Communication method and apparatus

Abstract

The present application relates to the field of communications, and discloses a communication method and apparatus, capable of improving the model inference energy efficiency. The communication method comprises: a first node sends, to a second node, first information for querying processing capability and / or energy supply information of at least one third node; the first node receives, from the second node, second information indicating the processing capability and / or energy supply information of the at least one third node; and the first node determines, on the basis of the second information, at least one target third node for executing a task.

Description

Communication methods and devices

[0001] This application claims priority to Chinese Patent Application No. 202311739904.1, filed with the State Intellectual Property Office of China on December 15, 2023, entitled "Communication Method and Apparatus", the entire contents of which are incorporated herein by reference. Technical Field

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

[0003] In the field of artificial intelligence (AI) / machine learning (ML), the number of model parameters is rapidly expanding. This trend brings more accurate results and better performance, but also leads to greater energy consumption. AI / ML energy consumption is influenced by various factors, but overall, model energy consumption is mainly distributed in the training and inference execution parts. Of these two parts, inference execution accounts for 80% to 90% of the energy consumption; that is, the model's energy consumption is primarily concentrated in the inference execution section.

[0004] The 3rd Generation Partnership Project (3GPP) R18 AIMLMGMT project studied model lifecycle management. Currently, the standard is under discussion for R19, and there is a preliminary agreement that R19 should further study the sustainability of AI / ML, with a focus on energy consumption and efficiency.

[0005] Therefore, how to reduce the energy consumption of inference execution in the AI / ML field and improve the energy efficiency of model inference is an urgent problem to be solved. Summary of the Invention

[0006] This application provides a communication method and apparatus for improving the energy efficiency of model inference.

[0007] To achieve the above objectives, this application adopts the following technical solution:

[0008] In a first aspect, a communication method is provided, applied to a first node. The execution subject of this method can be the first node, a component or device applied to the first node (e.g., a processor, chip, or chip system), or a logic module or software capable of implementing all or part of the functions of the first node. The communication method includes: sending first information to a second node for querying the processing capacity and / or power supply information of at least one third node; receiving second information from the second node indicating the processing capacity and / or power supply information of at least one third node; and determining at least one target third node for performing a task based on the second information.

[0009] In the first aspect, the first node first queries the second node for the processing capacity and / or energy supply information of the third node, and then determines the target third node for performing the task based on the processing capacity and / or energy supply information of the third node. This allows the selected target third node to perform the task with lower energy consumption (and / or lower carbon emissions), thereby improving energy efficiency during the task execution process.

[0010] In conjunction with the first aspect, in one possible implementation, the method may further include: sending third information to the second node, wherein the third information is used to indicate a third node that needs to obtain the execution result of the task and at least one target third node.

[0011] In this implementation, the first node notifies the second node of the third node that it needs to obtain the execution result of the task, as well as the target third node, through the third information, so that the task can be executed.

[0012] In conjunction with the first aspect, in one possible implementation, the second information includes identification information of at least one third node and corresponding power supply information; the method further includes: determining the time for each target third node to perform its task based on the power supply information in the second information.

[0013] In this implementation, the first node can determine the time for each target third node to execute its task based on the power supply information, thus ensuring the execution of the task.

[0014] In conjunction with the first aspect, in one possible implementation, the first information is also used to indicate the range information of the third node to be queried.

[0015] In this implementation, the first node also indicates the range information of at least one third node through the first information. After receiving the first information, the second node can feed back the processing capacity and / or power supply information of the third node within the range information indicated by the first information to the first node. The range information indicates the geographical range of the third node to be queried. By limiting the query to third nodes within a certain range, the subsequent transmission overhead can be reduced.

[0016] In conjunction with the first aspect, in one possible implementation, the method further includes: sending a sixth message to the second node, wherein the sixth message indicates the historical processing results of the invoked task.

[0017] In this implementation, the first node indicates the historical processing results of the task being called. Using the historical processing results to replace the actual task execution process can reduce the computing power required to execute the task, thereby improving the energy efficiency during the task execution process.

[0018] In conjunction with the first aspect, in one possible implementation, the task includes at least one of the following: model training, model testing, model simulation, model loading, or model inference.

[0019] In conjunction with the first aspect, processing capacity includes at least one of the following: hardware processing capacity, storage capacity, or hardware energy efficiency information; energy supply information includes at least one of the following: the type of energy supplied, the energy availability time, or the carbon emissions per unit of energy.

