside the operator's network, such as third-party networks. A DN element may include one or more EASs, which provide local services to the UE by transmitting data with the UE.

[0129] 7. Policy control network element: A unified policy framework used to guide network behavior, providing policy rule information to control plane functional network elements (such as AMF, SMF, etc.).

[0130] In 4G communication systems, this policy control network element can be a Policy and Charging Rules Function (PCRF) network element. In 5G communication systems, this policy control network element can be a PCF network element. In future communication systems, this policy control network element can still be a PCF network element, or it can have other names; this application does not limit its scope.

[0131] 8. Data management network element: used to handle user equipment identification, access authentication, registration, and mobility management, etc.

[0132] In 5G communication systems, this data management network element can be a unified data management (UDM) network element; in 4G communication systems, this data management network element can be a home subscriber server (HSS) network element. In future communication systems, the data management network element can still be a UDM network element, or it can have other names; this application does not limit this.

[0133] 9. Network Exposure Function (NEF) Element: Used to securely expose services and capabilities provided by the 3rd Generation Partnership Project (3GPP) network functions to the outside world.

[0134] 10. Application Function (AF) Network Element: Provides a specific application layer service to the UE. When providing services to the UE, the AF has requirements for QoS and charging policies and needs to notify the network. Simultaneously, the AF also needs to obtain application-related information from the core network. The AF can possess all the functions defined in the technical specification (TS) 23.701R-15, as well as related functions for application services. That is, in the user plane architecture, the application server (AS) and the UE communicate in the user plane via the UE-RAN-UPF-AS path. The AF can also communicate with other network function (NF) network elements in the 5G core network (5GC) in the control plane architecture via the NEF. For example, it can communicate with the PCF network element via the NEF network element. If the AF is deployed by the 5GC operator, the AF network element can also communicate directly with other NF network elements in the 5GC in the control plane architecture without going through the NEF network element, such as directly communicating with the PCF network element.

[0135] 11. Billing processing function network element: mainly responsible for providing users with calculation quotas or traffic configurations, and generating user billing bills based on the user's traffic consumption or calculation information.

[0136] 12. Computing Node: A node in the network used to process computing tasks. This computing node can be independent of other network devices in Figure 1, such as being located after the second UPF network element or after the RAN device; this computing node can also be deployed within the network devices in Figure 1, such as being deployed in the second UPF network element or in the RAN device.

[0137] It should be understood that the network architecture described above for the embodiments of this application is merely an example, and the network architecture applicable to the embodiments of this application is not limited thereto. Any network architecture capable of realizing the functions of the above-described network elements is applicable to the embodiments of this application.

[0138] It should also be understood that the AMF, SMF, UPF, NEF, PCF, UDM, and billing function network elements shown in Figure 1 can be understood as network elements in the core network used to implement different functions, such as network slices that can be combined as needed. These core network elements can be independent devices or integrated into the same device to implement different functions. This application does not limit the specific form of the above network elements.

[0139] It should also be understood that the above naming is defined only for the convenience of distinguishing different functions and should not constitute any limitation on this application. This application does not exclude the possibility of using other naming in 5G networks and other future networks. For example, in 6G networks or future networks, some or all of the above-mentioned network terms may be used, or other names may be used. The interface names between the various network elements in Figure 1 are just examples, and the interface names in specific implementations may be other names, which this application does not specifically limit. In addition, the names of the messages (or signaling) transmitted between the above-mentioned network elements are also just examples and do not constitute any limitation on the function of the messages themselves.

[0140] Referring to Figure 2, which is a flowchart illustrating a communication method provided in an embodiment of this application, the method is applied to a first communication device (or can be expressed as the method being executed by the first communication device). This first communication device can be an access network device or a communication module / processing module within the access network device, or a circuit or chip within the access network device (such as a modem chip, also known as a baseband chip, or a SoC chip or SIP chip containing a modem core, etc.), or a module or software capable of implementing all or part of the functions of the access network device; alternatively, the first communication device can be a network device, or a module within the network device (e.g., a module, circuit, chip, or chip system, etc.), or a logical node, logical module, or software capable of implementing all or part of the functions of the network device. For ease of explanation, the following description will use an access network device as an example, which can be the RAN device in the system shown in Figure 1. As shown in Figure 2, the method includes:

[0141] S201. The access network device receives the first information from the UE or computing node.

[0142] The first information is determined based on the information obtained by the UE and / or computing node in segmenting the first computing task using the first model. For example, the UE or computing node determines the first information based on the information obtained by the UE and / or computing node in segmenting the first computing task using the first model.

[0143] Among them, the information used by the UE and / or computing node to segment the first computing task using the first model can be the first segmentation information of the first model, such as the segmentation point of the first model, the termination point of the first model on the UE or the starting point on the computing node, the segmentation level of the first model, the segmentation grade of the first model, the segmentation identifier of the first model, the model layer identifier of the first model, and the end-cloud collaboration mode information, etc.

[0144] The split point of the first model indicates the model splitting information of the first model on the UE side. This information is used by the UE to determine the operating layer of the first model on the UE side when performing computation tasks. For example, assuming that split point A divides the first model into two parts, a and b, the UE uses part a of the first model to perform computation tasks and obtain a response. The computing node or application server uses part b of the first model to process the response sent by the UE to perform computation tasks.

[0145] The segmentation level of the first model indicates the segmentation level or identifier of the model. For example, for the first model, there are three different segmentation levels: A, B, and C. Different segmentation levels correspond to different segmentation methods of the model. Assuming that segmentation level A is selected, the UE uses the part of the model corresponding to level A to perform computational tasks and obtain a response. The computing node or application server uses the other part of the model corresponding to level A to process the response sent by the UE to perform computational tasks.

[0146] In this model, the layer number identifier indicates the number of layers in the first model. For example, if the first model has 100 layers, the layer number identifier indicates the first 20 layers. The UE uses the first 20 layers to perform computational tasks and obtain a response, while the computing node or application server uses the last 80 layers to process the response sent by the UE and perform computational tasks. Alternatively, the layer number identifier might indicate the last 70 layers. The UE uses the first 30 layers to perform computational tasks and obtain a response, while the computing node or application server uses the last 70 layers to process the response sent by the UE and perform computational tasks.

[0147] The edge-cloud collaboration mode information indicates the collaborative method of computing tasks between the UE and the computing node or application server. For example, the edge-cloud collaboration mode can indicate full offloading, in which case the computing node or application server performs all computing tasks, and the UE is only responsible for sending the corresponding data packets and not for performing computing tasks. Another example is that the edge-cloud collaboration mode can indicate partial offloading, in which case the UE and the computing node or application server jointly complete the computing tasks. In one possible implementation, the edge-cloud collaboration mode can indicate the percentage of partial offloading, such as 20%, in which case the UE performs 20% of the computing tasks, and the computing node or application server performs 80% of the computing tasks. The termination point of the first model indicates the model termination information on the UE side, which the UE can use to determine the inference termination point of the first model on the UE side when performing computing tasks. For example, for the first model, assuming termination point A is the Nth layer of the first model, the UE uses the first N layers of the model to perform computing tasks and obtain a response, while the computing node or application server uses the model after the Nth layer to process the response sent by the UE to perform computing tasks.

[0148] Specifically, the UE and / or computing node use the first model to segment the first computing task. The first model can be divided into a first sub-model and a second sub-model. When processing the first computing task, the UE uses the first sub-model to process the input data corresponding to the first computing task to obtain intermediate results. The UE transmits the intermediate results to the computing node, and the computing node uses the second sub-model to process the intermediate results. The processing result obtained is the processing result of the first computing task.

[0149] In one example, the first segmentation information of the first model is used to indicate the segmentation position where the first model is segmented into a first sub-model and a second sub-model.

[0150] It should be understood that a model can have one or more partitioning methods, meaning that one model corresponds to one partitioning information. One model partitioning information corresponds to one computational information.

[0151] Optionally, the first information includes either the first segmentation information of the first model or the information of the first QoS file. The information of the first QoS file is determined by the UE based on the first segmentation information of the first model and the first correspondence between the information of each QoS file in the multiple QoS files and the segmentation information of the multiple models.

[0152] The QoS file information includes the QoS file index or the QoS file identifier. Of course, the QoS file information also includes information from other files used to identify QoS, which is not limited here.

[0153] Optionally, the first segmentation information from the first model of the UE is determined by the UE based on the air interface status information between the UE and the access network equipment, or the first segmentation information from the first model of the UE is determined by the computing node and fed back to the UE.

[0154] In one example, the computing node determines the first segmentation information of the first model in the following way:

[0155] The computing node determines the first segmentation information of the first model based on one or more of the following: the computing node's load status, the interaction information between the computing node and the UE's application layer, and the air interface status information received by the computing node from the access network equipment side.

[0156] The aforementioned air interface status information includes, but is not limited to, one or more of the following: congestion status information, data rate, bandwidth, latency, and / or wireless channel quality.

[0157] For example, the access network device provides the UE with the current available data rate, such as the available data rate. Based on this information and the amount of data to be transmitted in the calculation information corresponding to the segmentation, the UE determines that the current available data rate can transmit a larger amount of data with less transmission latency. Therefore, the UE selects the segmentation method that transmits a larger amount of data. Alternatively, the access network device provides the UE with current congestion information, such as the percentage of congested data packets or based on ECN marking. The UE determines that the current base station congestion is severe, and transmitting data packets will cause significant transmission latency. Therefore, the UE can choose the segmentation method that transmits a smaller amount of data.

[0158] S202, The access network device determines the first QoS file based on the first information.

[0159] In one possible implementation, the method of this embodiment further includes: the access network device receiving multiple QoS files from a management function network element, and a first correspondence between the information of each QoS file and multiple model segmentation information; the first information includes first segmentation information of a first model from a UE or a computing node;

[0160] The access network device determines the first QoS file based on the first information, including: the access network device determines the information of the first QoS file based on the first segmentation information of the first model and the aforementioned first correspondence relationship; and the access network device determines the first QoS file from multiple QoS files based on the information of the first QoS file.

[0161] Specifically, the management function network element sends multiple QoS files and mapping rules to the access network device. These mapping rules establish a first correspondence between the information in each QoS file and the segmentation information of multiple models. Optionally, the mapping rules instruct the access network device to determine a first QoS file based on the first segmentation information of the first model sent by the UE, and to schedule data packets corresponding to the first computation task based on the QoS flow corresponding to the first QoS file. Optionally, the management function network element also sends third indication information to the access network device. This third indication information instructs the access network device to determine the first QoS file based on the first segmentation information of the first model sent by the UE, and to schedule data packets corresponding to the first computation task based on the QoS flow corresponding to the first QoS file. The management function network element also sends first indication information to the UE, instructing the UE to send the first segmentation information of the first model to the access network device.

[0162] After receiving the first segmentation information of the first model from the UE or computing node, the access network device determines the information of the first QoS file based on the first segmentation information of the first model and the first correspondence relationship, and determines the first QoS file from multiple QoS files based on the information of the first QoS file.

[0163] It can be seen that configuring multiple QoS files on the access network device, and the correspondence between multiple model segmentation information and the information of multiple QoS files, enables the access network device to determine the information of the first QoS file based on the first segmentation information of the first model from the UE and to determine the first QoS file based on the information of the first QoS file. This enables the access network device to use the QoS flow corresponding to the first QoS file to schedule the data packets corresponding to the first computing task, so as to meet the end-to-end QoS requirements of the computing task.

[0164] In one possible implementation, the first information includes information about a first QoS file. The information about the first QoS file is determined by the UE based on the first segmentation information of the first model and the first correspondence between the information of each QoS file and the segmentation information of the multiple models. The method in this embodiment further includes: the access network device receiving multiple QoS files from the management function network element.

[0165] The access network device determines a first QoS file based on first information, including: the access network device determines the first QoS file from multiple QoS files based on information from the first QoS file.

[0166] Specifically, the management function network element sends multiple QoS files to the access network device and a mapping rule to the UE. The mapping rule indicates the aforementioned first correspondence. Optionally, the mapping rule also instructs the UE to send information about the first QoS file to the access network device. Optionally, the management function network element also sends second indication information to the UE, which instructs the UE to send information about the first QoS file to the access network device. After receiving the information about the first QoS file from the UE, the access network device determines the first QoS file from the multiple QoS files based on the information in the first QoS file.

[0167] As can be seen, configuring multiple QoS files on the access network device enables the device to determine the first QoS file from among the multiple QoS files after receiving information about the first QoS file. This allows the access network device to schedule data packets corresponding to the first computing task using the QoS flow corresponding to the first QoS file, thereby meeting the end-to-end QoS requirements of the computing task. Furthermore, this method eliminates the need for the access network device to determine the information of the first QoS file, reducing its workload.

[0168] S203. The access network device schedules the data packets of the first computing task based on the QoS flow corresponding to the first QoS file.

[0169] Specifically, the access network device receives data packets corresponding to a first computing task from a UE or a computing node. The data packets corresponding to the first computing task are generated by the UE or computing node executing the first computing task using a first model based on the first segmentation information. The access network device creates a QoS stream based on a first QoS file and maps the data packets corresponding to the first computing task onto this QoS stream.

[0170] The data packets for the first computing task scheduled by the access network equipment include a first data packet carrying intermediate results sent by the UE to the computing node and a second data packet carrying the processing results of the first computing task sent by the computing node to the UE.

[0171] In one example, the first QoS file includes, but is not limited to, one or more of the following: delay budget, MFBR, MDBV, and GFBR. The delay budget includes the packet delay budget (PDB) or the calculation task transmission delay budget.

[0172] One possible interpretation is that the packet delay budget is used to indicate the upper limit of the delay for any data packet corresponding to the computing task to be transmitted or completed between the UE and the UPF network element or computing node; the computing task transmission delay budget is used to indicate the upper limit of the delay for multiple data packets corresponding to the computing task to be transmitted or completed between the UE and the UPF network element or computing node; and the computing delay is used to indicate one or more of the maximum delay, average delay, P95 delay, or SLA delay corresponding to the model executing the computing task.

[0173] It can be seen that the determination of the first QoS file takes into account the information of the UE and the computing node (or application server) using the first model to segment the first computing task, ensuring that the access network device can meet the end-to-end QoS requirements of the computing task when scheduling the data packets corresponding to the first computing task using the QoS flow corresponding to the first QoS file. The selection of the first segmentation information of the first model through air interface status information ensures that the data packets generated by the UE and / or computing node using the first model to segment the first computing task can be transmitted between the UE and the access network device. When determining the first segmentation information of the first model, the computing node considers its load status, ensuring that its computing resources can meet the requirements for executing the portion of the first computing task when the UE and / or computing node subsequently use the first model to segment the first computing task. The computing node negotiates with the UE's application layer when determining the first segmentation information of the first model, ensuring that the UE and computing node can successfully execute the first computing task when the UE and / or computing node subsequently use the first model to segment the first computing task. When determining the first segmentation information of the first model, the computing node considers the air interface status information on the access network device side, enabling the data packets generated by the UE and / or the computing node in segmenting the first computing task using the first model to be transmitted between the UE and the access network device. Furthermore, the first segmentation information of the first model is confirmed by the computing node, which helps reduce the workload of the UE.

[0174] Referring to Figure 3, which is a flowchart illustrating a communication method provided in an embodiment of this application, the method is applied to a second communication device (or can be expressed as the method being executed by the second communication device). This second communication device can be a management function network element or a communication module / processing module within a management function network element, or a circuit or chip within a management function network element (such as a modem chip, also known as a baseband chip, or a SoC chip or SIP chip containing a modem core, etc.), or a module or software capable of implementing all or part of the functions of the management function network element; alternatively, the second communication device can be a network device, or a module within a network device (e.g., a module, circuit, chip, or chip system, etc.), or a logical node, logical module, or software capable of implementing all or part of the functions of the network device. The management function network element can be an SMF network element in the system shown in Figure 1, or a network element in Figure 1 specifically used for computational management, such as a CMF network element. For ease of explanation, the following description will use the second communication device as a management function network element as an example. As shown in Figure 3, the method includes:

[0175] S301. The management function network element obtains multiple computing information from the UE and computing nodes and / or application servers. Each of the multiple computing information corresponds to a model segmentation information. The model segmentation information corresponding to the multiple computing information includes the first segmentation information of the first model. The first segmentation information of the first model is the information of the UE and / or computing nodes using the first model to segment the first computing task.

[0176] It should be noted that the above-mentioned computational information corresponds to the model's segmentation information. A model can have multiple segmentation methods, meaning a model has multiple segmentation information. One model segmentation information (i.e., one segmentation information of a model) corresponds to one computational information. The above-mentioned multiple computational information includes computational information corresponding to segmentation information of different models, and computational information corresponding to different segmentation information of the same model. The models corresponding to the above-mentioned multiple computational information are models that the UE has subscribed to, or models that can provide services that the UE has subscribed to.

[0177] The information used by the UE and / or computing node (or application server) to segment the first computing task using the first model can be the first segmentation information of the first model. This first segmentation information characterizes the segmentation method of the first model and includes one or more of the following: the segmentation point of the first model, the termination point of the first model on the UE or the starting point on the computing node, the segmentation level of the first model, the segmentation grade of the first model, the segmentation identifier of the first model, the model layer identifier of the first model, and end-cloud collaborative mode information. The computing information corresponding to the model segmentation information includes one or more of the following: the computing latency corresponding to the model and the data transmission rate corresponding to the model.

[0178] In one example, the amount of transmitted output data corresponding to the model is used to indicate the average or maximum amount of data or burst amount of data corresponding to the model's output results.

[0179] Specifically, the segmentation of the first computing task by the UE and / or computing node (or application server) using the first model means that the first model can be divided into a first sub-model and a second sub-model. When processing the first computing task, the UE uses the first sub-model to process the input data corresponding to the first computing task to obtain intermediate results. The UE transmits the intermediate results to the computing node, and the computing node uses the second sub-model to process the intermediate results. The processing result obtained is the processing result of the first computing task.

[0180] As described above, the UE and the computing node execute the first sub-model and the second sub-model of the first model, respectively. It should be understood that dividing the first model into the first sub-model and the second sub-model according to different partitioning methods will result in different latency for the UE and the computing node, different computing resources required, and different amounts of data transmitted between the UE and the computing node. Furthermore, due to the different amounts of data transmitted, the required bandwidth and data rate will also be different.

[0181] In one example, the first segmentation information of the first model is used to indicate the segmentation position where the first model is segmented into a first sub-model and a second sub-model.

[0182] S302, the management function network element determines multiple QoS files based on multiple calculation information.

[0183] The computational information corresponding to the segmentation information of each model includes, but is not limited to, one or more of the following: computational latency and data transmission volume.

[0184] In one possible implementation, each of the multiple QoS files includes a packet delay budget, and the calculation information includes the calculation delay corresponding to the model. The management function network element determines the multiple QoS files based on the calculation information corresponding to the multiple model segmentation information, including:

[0185] The management function network element determines the latency budget included in the QoS file based on the UE's QoE latency and the corresponding computational latency of the model. This latency budget includes the PDB or the computational task transmission latency budget.

[0186] The UE's QoE latency indicates the upper limit of the latency from when the UE sends a computation task to when it receives a response. The UE's QoE latency includes the RTT / round-trip time of data packets related to the computation task transmitted between the UE and the compute node (or application server) and the computation latency between the UE and the compute node (or application server).

[0187] In one example, the packet latency budget is half the difference between the UE's QoE latency and the computation latency corresponding to the model. It should be understood that the computation latency corresponding to the model is the time required for the UE and compute nodes to process computational tasks using the model, or the time required for the model to run on the UE and compute nodes.

[0188] In one example, Table 1 illustrates the correspondence between multiple model segmentation information and multiple computational information.

[0189] Table 1

[0190] The management function network element determines multiple first calculation information from the calculation information corresponding to multiple model segmentation information based on the UE's QoE latency. The calculation latency in the first calculation information is less than the UE's QoE latency. The management function network element determines multiple QoS files based on the multiple first calculation information.

[0191] In one example, each QoS file includes, but is not limited to, one or more of the following: delay budget, MFBR, MDBV, and GFBR. The delay budget includes the packet delay budget (PDB) or the calculation task transmission delay budget.

[0192] One possible interpretation is that the packet delay budget is used to indicate the upper limit of the delay for any data packet corresponding to the computing task to be transmitted or completed between the UE and the UPF network element or computing node; the computing task transmission delay budget is used to indicate the upper limit of the delay for multiple data packets corresponding to the computing task to be transmitted or completed between the UE and the UPF network element or computing node; and the computing delay is used to indicate one or more of the maximum delay, average delay, P95 delay, or SLA delay corresponding to the model executing the computing task.

[0193] S303. The management function network element sends multiple QoS files to the access network device. The multiple QoS files include a first QoS file. The QoS flow corresponding to the first QoS file is used by the access network device to schedule the first QoS file of the data packet corresponding to the first computing task.