[0020] Secondly, a communication method is provided, applied to a second node. The execution subject of this method can be the second node, a component or device applied to the second node (e.g., a processor, chip, or chip system), or a logic module or software capable of implementing all or part of the functions of the second node. The communication method includes: receiving first information, the first information being used to query the processing capacity and / or power supply information of at least one third node; determining second information based on the first information, the second information indicating the processing capacity and / or power supply information of at least one third node; and sending the second information.

[0021] In the second aspect, the second node first receives first information for querying the processing capacity and / or power supply information of at least one third node, and determines second information indicating the processing capacity and / or power supply information of at least one third node based on the first information. Then, it sends the second information to the first node, which can determine the target third node for performing the task based on the processing capacity and / or power supply information of the third node. This can make the selected target third node perform the task with lower energy consumption (and / or lower carbon emissions), thereby improving energy efficiency during the task execution process.

[0022] In conjunction with the second aspect, in one possible implementation, the method further includes: receiving third information, which indicates a third node that needs to obtain the execution result of the task and a target third node for executing the task; and sending fourth information to the target third node for executing the task, wherein the fourth information indicates a third node that needs to obtain the execution result of the task.

[0023] In this implementation, the second node determines the third node that needs to obtain the execution result of the task and the target third node based on the third information. Then, based on the fourth information, it notifies the target third node that the task execution result needs to be obtained, thus enabling the execution of the task.

[0024] In conjunction with the second aspect, in one possible implementation, the method further includes: receiving a sixth message, wherein the sixth message indicates the invocation of historical processing results; and sending the sixth message to the target third node.

[0025] In this implementation, the sixth information, which indicates the historical processing results of the task to be invoked, is received and sent to the target third node. The target third node can use the historical processing results to replace the actual task execution process, thereby reducing the computing power required to execute the task and improving the energy efficiency during the task execution process.

[0026] In conjunction with the second aspect, in one possible implementation, the first information is also used to indicate the range information of the third node to be queried.

[0027] In this implementation, after receiving the first information, the second node can feed back the processing capacity and / or power supply information of the third node within the range indicated by the first information to the first node. The range information indicates the geographical range of the third node to be queried. By limiting the query to the third node within a certain range, the subsequent transmission overhead can be reduced.

[0028] In conjunction with the second aspect, in one possible implementation, the method also includes:

[0029] Obtain the processing capacity and / or power supply information of at least one third node.

[0030] In this implementation, the second node obtains the processing capacity and / or power supply information of the third node, which can be used by the first node to determine the target third node for executing the task based on strategies to reduce energy consumption / carbon emissions, thereby improving the energy efficiency of task processing.

[0031] Thirdly, a communication method is provided, applied to a third node executing a task (i.e., a target third node). The executing entity of this method can be the target third node, a component or device applied to the target third node (e.g., a processor, chip, or chip system), or a logic module or software capable of implementing all or part of the target third node's functions. The communication method includes: receiving fourth information, wherein the fourth information indicates a third node that needs to obtain the execution result of a task; receiving task parameters required to execute the task; executing the task according to the task parameters to obtain the task execution result; and sending the execution result to the third node that needs to obtain the task execution result.

[0032] In the third aspect, the target third node executes the task based on the task parameters and sends the execution result to the third node that needs to obtain the execution result of the task, thus satisfying the task requirements of the third node that needs to obtain the execution result of the task.

[0033] In conjunction with the third aspect, in one possible implementation, the third node executing the task stores historical task parameters and corresponding historical processing results; the task is executed according to the task parameters to obtain the task execution result, including: determining the target historical processing result as the task execution result, wherein the historical task parameters corresponding to the target historical processing result match the task parameters.

[0034] In this implementation, the target third node uses historical processing results to replace the actual task execution process, which can reduce the computing power required to execute the task and thus improve the energy efficiency during the task execution process.

[0035] Optionally, in this implementation, the target third node may determine the target historical processing result as the task execution result after receiving the sixth information indicating that the historical processing result should be invoked.