[0194] In one possible implementation, the management function network element further sends a mapping rule to the access network device. This mapping rule indicates a first correspondence between the information of each QoS file in multiple QoS files and the segmentation information of multiple models. Optionally, the mapping rule is further used to instruct the access network device to schedule the data packets corresponding to the first computing task based on the QoS flow corresponding to the first QoS file determined by the first segmentation information of the first model determined by the UE. Optionally, the management function network element further sends third indication information to the access network device. This third indication information instructs the access network device to schedule the data packets corresponding to the first computing task based on the QoS flow corresponding to the first QoS file determined by the first segmentation information of the first model determined by the UE or the computing node.

[0195] The management function network element sends a first indication information to the UE, which instructs the UE to send the first segmentation information of the first model to the access network equipment.

[0196] In one example, Table 2 illustrates the correspondence between model segmentation information and the index of QoS files.

[0197] Table 2

[0198] As shown in Table 2, the same model has different segmentation information, corresponding to different QoS file indexes.

[0199] It can be seen that by sending the first correspondence to the access network device, the access network device can determine the information of the first QoS file based on the first segmentation information of the first model sent by the UE, and then determine the first QoS file from multiple QoS files based on the information of the first QoS file. By instructing the access network device to schedule the data packets corresponding to the first computing task based on the QoS flow corresponding to the first QoS file determined according to the first segmentation information of the first model, the access network device can meet the end-to-end QoS requirements of the computing task when scheduling the data packets corresponding to the first computing task using the QoS flow corresponding to the first QoS file. Furthermore, the implicit indication through the mapping rule eliminates the need to send additional indication information to the access network device, which helps reduce the amount of data transmitted and thus reduces transmission resource overhead. By instructing the UE to send the first segmentation information of the first model to the access network device, the access network device can select the first QoS file from multiple QoS files based on the first segmentation information of the first model and the information of the first QoS file, thereby enabling the access network device to meet the end-to-end QoS requirements of the computing task when scheduling the data packets corresponding to the first computing task using the QoS flow corresponding to the first QoS file.

[0200] In one possible implementation, the management function network element sends a second correspondence between multiple QoS flow information and multiple QoS file information to the access network device. The management function network element sends a QoS rule to the UE, which is used to indicate a third correspondence between multiple QoS flow information and multiple model segmentation information and / or to instruct the UE to determine the first QoS flow information based on the first segmentation information and the third correspondence of the first model, and to map the data packet corresponding to the first computing task to the first QoS flow for transmission based on the first QoS flow information.

[0201] The first QoS flow information corresponds to the first segmentation information of the first model, and the second correspondence is used by the access network device to determine the first QoS file based on the first QoS flow information. There is a correspondence between the first QoS flow information and the first QoS file.

[0202] In one example, QoS flow information includes the QoS flow identifier.

[0203] In one example, Table 3 illustrates the correspondence between a QoS flow identifier and an index of a QoS file.

[0204] Table 3

[0205] In one example, Table 4 illustrates the correspondence between model segmentation information and QoS flow representation.

[0206] Table 4

[0207] As shown in Table 4, different segmentation information for the same model corresponds to different QoS requirements, and therefore different QoS flow identifiers.

[0208] It can be seen that by sending QoS rules to the UE, the UE can determine the first QoS information corresponding to the first segmentation information of the first model based on the first segmentation information of the first model and the third correspondence indicated by the QoS rules. Then, the UE maps the data packets corresponding to the first computing task to the first QoS flow for transmission based on the first QoS flow information. By sending the second correspondence to the access network device, the access network device can determine the first QoS file based on the first QoS flow information and the second correspondence. This enables the access network device to meet the end-to-end QoS requirements of the computing task when scheduling the data packets corresponding to the first computing task using the QoS flow corresponding to the first QoS file.

[0209] In one possible implementation, the management function network element sends a mapping rule and a second indication information to the UE. The mapping rule is used to indicate a first correspondence between the information of each QoS file in the multiple QoS files and the multiple model segmentation information. The second indication information is used to instruct the UE to send the information of the first QoS file to the access network device. The information of the first QoS file is determined by the UE based on the first segmentation information of the first model and the first correspondence.

[0210] It should be noted that the function of the second indication information can be implicitly indicated through mapping rules, in which case the management function network element does not need to send the second indication information to the UE.

[0211] By sending mapping rules to the UE and instructing the UE to send the first segmentation information of the first model to the access network device, the UE can determine the information of the first QoS file based on the first segmentation information of the first model and send the information of the first QoS file to the access network device. This enables the access network device to select the first QoS file from multiple QoS files based on the information of the first QoS file. When the access network device schedules the data packets corresponding to the first computing task using the QoS flow corresponding to the first QoS file, it can meet the end-to-end QoS requirements of the computing task.

[0212] In one possible implementation, the method of this embodiment further includes:

[0213] The management function network element sends a fourth indication message to the computing node. This fourth indication message instructs the computing node to send the first segmentation information of the first model to the access network device or the UE. It can be seen that by instructing the computing node to send the first segmentation information of the first model to the access network device, the UE does not need to report, reducing the workload of the UE and also reducing transmission overhead.

[0214] In one possible implementation, the method of this embodiment further includes:

[0215] The management function network element sends a fifth indication message to the user plane function network element. This fifth indication message is used by the user plane function network element to send the first segmentation information of the first model to the access network device or UE. In one possible implementation, the user plane function network element identifies the first segmentation information of the first model sent by the application server or computing node, adds it to the GTP-U header of the data packet, and sends it to the access network device.

[0216] It can be seen that by instructing the user plane function network element to send the first segmentation information of the first model to the access network device, the UE does not need to report, which reduces the workload of the UE and reduces the transmission overhead.

[0217] In one possible implementation, the first segmentation information of the first model is determined by the computing node based on one or more of the following: the computing node's load status, the application layer interaction information between the computing node and the UE, and the air interface status information received by the computing node from the access network device. Alternatively, the first segmentation information of the first model is determined by the UE based on the air interface status information between the UE and the access network device.

[0218] It can be seen that when determining the first segmentation information of the first model, the computing node considers its load status, ensuring that its computing resources are sufficient to execute the corresponding portion of the first computing task when the UE and / or computing node subsequently use the first model to segment the first computing task. The computing node negotiates with the UE's application layer to determine the first segmentation information, ensuring that the UE and / or computing node can successfully execute the first computing task when using the first model to segment it. When determining the first segmentation information of the first model, air interface status information is considered, ensuring that data packets generated by the UE and / or computing node in segmenting the first computing task using the first model can be transmitted between the UE and the access network equipment. Furthermore, when the first segmentation information of the first model is determined by the computing node, it reduces the workload of the UE. Conversely, when the first segmentation information of the first model is determined by the UE, it reduces the workload of the computing node.

[0219] In one possible implementation, the air interface state information includes one or more of the following: congestion state information, data rate, bandwidth, latency, and wireless channel quality.

[0220] For example, the access network device provides the computing node with the current available data rate, such as the available data rate. Based on this information and the amount of data to be transmitted in the computing information corresponding to the segmentation, the computing node determines that the current available data rate can transmit a larger amount of data with lower transmission latency. Therefore, the computing node chooses the segmentation method that transmits a larger amount of data. As another example, the access network device provides the computing node with current congestion information, such as the percentage of congested data packets or based on ECN marking. The computing node determines that the current base station congestion is severe, and transmitting data packets will cause significant transmission latency. Therefore, the computing node can choose the segmentation method that transmits a smaller amount of data.

[0221] In one feasible implementation, the management function network element also sends a sixth instruction message to the access network device, which is used to instruct the access network device to open the network status information to the UE, computing node or application server.

[0222] In one feasible implementation, the management function network element sends a seventh instruction message to the computing node or server. This sixth instruction message is used to instruct the computing node or application server to send the model segmentation information to the access network device through the Artificial Intelligence Control Plane (AICP) layer.

[0223] As can be seen, in the scheme of this embodiment, the management function network element sets multiple QoS files corresponding to multiple model segmentation information and sends the multiple QoS files to the access network device, so that the access network device can select the first QoS file from the multiple QoS files based on the information fed back by the UE, thereby enabling the access network device to meet the end-to-end QoS requirements of the computing task when scheduling the data packets corresponding to the first computing task using the QoS flow corresponding to the first QoS file.

[0224] Referring to Figure 4, which is a flowchart illustrating another communication method provided in an embodiment of this application, this method is applied to a third communication device (or can be expressed as the method being executed by a third communication device). This third communication device can be a UE or a communication module / processing module within the UE, or a circuit or chip within the UE (such as a modem chip, also known as a baseband chip, or a system-on-a-chip (SoC) chip containing a modem core, or a system-in-package (SIP) chip, etc.), or a module or software capable of implementing all or part of the UE's functions; alternatively, the third communication device can be a network device, or a module within a network device (e.g., a module, circuit, chip, or chip system, etc.), or a logical node, logical module, or software capable of implementing all or part of the network device's functions. For ease of explanation, the UE in the system shown in Figure 1 will be used as an example for the following description. As shown in Figure 4, the method includes:

[0225] S401, the UE determines the first information, which is determined based on the information obtained by the UE and / or the computing node using the first model to divide the first computing task.

[0226] For example, the UE determines the first information based on the information obtained by the UE and / or computing nodes using the first model to segment the first computing task.

[0227] In one possible implementation, the first information includes the first segmentation information of the first model.

[0228] The first segmentation information of the first model includes one or more of the following: the segmentation point of the first model, the termination point of the first model on the UE or the starting point on the computing node, the segmentation level of the first model, the segmentation grade of the first model, the segmentation identifier of the first model, the model layer identifier of the first model, and the edge-cloud collaboration mode information. For a detailed explanation, please refer to the relevant description in S201, which will not be repeated here.

[0229] In one example, the first segmentation information of the first model is determined by the UE based on the air interface state information between the UE and the access network device. This air interface state information includes, but is not limited to, one or more of congestion state information, data rate, bandwidth, latency, and radio channel quality. Specifically, the management function network element deploys multiple model segmentation information on the UE, including different segmentation information for the same model and segmentation information for different models. The requirements for the air interface state information between the UE and the access network device differ when the UE and the access network device process computation tasks based on different segmentation information. The UE can perceive the air interface state information between itself and the access network device. Based on this information and the first computation task it needs to perform, the UE determines the first segmentation information of the first model. The first model is used to process the first computation task, and the air interface state information between the UE and the access network device satisfies the requirements of the UE and the computing node to process the first computation task based on the first segmentation information of the first model.

[0230] In another example, the first segmentation information of the first model is determined by the computing node based on one or more of the following: the computing node's load status, the interaction information between the computing node and the UE's application layer, and the air interface status information received by the computing node from the access network equipment side.

[0231] It should be understood that different model segmentation methods have different requirements for computing resources of computing nodes. For example, if the first model has 80 layers, the computing resources required for the computing node to execute 50 layers are different from those required to execute 80 layers. Therefore, when selecting model segmentation information, the computing node will consider the load status of the computing node or the remaining computing resources of the computing node, so that when the UE and the computing node use the first model to execute the first computing task, the computing node has sufficient computing resources to execute the part of the first model that needs to be executed by the computing node.

[0232] When determining the first segmentation information of the first model, the computing node interacts and negotiates with the application layer of the UE to determine the first segmentation information, so that when the UE and / or the computing node use the first model to segment the first computing task, the UE and the computing node can successfully execute the first computing task.

[0233] The management function network element deploys multiple model segmentation information on the computing node, including different segmentation information for the same model and segmentation information for different models. The computing node obtains the air interface status information from the access network device side, i.e., the air interface status information between the UE and the access network device, through a subscription method. Based on the information of the first computing task and the air interface status information, the computing node determines the first segmentation information of the first model from the deployed multiple model segmentation information. The first model is used to process the first computing task, and the air interface status information between the UE and the access network device can meet the needs of the UE and the computing node to process the first computing task based on the first segmentation information of the first model.

[0234] In one possible implementation, the first information includes information from a first QoS file, and the UE receives mapping rules from a management function network element; the mapping rules are used to indicate a first correspondence between the information of each QoS file in the multiple QoS files and the multiple model segmentation information.

[0235] The UE determines the first information, including: the UE determines the first segmentation information of the first model; and determines the information of the first QoS file based on the first segmentation information of the first model and the first correspondence relationship. The information of the first QoS file has a correspondence relationship with the first segmentation information of the first model. The first QoS file is used to schedule the data packets corresponding to the first computing task.

[0236] S402, the UE sends first information to the access network device, the first information being used to determine the QoS file of the QoS stream for transmitting the data packet corresponding to the first computing task.

[0237] In one possible implementation, the UE receives first indication information from a management function network element, which instructs the UE to send first segmentation information of the first model to the access network device.

[0238] It can be seen that by instructing the UE to report the first segmentation information of the first model, there is no need for the computing node to report, which helps to reduce the workload of the UE.

[0239] In one possible implementation, the UE receives second indication information from the management function network element. This second indication information instructs the UE to send information about the first QoS file to the access network device. Optionally, the function of the first indication information can be implicitly indicated through mapping rules, eliminating the need for the management function network element to send the first indication information to the UE separately, thus reducing the workload of the management function and lowering transmission resource overhead.

[0240] In one feasible implementation, the UE receives QoS rules from management function network elements. These QoS rules indicate a third correspondence between multiple QoS flow information and multiple model segmentation information, and / or instruct the UE to determine first QoS flow information based on first segmentation information and a second correspondence of a first model. The UE then maps data packets corresponding to a first computing task onto the first QoS flow for transmission based on the first QoS flow information. After determining the first segmentation information of the first model, the UE determines the first QoS flow information based on the second correspondence and the first segmentation information of the first model, where the first QoS flow information corresponds to the first segmentation information of the first model. The UE then maps the data packets corresponding to the first computing task processed by the UE onto the first QoS flow based on the first QoS information and transmits them to the access network device.

[0241] For the data packet corresponding to the first computation task, please refer to the following explanation:

[0242] The first model can be divided into a first sub-model and a second sub-model. When processing the first computing task, the UE uses the first sub-model to process the input data corresponding to the first computing task to obtain intermediate results. The UE transmits the intermediate results to the computing node, and the computing node uses the second sub-model to process the intermediate results. The resulting result is the processing result of the first computing task. The data packet corresponding to the first computing task processed by the UE is the aforementioned intermediate result.

[0243] As can be seen, in the embodiment, by having the UE feed back first information to the access network device, the access network device can determine the first QoS file based on the first information. This enables the access network device to meet the end-to-end QoS requirements of the computing task when scheduling data packets corresponding to the first computing task using the QoS flow corresponding to the first QoS file. By having the UE determine the information of the first QoS file based on the first segmentation information of the first model and send this information to the access network device, the access network device can select the first QoS file from multiple QoS files based on this information. When scheduling data packets corresponding to the first computing task using the QoS flow corresponding to the first QoS file, the access network device can meet the end-to-end QoS requirements of the computing task. Furthermore, the access network device can determine the first QoS file based solely on its information without performing other operations, which helps reduce the workload of the access network device.

[0244] Referring to Figure 5, Figure 5 is an interactive flowchart illustrating a communication method provided in an embodiment of this application. This method is applied to the system shown in Figure 1. As shown in Figure 5, the method includes:

[0245] S500 and UE complete the attachment process.

[0246] Optionally, during the UE attach process, the AMF network element carries a task identifier or service type identifier in the attach response message sent back to the UE. The task identifier or service identifier indicates the tasks the UE is allowed to perform or the service types the UE is allowed to use, such as text-to-text services, text-to-image services, or interactive AI applications. During the UE attach process, the AMF network element obtains the UE's subscription data from the UDM network element. The UE's subscription data includes indication information indicating the task / service types the UE has subscribed to. Based on this indication information, the AMF network element can determine the task / service types that the UE can use.

[0247] S501, the UE sends a computing plane connection request to the AMF network element.

[0248] The computation plane connection request includes the UE's identifier. Optionally, the request may also include the UE's partitioning capability indication information, which indicates that the UE supports model partitioning capabilities. Optionally, the request may also include the UE's requested service type or computation task type.

[0249] In one example, the service type requested by the UE is a newly defined service type based on the computing service, such as a task / service type in the attach response message fed back by the AMF network element. In this case, the computing plane connection request is a newly defined connection request, and new signaling is required to represent this computing plane connection request.

[0250] In another example, the service type requested by the UE is determined based on the DNN / S-NSSAI assigned to the computing service. In other words, the service type requested by the UE is indicated by the DNN / S-NSSAI assigned to the computing service. In this case, the computing plane connection request reuses an existing session connection request; that is, the computing plane connection request is an existing session connection request that includes the DNN / S-NSSAI assigned to the computing service. This DNN / S-NSSAI is used to determine the service type requested by the UE, and can be seen as an identifier of the service type requested by the UE.

[0251] S502, the AMF network element forwards the computing plane connection request from the UE to the management function network element.

[0252] Specifically, when the AMF network element determines that the UE requests computing services based on the computing plane connection request represented by the new signaling or DNN / S-NSSAI, the AMF network element forwards the computing plane connection request from the UE to the management function network element.

[0253] Optionally, the AMF network element selects a management function network element based on the UE's location information or the DNN / S-NSSAI allocated for computing services. Optionally, when selecting a management function network element, the AMF will consider the service type requested by the UE carried in the computing plane connection request, such as whether the selected management function network element meets the latency requirements of the UE's requested service type. Before this, the AMF network element determines whether the UE has permission to initiate a computing plane connection request; if it determines that the UE has permission to initiate a computing plane connection request, the AMF network element performs the operation of selecting a management function network element. Specifically, the AMF network element determines whether the UE has permission to initiate a computing plane connection request based on whether the UE has subscribed to a computing session. If it determines that the UE has subscribed to a computing session, the AMF network element determines that the UE has permission to initiate a computing plane connection request; if it determines that the UE has not subscribed to a computing session, the AMF network element determines that the UE does not have permission to initiate a computing plane connection request.

[0254] In one possible implementation architecture, the UE can directly send a connection request for the computing plane to the management function network element.

[0255] S503, the management function network element sends a compute management (CM) policy negotiation creation request to the PCF network element.

[0256] S504, the PCF network element obtains the UE's subscription data from the UDM network element.

[0257] The UE's subscription data calculation includes the UE's QoE latency. The PCF network element generates a first rule based on the UE's QoE latency. In one example, the first rule is a PCC rule.

[0258] Optionally, the UE's computational subscription data may include one or more of the UE's subscription model information, subscribed service type, and subscribed computational load. The first rule may also include one or more of the UE's subscription model information, subscribed service type, and subscribed computational load.

[0259] S505, PCF network element sends CM policy negotiation creation response to management function network element.

[0260] The CM strategy negotiation creation response is used to respond to the CM strategy negotiation creation request, and the CM strategy negotiation creation response includes the first rule.

[0261] It should be noted that if the local configuration of the management function network element includes the UE's QoE latency, then the management function network element does not need to initiate S503, that is, S503-S505 do not need to be executed.

[0262] S506. Management function network elements obtain model deployment information of UE and / or computing nodes based on PCC rules or local configuration.

[0263] The model deployment information includes one or more of the following: model information, corresponding computation latency, corresponding segmentation information, and corresponding data transmission volume.

[0264] It should be noted that the model deployment information obtained by the management function network element is the deployment information of the model that the UE has subscribed to, or the deployment information of the model that can provide the UE's subscribed services, or the deployment information of the model that provides the UE's subscribed computing load.

[0265] In one example, the management function network element can obtain the above model deployment information from the OAM network element.

[0266] In another example, the management function network element can obtain model deployment information of computing nodes within its service range based on local configuration. The management function network element selects computing nodes to serve the UE based on the UE's location information, the service range of the computing nodes, and whether the computing nodes support model partitioning. Specifically, the UE is located within the service range of the selected computing node, and the computing node supports model partitioning. The computing node shown in Figure 5 is the computing node serving the UE. After determining the computing node to serve the UE, the management function network element can obtain the model deployment information of that computing node.

[0267] In another example, the management function network element determines a first model or computing node based on the UE's subscribed computing data. This first model or computing node can be used to execute the UE's first computing task, i.e., the service subscribed to by the UE.

[0268] Specifically, when the compute plane connection request carries the service type requested by the UE, the management function network element determines, based on the service type requested by the UE, a first model capable of providing the requested service or a first model processing the UE's first computing service from the models subscribed to by the UE, and determines the IP address of the first model or the IP address of the compute node deployed with the first model. When the compute plane connection request does not carry the service requested by the UE, the management function network element uses the default model from the models subscribed to by the UE as the first model, and determines the IP address of the first model or the IP address of the compute node deployed with the first model.