[0036] Fourthly, this application provides a communication device, which can be a first node or a chip or system-on-a-chip within the first node. This communication device can implement the functions performed by the first node in the first aspect or possible designs described above. These functions can be implemented in hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the aforementioned functions. For example, the communication device includes: a transceiver module for sending first information to a second node, the first information being used to query the processing capacity and / or power supply information of at least one third node; the transceiver module is also used to receive second information, the second information indicating the processing capacity and / or power supply information of at least one third node; and a processing module for determining at least one target third node for performing a task based on the second information.

[0037] In conjunction with the fourth aspect, in one possible implementation, the apparatus further includes: a transceiver module for sending third information to the second node, wherein the third information is used to indicate a third node that needs to obtain the execution result of the task and at least one target third node.

[0038] In conjunction with the fourth aspect, in one possible implementation, the second information includes identification information of at least one third node and corresponding power supply information; the processing module is further configured to: determine the time for each target third node to perform its task based on the power supply information in the second information.

[0039] In conjunction with the fourth aspect, in one possible implementation, the first information is also used to indicate the range information of the third node to be queried.

[0040] In conjunction with the fourth aspect, in one possible implementation, the transceiver module is also used to: send a sixth message to the second node, wherein the sixth message indicates the historical processing result of the invoked task.

[0041] In conjunction with the fourth aspect, in one possible implementation, the task includes at least one of the following: model training, model testing, model simulation, model loading, or model inference.

[0042] In conjunction with the fourth aspect, processing capacity includes at least one of the following: hardware processing capacity, storage capacity, or hardware energy efficiency information; energy supply information includes at least one of the following: the type of energy supplied, the energy availability time, or the carbon emissions per unit of energy.

[0043] Fifthly, this application provides a communication device, which can be a second node or a chip or system-on-a-chip within the second node. This communication device can implement the functions performed by the second node in the second aspect or possible designs described above. These functions can be implemented in hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the aforementioned functions. For example, the communication device includes: a transceiver module for receiving first information, which is used to query the processing capacity and / or power supply information of at least one third node; a processing module for determining second information based on the first information, which indicates the processing capacity and / or power supply information of at least one third node; and the transceiver module is further used to send the second information.

[0044] In conjunction with the fifth aspect, in one possible implementation, the transceiver module is also used to receive third information, which indicates the third node that needs to obtain the execution result of the task and the target third node for executing the task.

[0045] The transceiver module is also used to send a fourth message to the target third node that is executing the task, wherein the fourth message indicates the third node that needs to obtain the execution result of the task.

[0046] In conjunction with the fifth aspect, in one possible implementation, the transceiver module is also used for:

[0047] Receive the sixth message, which indicates the retrieval of historical processing results;

[0048] Send the sixth message to the target third node.

[0049] In conjunction with the fifth aspect, in one possible implementation, the first information is also used to indicate the range information of the third node to be queried.

[0050] In conjunction with the fifth aspect, in one possible implementation, the processing module is also used for:

[0051] The transceiver module obtains the processing capacity and / or power supply information of at least one third node.

[0052] Sixthly, this application provides a communication device, which can be a target third node or a chip or system-on-a-chip within the target third node. This communication device can realize the functions performed by the target third node in the aforementioned third aspect or possible designs of the third aspect. These functions can be implemented in hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the aforementioned functions. For example, the communication device includes: a transceiver module for receiving fourth information, wherein the fourth information indicates a third node that needs to obtain the execution result of a task; the transceiver module is also used to receive task parameters required for executing the task; a processing module is used to execute the task according to the task parameters and obtain the execution result of the task; the transceiver module is also used to send the execution result to the third node that needs to obtain the execution result of the task.

[0053] In conjunction with the sixth aspect, in one possible implementation, the third node executing the task stores historical task parameters and corresponding historical processing results; the processing module is specifically used to determine the target historical processing result as the execution result of the task, wherein the historical task parameters corresponding to the target historical processing result match the task parameters.

[0054] In a seventh aspect, this application provides a communication device including a processor and a transceiver, the processor and transceiver being configured to support the communication device in executing the method of the first aspect. Further, the communication device may also include a memory storing computer instructions, the processor being able to execute the computer instructions to perform the method of the first aspect, the second aspect, or the third aspect.

[0055] Eighthly, this application provides a computer-readable storage medium that stores computer instructions, wherein when the computer instructions are executed, the methods of the first, second, or third aspects are performed.