[0269] It should be noted that if there are multiple computing nodes with the first model deployed, the management function network element can select the computing node closest to the UE from the multiple computing nodes with the first model deployed as the computing node.

[0270] The management function network element feeds back the IP address of the first model or the IP address of the computing node to the UE.

[0271] After determining the computing node serving the UE, the management function network element can obtain the model deployment information of the computing node, or after determining the first model, the management function network element can obtain the deployment information corresponding to the first model, including model splitting information, computing latency, and data transmission volume, or one or more of these.

[0272] S507, Management function network elements generate multiple QoS files based on the model deployment information of UE and / or computing nodes.

[0273] The model deployment information includes multiple model segmentation information and corresponding calculation information. The specific process by which management function network elements determine multiple QoS files based on this calculation information can be found in the relevant description in S302, and will not be repeated here.

[0274] In one example, the computational information includes, but is not limited to, one or more of computational latency and data transfer volume.

[0275] S508a: Management function network elements send multiple QoS files and mapping rules to access network devices.

[0276] The mapping rule is used to indicate the first correspondence between the information of each QoS file in multiple QoS files and the segmentation information of multiple models. Optionally, the management function network element also sends third indication information to the access network device. The third indication information is used to instruct the access network device to schedule the data packets corresponding to the first calculation task corresponding to the QoS file determined by the first segmentation information of the first model from the UE.

[0277] Optionally, the management function network element also sends the first QoS flow information to the access network device.

[0278] S509a, the management function network element sends the first instruction information to the UE.

[0279] The first indication information is used to instruct the UE to send the first segmentation information of the first model to the access network device. A detailed description of the first segmentation information of the first model can be found in the relevant descriptions of the embodiments corresponding to Figures 2-4, and will not be repeated here.

[0280] S510a, the UE sends the first segmentation information of the first model to the access network equipment.

[0281] Specifically, the UE determines the first segmentation information of the first model based on the air interface status information between the UE and the access network device, and the identifier of the first computing task. The first model is used to process the first computing task, and the air interface status information between the UE and the access network device is sufficient to meet the needs of the UE and the computing node to process the first computing task based on the first segmentation information of the first model. The UE sends a buffer status reporting (BSR) to the access network device, which carries the first segmentation information of the first model.

[0282] The UE determines the first segmentation information of the first model, processes the first computation task based on the first model using the first segmentation information, and obtains intermediate results. The UE sends a first data packet to the access network device, which carries the intermediate results.

[0283] Optionally, the management function network element sends a fourth indication information to the computing node, which instructs the computing node to send the first segmentation information of the first model to the access network device or UE.

[0284] It should be noted that the first segmentation information of the first model is determined by the computing node based on one or more of the following: the computing node's load status, the interaction information between the computing node and the UE's application layer, and the air interface status information of the access network equipment subscribed to by the computing node.

[0285] It should be understood that different model segmentation methods have different requirements for computing resources of computing nodes. For example, if the first model has 80 layers, the computing resources required for the computing node to execute 50 layers are different from those required to execute 80 layers. Therefore, when selecting model segmentation information, the computing node will consider the load status of the computing node or the remaining computing resources of the computing node, so that when the UE and the computing node use the first model to execute the first computing task, the computing node has sufficient computing resources to execute the part of the first model that needs to be executed by the computing node.

[0286] When determining the first segmentation information of the first model, the computing node interacts and negotiates with the application layer of the UE to determine the first segmentation information, so that when the UE and / or the computing node use the first model to segment the first computing task, the UE and the computing node can successfully execute the first computing task.

[0287] The management function network element deploys multiple model segmentation information on the computing node, including different segmentation information for the same model and segmentation information for different models. The computing node obtains the air interface status information from the access network device side, i.e., the air interface status information between the UE and the access network device, through a subscription method. Based on the information of the first computing task and the air interface status information, the computing node determines the first segmentation information of the first model from the deployed multiple model segmentation information. The first model is used to process the first computing task, and the air interface status information between the UE and the access network device can meet the needs of the UE and the computing node to process the first computing task based on the first segmentation information of the first model.

[0288] The computing node subscribes to the air interface status information between the access network device and the UE from the access network device in the following manner:

[0289] Method 1: The access network device feeds back the QoS file currently used by the access network device to the UE through the QNC that replaces the QoS file. The computing node can determine the air interface status information between the access network device and the UE based on the QoS file.

[0290] Method 2: The access network device feeds back real-time air interface status information between the access network device and the UE to the computing node.

[0291] Method 3: The computing node obtains the air interface status information between the access network device and the UE through the QoS monitoring mechanism.

[0292] After determining the first segmentation information of the first model, the computing node can transmit the first segmentation information of the first model to the UE in the following manner:

[0293] Method 1: The computing node sends an empty packet to the UE, and the empty packet carries the first segmentation information of the first model.

[0294] Method 2: The computing node sends a third data packet to the UE. This third data packet carries the processing result of the second computing task and the first segmentation information of the first model. The second computing task is the task executed before the first computing task.

[0295] Method 3: The computing node sends the first segmentation information of the first model to the UE through the AICP layer.

[0296] When a computing node sends an empty packet or a third data packet to a UE, it passes through a UPF network element. The UPF network element obtains the first segmentation information of the first model from the transport layer of the empty packet or the third data packet, and adds the first segmentation information of the first model to the header of the empty packet or the header of the third data packet. For example, the first segmentation information of the first model is carried in the GTP-U header of the empty packet or the GTP-U header of the third data packet. Then, the empty packet or the third data packet is sent to the access network device, thereby realizing the transmission of the first segmentation information of the first model to the access network device. In this case, the UE does not need to send the first segmentation information of the first model to the access network device.

[0297] It should be noted that if the computing node is deployed on the UPF network element, the UPF network element does not need to perform the operation of obtaining the first segmentation information of the first model from the transport layer of the empty packet or the third data packet. It only needs to add the first segmentation information of the first model to the header of the empty packet or the header of the third data packet and send it to the access network device.

[0298] S511a, The access network device determines the first QoS file based on the first segmentation information of the first model.

[0299] Specifically, the management function network element determines the information of the first QoS file based on the first segmentation information of the first model and the correspondence between the segmentation information of multiple models and multiple QoS files. The information of the first QoS file corresponds to the first segmentation information of the first model. The management function network element determines the first QoS file from multiple QoS files based on the information of the first QoS file.

[0300] S508b: The management function network element sends a second correspondence between multiple QoS flow information and multiple QoS file information to the access network device.

[0301] S509b: Management function network elements send QoS rules to the UE.

[0302] The QoS rules are used to indicate the third correspondence between multiple QoS flow information and multiple model segmentation information and / or to instruct the UE to determine the first QoS flow information based on the first segmentation information and the third correspondence of the first model, and to map the data packet corresponding to the first computing task to the first QoS flow for transmission based on the first QoS flow information.

[0303] S510b: The UE sends the first QoS flow to the access network equipment.

[0304] Specifically, the UE determines the first segmentation information of the first model based on the air interface state information between the UE and the access network device, and the identifier of the first computing task. The first model is used to process the first computing task, and the air interface state information between the UE and the access network device can meet the needs of the UE and the computing node to process the first computing task based on the first segmentation information of the first model. The UE determines the first QoS flow information based on the first segmentation information of the first model and the third correspondence between multiple QoS flow information and multiple model segmentation information, wherein the first QoS flow information corresponds to the first segmentation information of the first model.

[0305] The UE determines the first segmentation information of the first model, processes the first computational task based on the first model using the first segmentation information, and obtains intermediate results. The UE maps the first data packet to the first QoS flow and transmits it to the access network device through the first QoS flow information. This first data packet carries the intermediate results. Optionally, the first data packet also carries the first segmentation information of the first model.

[0306] S511b: The access network device determines the first QoS file based on the first QoS flow information.

[0307] Specifically, after receiving the first QoS stream, the access network device determines the first QoS stream information from the first QoS stream, and determines the information of the first QoS file based on the first QoS stream information and a second correspondence between the information of multiple QoS streams and the information of multiple QoS files; the first QoS stream information corresponds to the information of the first QoS file. The access network device determines the first QoS file from multiple QoS files based on the information of the first QoS file.

[0308] It should be noted that S508a-S511a and S508b-S511b are to be executed selectively.

[0309] S512, The access network device schedules the first data packet based on the QoS flow corresponding to the first QoS file.

[0310] Specifically, the access network device schedules a first data packet based on the QoS flow corresponding to the first QoS file to send the first data packet to the computing node. The computing node obtains intermediate results and first segmentation information of the first model from the first data packet. The computing node processes the intermediate results using the portion of the first model indicated by the first segmentation information of the first model to obtain a processing result, which is the processing result of the first computing task. The computing node sends a second data packet to the access network device, which carries the processing result of the first computing task. The access network device schedules the second data packet based on the QoS flow corresponding to the first QoS file to send the second data packet to the UE.

[0311] In one example, after the UE knows the IP address of the computing node where the first model resides, the UE sets the destination IP address of the first data packet to the IP address of the computing node. Upon receiving the first data packet, the access network device schedules it to the computing node based on the destination IP address and the first QoS file. The first data packet carries the identification information of the first model and the first segmentation information of the first model. The identification information of the first model can be the ID of the first model, the ID of the task executed by the first model, or the port number corresponding to the first model. After receiving the first data packet, the computing node parses the identification information, the first segmentation information, and the intermediate results from it. Based on the identification information, it determines the first model; based on the first segmentation information, it uses the first model to process the intermediate results, i.e., executes the UE's computing services to obtain the processing result of the first computing task.

[0312] In another example, the UE sets the destination address of the first data packet to the IP address of the first model. The management function network element is aware of the model information deployed on the computing node. Based on the IP address of the first model and the first QoS file, the management function network element schedules the first data packet to the computing node. After receiving the first data packet, the computing node parses the IP address of the first model, the first segmentation information of the first model, and the intermediate results from the first data packet. Based on the IP address of the first model, it determines the first model. Based on the first segmentation information of the first model, it uses the first model to process the intermediate results, that is, to execute the UE's computing services to obtain the processing result of the first computing task.

[0313] In another example, the destination address of the first data packet is the IP address of the computing node, and the source address is the IP address of the UE. After receiving the first data packet, the access network device schedules the first data packet to the computing node based on the destination IP address and the first QoS file. The first data packet carries the service type requested by the UE or the ID of the model that the UE needs to use, as well as the first segmentation information and intermediate results of the first model. After receiving the first data packet, the computing node parses the service type requested by the UE or the ID of the model that the UE needs to use, as well as the first segmentation information and intermediate results of the first model from the first data packet; the computing node determines the first model based on the service type requested by the UE or the ID of the model that the UE needs to use, and the first model can provide the service corresponding to the service type requested by the UE, or the ID of the first model is the ID of the model that the UE needs to use; the computing node processes the intermediate results using the first model based on the first segmentation information of the first model, that is, executes the UE's computing service to obtain the processing result of the first computing task.

[0314] It should be noted that when scheduling data packets corresponding to the first computing task, the access network device can schedule them based on the QoS stream corresponding to the QoS file fed back to the computing node by the access network device through the QNC mechanism.

[0315] It should be noted that if the first model has only one segmentation method, after receiving the IP address of the first model sent by the management function network element, the UE determines the first model based on the IP address of the first model, and then determines the first segmentation information of the first model. If the first model has multiple segmentation methods, after receiving the IP address of the first model sent by the management function network element, the UE determines the first model based on the IP address of the first model, and then determines the first segmentation information of the first model in the manner described above.

[0316] It should be noted that the beneficial effects of the embodiment corresponding to Figure 5 can be found in the relevant descriptions of the embodiments corresponding to Figures 2-4, and will not be repeated here.

[0317] Referring to Figure 6, Figure 6 is an interactive flowchart illustrating a communication method provided in an embodiment of this application. This method is applied to the system shown in Figure 1. As shown in Figure 6, the method includes:

[0318] S600 and UE complete the attachment process.

[0319] S601, the UE sends a computing plane connection request to the AMF network element.

[0320] S602, the AMF network element forwards the computing plane connection request from the UE to the management function network element.

[0321] S603, The management function network element sends a request to the PCF network element to negotiate and create a computing management (CM) policy.

[0322] S604, the PCF network element obtains the UE's subscription data from the UDM network element.

[0323] S605, PCF network element sends CM policy negotiation creation response to management function network element.

[0324] S606. Management function network elements obtain model deployment information of UE and / or computing nodes based on PCC rules or local configuration.

[0325] S607. Management function network elements generate multiple QoS files based on the model deployment information of UE and / or computing nodes.

[0326] It should be noted that the specific implementation process of S600-S607 can be found in the relevant descriptions of S500-S507, and will not be described here.

[0327] S608, the management function network element sends multiple QoS files and third indication information to the access network equipment.

[0328] The third indication information is used to instruct the access network device to schedule the data packets corresponding to the first calculation task based on the QoS flow corresponding to the first QoS file determined by the first segmentation information from the first model of the UE.

[0329] S609. Management function network elements send mapping rules to UE.

[0330] The mapping rule is used to indicate the first correspondence between the information in each QoS file and the multiple model segmentation information. Optionally, the mapping rule is used to instruct the UE to send the QoS file information to the access network device, or the management function network element further sends second indication information to the UE, which instructs the UE to send the first QoS file information to the access network device. Optionally, the second indication information is also used to instruct the UE to determine the QoS file information based on the selected model segmentation information.

[0331] S610, the UE sends the first QoS file information to the access network equipment.

[0332] Specifically, the UE determines the first segmentation information of the first model based on the air interface state information between the UE and the access network device, and the identifier of the first computing task. The first model is used to process the first computing task, and the air interface state information between the UE and the access network device can meet the needs of the UE and the computing node to process the first computing task based on the first segmentation information of the first model. The UE determines the information of the first QoS file based on the first segmentation information of the first model and the correspondence between multiple model segmentation information and multiple QoS file information. The UE sends a buffer status reporting (BSR) to the access network device, which carries the information of the first QoS file.

[0333] After determining the first segmentation information of the first model, the UE processes the first computational task based on the first segmentation information of the first model to obtain intermediate results. The UE then sends a first data packet to the access network device, which carries the intermediate results.

[0334] It should be noted that the segmentation information of the first mode can be determined by the computing node and sent to the UE. The process of the computing node determining the first segmentation information of the first mode can be found in the relevant description of S510a, and will not be described here.

[0335] S611. The access network device determines the first QoS file based on the information in the first QoS file.

[0336] Specifically, the management function network element determines the first QoS file from multiple QoS files based on the information in the first QoS file.

[0337] S612, The access network device schedules the first data packet based on the QoS flow corresponding to the first QoS file.

[0338] Specifically, the access network device schedules a first data packet based on the QoS flow corresponding to the first QoS file to send the first data packet to the computing node. The computing node obtains intermediate data packets related to the intermediate results and first segmentation information of the first model from the first data packet. The computing node processes the intermediate data packet using the portion of the first model indicated by the first segmentation information of the first model to obtain a result data packet related to the processing result. This processing result is the processing result of the first computing task. The computing node sends a second data packet to the access network device, which carries the processing result of the first computing task. The access network device schedules the second data packet based on the QoS flow corresponding to the first QoS file to send the second data packet to the UE.

[0339] It should be noted that the beneficial effects of the embodiment corresponding to Figure 6 can be found in the relevant descriptions of the embodiments corresponding to Figures 2-4, and will not be repeated here.

[0340] Referring to Figure 7, which is a schematic diagram of the structure of an access network device according to an embodiment of this application, the access network device 700 includes:

[0341] The receiving unit 701 is configured to receive first information from the UE or the computing node, the first information being determined based on information obtained by the UE and / or the computing node using the first model to divide the first computing task;

[0342] Determining unit 702 is used by the access network device to determine the first QoS file based on the first information;

[0343] The processing unit 703 is used to schedule data packets corresponding to the first computing task based on the QoS stream corresponding to the first QoS file.

[0344] In one possible implementation, the receiving unit 701 is further configured to receive multiple QoS files from the management function network element, and a first correspondence between the information of each QoS file and multiple model segmentation information; the first information includes first segmentation information of a first model from the UE or the computing node;

[0345] The determining unit 702 is specifically used to: determine the information of the first QoS file based on the first segmentation information of the first model and the aforementioned first correspondence, and determine the first QoS file from multiple QoS files based on the information of the first QoS file.

[0346] In one possible implementation, the first information includes information about the first QoS file. The information about the first QoS file is determined by the UE based on the first segmentation information of the first model and the first correspondence between the information of each QoS file and the segmentation information of the multiple models. The receiving unit 701 is also used to receive multiple QoS files from the management function network element.

[0347] The determining unit 702 is specifically used for: the access network device determining the first QoS file from multiple QoS files based on the information of the first QoS file.

[0348] In one possible implementation, the first segmentation information of the first model is determined by the UE based on the air interface state information between the UE and the access network equipment.

[0349] In one possible implementation, the first segmentation information of the first model is determined by the computing node based on one or more of the following: the computing node's load status, the interaction information between the computing node and the UE's application layer, and the air interface status information of the access network equipment subscribed to by the computing node.

[0350] In one possible implementation, the air interface state information includes one or more of the following: congestion state information, data rate, bandwidth, latency, and wireless channel quality.

[0351] In one possible implementation, the receiving unit 701 is further configured to receive a data packet corresponding to a first computing task from the UE or the computing node. The data packet corresponding to the first computing task is a data packet generated by the UE or the computing node using a first model to execute the first computing task based on the first segmentation information.

[0352] It is worth noting that the specific functional implementation of the access network device 700 is described in detail in the embodiment shown in Figure 2. Each unit or module in the access network device 700 can be individually or entirely merged into one or more other units or modules, or some of the units or modules can be further divided into multiple functionally smaller units or modules. This achieves the same operation without affecting the technical effects of the embodiments of this application. The aforementioned units or modules are based on logical functional division. In practical applications, the function of one unit (or module) is implemented by multiple units (or modules), or the function of multiple units (or modules) is implemented by one unit (or module).

[0353] Referring to Figure 8, which is a structural schematic diagram of a management function network element provided in an embodiment of this application, the management function network element 800 includes:

[0354] The acquisition unit 801 is used to acquire multiple computing information of UE, computing node and / or application server. Each computing information corresponds to a model segmentation information. The model segmentation information corresponding to the multiple computing information includes the first segmentation information of the first model. The first segmentation information of the first model is the information of UE and / or computing node segmenting the first computing task using the first model.

[0355] The determining unit 802 is used to determine multiple QoS files based on multiple calculation information;

[0356] The sending unit 803 is used to send multiple QoS files to the access network device. The multiple QoS files include a first QoS file. The QoS stream of the first QoS file is used by the access network device to schedule the data packets corresponding to the first computing task.

[0357] In one possible implementation, each of the multiple QoS files includes a latency budget, and the calculation information includes the computation latency corresponding to the model. The determination unit 802 is specifically used for:

[0358] The management function network element determines the latency budget included in the QoS file based on the UE's QoE latency and the calculated latency corresponding to the model.

[0359] In one possible implementation, the sending unit 803 is further configured to send a mapping rule to the access network device, the mapping rule indicating a first correspondence between the information of each QoS file in multiple QoS files and multiple model segmentation information.

[0360] In one possible implementation, the mapping rule is also used to instruct the access network device to schedule the data packets corresponding to the first computation task based on the QoS flow corresponding to the first QoS file determined according to the first segmentation information of the first model determined by the UE.

[0361] In one possible implementation, the sending unit 803 is further configured to send first indication information to the UE, the first indication information being used to instruct the UE to send first segmentation information of the first model to the access network device.

[0362] In one possible implementation, the sending unit 803 is further configured to send a second correspondence between information of multiple QoS streams and information of multiple QoS files to the access network device; send a QoS rule to the UE, the QoS rule being used to indicate a third correspondence between multiple QoS streams and information of multiple model segmentation and / or to instruct the UE to determine first QoS stream information based on the first segmentation information and the third correspondence of the first model, and to map the data packet corresponding to the first computing task onto the first QoS stream for transmission based on the first QoS stream information;

[0363] The first QoS flow information corresponds to the first segmentation information of the first model, and the second correspondence is used by the access network device to determine the first QoS file based on the first QoS flow information. There is a correspondence between the first QoS flow information and the first QoS file.

[0364] In one possible implementation, the sending unit 803 is further configured to send a mapping rule and a second indication information to the UE, wherein the mapping rule is used to indicate a first correspondence between the information of each QoS file of the multiple QoS files and the multiple model segmentation information; the second indication information is used to instruct the UE to send the information of the first QoS file to the access network device, wherein the information of the first QoS file is determined by the UE based on the first segmentation information of the first model and the first correspondence.

[0365] In one possible implementation, the sending unit 803 is further configured to send third indication information to the access network device, the third indication information being used to instruct the access network device to schedule the data packets corresponding to the first computing task based on the QoS flow corresponding to the first QoS file determined by the first segmentation information from the first model of the UE or the computing node.

[0366] In one possible implementation, the sending unit 803 is further configured to send a fourth indication information to the computing node, the fourth indication information being used to instruct the computing node to send the first segmentation information of the first model to the access network device or UE.

[0367] In one possible implementation, the sending unit 803 is further configured to send a fifth indication information to the user plane function network element, the fifth indication information being used by the user plane function network element to send the first segmentation information of the first model to the access network device or UE.

[0368] In one possible implementation, the first segmentation information of the first model is determined by the computing node based on one or more of the following: the computing node's load status, the application layer interaction information between the computing node and the UE, and the air interface status information subscribed by the computing node to the access network device. Alternatively, the first segmentation information of the first model is determined by the UE based on the air interface status information between the UE and the access network device.

[0369] In one possible implementation, the air interface state information includes one or more of the following: congestion state information, data rate, bandwidth, latency, and / or wireless channel quality.

[0370] It is worth noting that the specific functional implementation of the management function network element 800 is described in detail in the embodiment shown in Figure 3. Each unit or module in the management function network element 800 can be individually or entirely merged into one or more other units or modules, or some of these units or modules can be further divided into multiple functionally smaller units or modules. This achieves the same operation without affecting the technical effects of the embodiments of this application. The aforementioned units or modules are based on logical functional division. In practical applications, the function of one unit (or module) is implemented by multiple units (or modules), or the function of multiple units (or modules) is implemented by one unit (or module).

[0371] Referring to Figure 9, which is a schematic diagram of a UE provided in an embodiment of this application, the UE 900 includes:

[0372] The determining unit 901 is used to determine the first information, which is determined based on the information obtained by the UE and / or the computing node using the first model to divide the first computing task;

[0373] The sending unit 902 is used to send first information to the access network device. The first information is used to determine the QoS file of the QoS stream of the data packet corresponding to the first computing task.

[0374] In one possible implementation, the first information includes the first segmentation information of the first model.

[0375] In one possible implementation, the first information includes information from the first QoS file, UE 900, and also includes:

[0376] The receiving unit 903 is used to receive mapping rules from the management function network element; the mapping rules are used to indicate the first correspondence between the information of each QoS file in multiple QoS files and the multiple model segmentation information;

[0377] The determining unit 901 is specifically used for: the UE determining the first segmentation information of the first model; determining the information of the first QoS file based on the first segmentation information of the first model and the first correspondence relationship, wherein the information of the first QoS file has a correspondence relationship with the first segmentation information of the first model, and the first QoS file is used to schedule the data packets corresponding to the first computing task.

[0378] In one possible implementation, the first segmentation information of the first model is determined by the computing node based on one or more of the following: the computing node's load status, the application layer interaction information between the computing node and the UE, and the air interface status information subscribed by the computing node to the access network device. Alternatively, the first segmentation information of the first model is determined by the UE based on the air interface status information between the UE and the access network device.

[0379] In one possible implementation, the air interface state information includes one or more of the following: congestion state information, data rate, bandwidth, latency, and wireless channel quality.

[0380] In one possible implementation, the receiving unit 903 is further configured to: receive first indication information sent from the management function network element, the first indication information being used to instruct the UE to send first segmentation information of the first model to the access network device.

[0381] In one possible implementation, the receiving unit 903 is further configured to: receive second indication information sent from the management function network element, the second indication information being used to instruct the UE to send a first QoS file to the access network device.

[0382] In one possible implementation, the sending unit 902 is further configured to send a data packet corresponding to the first computing task to the access network device. The data packet corresponding to the first computing task is a data packet generated by the UE using the first model to execute the first computing task based on the first segmentation information.

[0383] It is worth noting that the specific functional implementation of UE 900 is described in detail in the embodiment shown in Figure 4. Each unit or module in UE 900 can be individually or entirely merged into one or more other units or modules, or some of these units or modules can be further divided into multiple functionally smaller units or modules. This achieves the same operation without affecting the technical effects of the embodiments of this application. The aforementioned units or modules are based on logical functional division. In practical applications, the function of one unit (or module) is implemented by multiple units (or modules), or the function of multiple units (or modules) is implemented by one unit (or module).

[0384] Based on the description of the above method embodiments and related device embodiments, please refer to FIG10, which provides a schematic diagram of the structure of a communication device 1000. The communication device 1000 shown in FIG10 includes a memory 1001, a processor 1002, a communication interface 1003, and a bus 1004. The memory 1001, the processor 1002, and the communication interface 1003 are interconnected through the bus 1004.

[0385] Optionally, the memory 1001 may be a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM).

[0386] The memory 1001 is capable of storing programs. When the program stored in the memory 1001 is executed by the processor 1002, the processor 1002 and the communication interface 1003 are used to execute the various steps of the communication method of any embodiment shown in Figures 2-6.

[0387] The processor 1002 employs a general-purpose central processing unit (CPU), microprocessor, application-specific integrated circuit (ASIC), graphics processing unit (GPU), or one or more integrated circuits to execute relevant programs to achieve the functions required by the units in the access network device 700, management function network element 800, or UE 900 of this application embodiment, or to execute the communication method of any embodiment shown in Figures 2-6 of this application.

[0388] The processor 1002 can also be an integrated circuit chip with signal processing capabilities. In implementation, each step of the communication method shown in Figures 2-8 of this application can be completed by the integrated logic circuitry in the hardware of the processor 1002 or by instructions in software form. Optionally, the processor 1002 can be a general-purpose processor, a digital signal processor (DSP), an ASIC, a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The processor 1002 can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor is a microprocessor or any conventional processor, etc. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or can be executed by a combination of hardware and software modules in the decoding processor. Optionally, the software modules are located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in the memory 1001. The processor 1002 reads the information in the memory 1001 and, in conjunction with its hardware, performs the functions required by the units included in the access network device 700, management function network element 800, or UE900 in the embodiments of this application, or executes the communication method of any of the embodiments shown in Figures 2-6.

[0389] The communication interface 1003 uses transceiver devices, such as, but not limited to, transceivers, to enable communication between the communication device 1000 and other devices or communication networks.

[0390] Bus 1004 may include a path for transmitting information between various components of communication device 1000 (e.g., memory 1001, processor 1002, communication interface 1003).

[0391] It should be noted that although the communication device 1000 shown in Figure 10 only illustrates the memory, processor, and communication interface, those skilled in the art should understand that in specific implementations, the communication device 1000 may also include other devices necessary for normal operation. Furthermore, depending on specific needs, those skilled in the art should understand that the communication device 1000 may also include hardware devices for implementing other additional functions. In addition, those skilled in the art should understand that the communication device 1000 may only include the devices necessary for implementing the embodiments of this application, and not necessarily all the devices shown in Figure 10.

[0392] This application also provides a chip, which includes a processor and a data interface. The processor reads instructions stored in a memory through the data interface to implement the communication method of this application.

[0393] Optionally, as one implementation, the chip further includes a memory storing instructions, and the processor is used to execute the instructions stored in the memory. When the instructions are executed, the processor is used to execute the communication method.

[0394] This application also provides a computer-readable storage medium storing instructions that, when executed on a computer or processor, cause the computer or processor to perform one or more steps of any of the above methods.

[0395] This application also provides a computer program product containing instructions. When the computer program product is run on a computer or processor, it causes the computer or processor to perform one or more steps of any of the methods described above.

[0396] Those skilled in the art will appreciate that the functionality described in conjunction with the various illustrative logic blocks, modules, and algorithmic steps disclosed herein can be implemented by hardware, software, firmware, or any combination thereof. If implemented in software, the functionality described by the various illustrative logic blocks, modules, and steps can be stored or transmitted as one or more instructions or codes on a computer-readable medium and executed by a hardware-based processing unit. The computer-readable medium may comprise a computer-readable storage medium, which corresponds to a tangible medium, such as a data storage medium, or a communication medium that includes any medium facilitating the transfer of a computer program from one place to another (e.g., based on a communication protocol). In this way, the computer-readable medium may substantially correspond to (1) a non-transitory tangible computer-readable storage medium, or (2) a communication medium, such as a signal or carrier wave. The data storage medium may be any available medium accessible by one or more computers or one or more processors to retrieve instructions, code, and / or data structures for implementing the techniques described in this application. A computer program product may comprise a computer-readable medium.

[0397] By way of example and not limitation, such computer-readable storage media includes RAM, ROM, EEPROM, CD-ROM or other optical disc storage devices, magnetic disk storage devices or other magnetic storage devices, flash memory, or any other media that can be used to store desired program code in the form of instructions or data structures and is accessible by a computer. Furthermore, any connection is properly referred to as computer-readable media. For example, if instructions are transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of media. However, it should be understood that the computer-readable storage media and data storage media do not include connections, carrier waves, signals, or other temporary media, but are specifically referring to non-temporary tangible storage media. As used herein, disks and optical discs include Compact Discs (CDs), Laser Discs, Optical Discs, Digital Versatile Discs (DVDs), and Blu-ray Discs, where disks typically reproduce data magnetically, while optical discs reproduce data optically using lasers. Combinations of these should also be included within the scope of computer-readable media.

[0398] Instructions can be executed by one or more processors, such as one or more DSPs, general-purpose microprocessors, ASICs, FPGAs, or other equivalent integrated or discrete logic circuits. Therefore, the term "processor" as used herein can refer to any of the aforementioned structures or any other structures suitable for implementing the techniques described herein. Furthermore, in some aspects, the functions described in the various illustrative logic blocks, modules, and steps described herein are provided within dedicated hardware and / or software modules configured for encoding and decoding, or incorporated into combined codecs. Moreover, the techniques can be fully implemented in one or more circuit or logic elements.

[0399] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the division of units is merely a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Optionally, the coupling, direct coupling, or communication connection shown or discussed between them may be through some interfaces, indirect coupling or communication connection of devices or units, such as electrical, mechanical, or other forms.

[0400] Optionally, the units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0401] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented, in whole or in part, as a computer program product. This computer program product includes one or more computer instructions. When these computer program instructions are loaded and executed on a computer, all or part of the processes or functions according to the embodiments of this application are generated.

[0402] The above description is merely a specific implementation of the embodiments of this application, but the protection scope of the embodiments of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in the embodiments of this application should be covered within the protection scope of the embodiments of this application. Therefore, the protection scope of the embodiments of this application should be determined by the protection scope of the claims.

Claims

1. A communication method, characterized in that, The method includes: Receive first information from a user equipment (UE) or a computing node, wherein the first information is determined based on information obtained by the UE and / or the computing node using a first model to divide a first computing task; Based on the first information, a first Quality of Service (QoS) file is determined; Data packets corresponding to the first computing task are scheduled based on the QoS stream corresponding to the first QoS file.

2. The method according to claim 1, characterized in that, The method further includes: Receive multiple QoS files from management function network elements, and the first correspondence between the information of each QoS file and the multiple model segmentation information; The first information includes first segmentation information of the first model from the UE or the computing node, and determining the first QoS file based on the first information includes: The information of the first QoS file is determined based on the first segmentation information of the first model and the first correspondence relationship; The first QoS file is determined from the plurality of QoS files based on the information in the first QoS file.

3. The method according to claim 1, characterized in that, The first information includes information from the first QoS file, which is determined by the UE based on first segmentation information of the first model and a first correspondence between the information of each QoS file in the multiple QoS files and the segmentation information of the multiple models. The method further includes: Receive multiple QoS files from management function network elements; The step of determining the first QoS file based on the first information includes: The first QoS file is determined from the plurality of QoS files based on the information in the first QoS file.

4. The method according to claim 2 or 3, characterized in that, The first segmentation information of the first model is determined by the UE based on the air interface status information between the UE and the access network device.

5. The method according to claim 2, characterized in that, The first segmentation information of the first model is determined by the computing node based on one or more of the following: the load status of the computing node, the interaction information between the computing node and the application layer of the UE, and the air interface status information received by the computing node from the access network device side.

6. The method according to claim 2 or 5, characterized in that, The first segmentation information of the first model is carried in an empty packet sent to the access network device, or in a data packet containing the execution result of the second computing task sent to the UE, wherein the second computing task is a task whose execution order is before the first computing task.

7. The method according to any one of claims 4-6, characterized in that, The air interface status information includes one or more of the following: congestion status information, data rate, bandwidth, and / or latency.

8. The method according to any one of claims 1-7, characterized in that, The method includes: The system receives a data packet corresponding to the first computing task from the UE or the computing node. The data packet corresponding to the first computing task includes a first data packet corresponding to the intermediate result generated by the UE or the computing node using the first model to execute the first computing task based on the first segmentation information, and a second data packet generated by the computing node or the UE using the data packet corresponding to the intermediate result to execute the computing task using the first model.

9. A communication method, characterized in that, The method includes: Obtain multiple computing information from user equipment (UE) and / or computing nodes, where each computing information corresponds to a model segmentation information. The model segmentation information corresponding to the multiple computing information includes a first segmentation information of a first model, where the first segmentation information of the first model is information on how the UE and / or the computing node segment a first computing task using the first model. Multiple Quality of Service (QoS) files are determined based on the aforementioned calculation information; The plurality of QoS files are sent to the access network device, the plurality of QoS files including a first QoS file, and the QoS stream corresponding to the first QoS file is used by the access network device to schedule the data packets corresponding to the first computing task.

10. The method according to claim 9, characterized in that, Each of the plurality of QoS files includes a latency budget, and the calculation information includes the computation latency corresponding to the model. Determining the plurality of QoS files based on the plurality of calculation information includes: The latency budget included in the QoS file is determined based on the UE's Quality of Experience (QoE) latency and the computation latency corresponding to the model.

11. The method according to claim 9 or 10, characterized in that, The method further includes: A mapping rule is sent to the access network device, the mapping rule indicating the first correspondence between the information of each QoS file in the plurality of QoS files and the multiple model segmentation information.

12. The method according to claim 11, characterized in that, The mapping rule is also used to instruct the access network device to schedule the data packets corresponding to the first computing task based on the QoS flow corresponding to the first QoS file determined according to the first segmentation information of the first model determined by the UE.

13. The method according to claim 12, characterized in that, The method further includes: Send a first indication message to the UE, the first indication message being used to instruct the UE to send the first segmentation information of the first model to the access network device.

14. The method according to claim 9 or 10, characterized in that, The method further includes: Send a second correspondence between multiple QoS flow information and the multiple QoS files to the access network device. Send QoS rules to the UE, the QoS rules are used to indicate the third correspondence between the plurality of QoS flow information and the plurality of model segmentation information and / or instruct the UE to determine the first QoS flow information based on the first segmentation information of the first model and the third correspondence, and to map the data packet corresponding to the first computing task to the first QoS flow for transmission based on the first QoS flow information; Wherein, the first QoS flow information corresponds to the first segmentation information of the first model, and the second correspondence is used by the access network device to determine the first QoS file based on the first QoS flow information, and there is a correspondence between the first QoS flow information and the first QoS file.

15. The method according to claim 9 or 10, characterized in that, The method further includes: Send mapping rules and second indication information to the UE. The mapping rule is used to indicate the first correspondence between the information of each QoS file in the plurality of QoS files and the segmentation information of the plurality of models; the second indication information is used to instruct the UE to send the information of the first QoS file to the access network device, wherein the information of the first QoS file is determined by the UE based on the first segmentation information of the first model and the first correspondence.

16. The method according to claim 11 or 15, characterized in that, The method further includes: Send a third indication message to the access network device, the third indication message being used to instruct the access network device to schedule the data packets corresponding to the first computing task based on the QoS flow corresponding to the first QoS file determined by the first segmentation information from the first model of the UE or the computing node.

17. The method according to claim 16, characterized in that, The method further includes: A fourth indication message is sent to the computing node, the fourth indication message being used to instruct the computing node to send the first segmentation information of the first model to the access network device or UE.

18. The method according to any one of claims 9-16, characterized in that, The first segmentation information of the first model is determined by the computing node based on one or more of the following: the load status of the computing node, the interaction information between the computing node and the UE at the application layer, and the air interface status information subscribed by the computing node to the access network device. Alternatively, the first segmentation information of the first model is determined by the UE based on the air interface status information between the UE and the access network device.

19. The method according to claim 18, characterized in that, The air interface status information includes one or more of the following: congestion status information, data rate, bandwidth, and / or latency.

20. A communication method, characterized in that, The method includes: The first information is determined based on information obtained by the UE and / or the computing node using the first model to divide the first computing task; The first information is sent to the access network device, and the first information is used to determine the QoS file of the QoS stream for transmitting the data packets corresponding to the first computing task.

21. The method according to claim 20, characterized in that, The first information includes the first segmentation information of the first model.

22. The method according to claim 20, characterized in that, The first information includes information from the first QoS file, and the method includes: Receive mapping rules from management function network elements; the mapping rules are used to indicate the first correspondence between the information of each QoS file in the plurality of QoS files and the multiple model segmentation information; The determination of the first information includes: Determine the first segmentation information of the first model; The information of the first QoS file is determined based on the first segmentation information of the first model and the first correspondence relationship. The information of the first QoS file has a correspondence relationship with the first segmentation information of the first model. The QoS stream corresponding to the first QoS file is used to schedule the data packets corresponding to the first computing task.

23. The method according to claim 21 or 22, characterized in that, The first segmentation information of the first model is determined by the computing node based on one or more of the following: the load status of the computing node, the interaction information between the computing node and the UE at the application layer, and the air interface status information received by the computing node from the access network device. Alternatively, the first segmentation information of the first model is determined by the UE based on the air interface status information between the UE and the access network device.

24. The method according to claim 23, characterized in that, The air interface status information includes one or more of the following: congestion status information, data rate, bandwidth, and / or latency.

25. The method according to claim 21, characterized in that, The method further includes: The system receives a first indication message from a management function network element, the first indication message being used to instruct the UE to report the first segmentation information of the first model to the access network device.

26. The method according to claim 22, characterized in that, The method further includes: The UE receives a second indication message from a management function network element, the second indication message being used to instruct the UE to send information of the first QoS file to the access network device.

27. The method according to any one of claims 20-26, characterized in that, The method further includes: The data packet corresponding to the first computing task is sent to the access network device. The data packet corresponding to the first computing task is a first data packet of intermediate results generated by the UE using the first model to execute the first computing task based on the first segmentation information. The first data packet is used as a second data packet of processing results generated by the computing node after executing the first task using the first model.

28. A communication device, characterized in that, The communication device includes at least one processor coupled to at least one memory, the at least one processor being configured to execute a computer program or instructions stored in the at least one memory to cause the communication device to perform the method of any one of claims 1-8, or the method of any one of claims 9-20, or the method of any one of claims 21-27.

29. A communication device, characterized in that, The method includes a processor and a memory, wherein the memory is used to store program code, and the processor is used to execute the program code to implement the method of any one of claims 1-27.

30. A communication method, characterized in that, The method is applied to a communication system, which includes access network equipment, management function network elements, computing nodes, and user equipment (UE). The method includes: The management function network element obtains multiple computing information from the UE and / or the computing node. Each of the multiple computing information corresponds to a segmentation information of a model. The segmentation information of the model corresponding to the multiple computing information includes the first segmentation information of the first model. The first segmentation information of the first model is the information of the UE and the computing node using the first model to segment the first computing task. The management function network element determines multiple Quality of Service (QoS) files based on the multiple calculation information; The management function network element sends the multiple QoS files to the access network device; The UE or the computing node sends first information to the access network device, the first information being determined based on information obtained by the UE and the computing node using a first model to divide the first computing task; The access network device determines the first QoS file based on the first information and the plurality of QoS files; and schedules the data packets corresponding to the first computing task based on the QoS flow corresponding to the first QoS file.

31. A communication system, characterized in that, The communication system includes access network equipment, management function network elements, computing nodes, and user equipment (UE). The management function network element is used to obtain multiple computing information of the UE and / or the computing node. Each of the multiple computing information corresponds to a model segmentation information. The model segmentation information corresponding to the multiple computing information includes the first segmentation information of the first model. The first segmentation information of the first model is the information of the UE and the computing node using the first model to segment the first computing task. The management function network element is also used to determine multiple Quality of Service (QoS) files based on the multiple calculation information; The management function network element is also used to send the multiple QoS files to the access network device; The UE or the computing node is configured to send first information to the access network device, wherein the first information is determined based on information obtained by the UE and the computing node using a first model to divide the first computing task; The access network device is configured to determine the first QoS file based on the first information and the plurality of QoS files; Data packets corresponding to the first computing task are scheduled based on the QoS stream corresponding to the first QoS file.

32. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method according to any one of claims 1-27.

33. A computer program product, characterized in that, When the computer program product is run on a computer, it causes the computer to perform the method as described in any one of claims 1-27.

Images