[0056] Ninthly, this application provides a computer program product containing instructions that, when run on a computer, enable the computer to perform the methods described in the first, second, or third aspects above.

[0057] In a tenth aspect, this application provides a chip including a processor and a transceiver, the processor and transceiver being configured to support a communication device in performing the methods of the first aspect, the second aspect, or the third aspect.

[0058] Eleventhly, this application provides a communication system, which includes a first node and a second node. The first node is used to execute the method of the first aspect, and the second node is used to execute the method of the second aspect.

[0059] Optionally, the communication system may also include a third node for performing the third aspect of the method.

[0060] The beneficial effects described in aspects four through eleven of this application can be referred to the beneficial effect analysis of aspects one, two or three, and will not be repeated here. Attached Figure Description

[0061] Figure 1 is a schematic diagram of the AI / ML workflow provided in an embodiment of this application;

[0062] Figure 2 is a schematic diagram of the architecture of a communication system provided in an embodiment of this application;

[0063] Figure 3 is a schematic diagram of the architecture of another communication system provided in an embodiment of this application;

[0064] Figure 4 is a schematic diagram of the architecture of another communication system provided in an embodiment of this application;

[0065] Figure 5 is a schematic diagram of the architecture of another communication system provided in an embodiment of this application;

[0066] Figure 6 is a schematic diagram of the architecture of another communication system provided in an embodiment of this application;

[0067] Figure 7 is a schematic diagram of the architecture of another communication system provided in an embodiment of this application;

[0068] Figure 8 is a flowchart illustrating a communication method provided in an embodiment of this application;

[0069] Figure 9 is a flowchart illustrating another communication method provided in an embodiment of this application;

[0070] Figure 10 is a flowchart illustrating another communication method provided in an embodiment of this application;

[0071] Figure 11 is a flowchart illustrating another communication method provided in an embodiment of this application;

[0072] Figure 12 is a flowchart illustrating another communication method provided in an embodiment of this application;

[0073] Figure 13 is a flowchart illustrating another communication method provided in an embodiment of this application;

[0074] Figure 14 is a flowchart illustrating another communication method provided in an embodiment of this application;

[0075] Figure 15 is a schematic diagram of the structure of a communication device provided in an embodiment of this application;

[0076] Figure 16 is a schematic diagram of another communication device provided in an embodiment of this application;

[0077] Figure 17 is a schematic diagram of another communication device provided in an embodiment of this application;

[0078] Figure 18 is a schematic diagram of another communication device provided in an embodiment of this application. Detailed Implementation

[0079] Before introducing the embodiments of this application, some terms involved in the embodiments of this application are explained as shown in Table 1.

[0080] Table 1

[0081] The existing AI / ML workflow, as shown in Figure 1, mainly includes the training phase (model training and testing), the simulation phase (model simulation), the deployment phase (model loading), and the inference phase (model inference). In the deployment phase, the model obtained during training is loaded into the target inference function for execution during the inference phase. Current AI / ML workflows primarily prioritize performance improvement without considering energy consumption constraints, resulting in significant energy consumption during model usage and a severe environmental impact.

[0082] AI refers to enabling machines to possess human-like intelligence, such as allowing machines to use computer hardware and software to simulate certain intelligent human behaviors. To achieve AI, machine learning (ML) methods can be employed, such as machines learning or training ML models using training data (model training). The trained model can represent the mapping between input and output data and can be used for reasoning (model inference). In other words, the model can be used to obtain the output data corresponding to given input data. It can be understood that the trained model can be an AI model, an ML model, or other models, and AI models can include ML models; that is, an ML model is a type of AI model. An AI model can be considered a specific method for implementing AI functions, representing the mapping relationship or function between the input and output of the AI ​​model. This AI function can also be called AI operation, which can include at least one of the following functions: data collection, model training, model learning, model information dissemination, model inference, model reasoning, or model prediction.

[0083] Currently, energy consumption during model usage can be reduced through model lightweighting. Model lightweighting is a technique that reduces the number of model parameters while maintaining necessary accuracy. It includes techniques such as pruning, quantization, and distillation. Model pruning removes "unimportant" weights from the model, reducing the number of parameters and computational cost while minimizing any impact on accuracy. Model quantization transforms trained model parameters from high-precision to low-precision, effectively reducing computational overhead, parameter size, and memory consumption. Model distillation involves transferring knowledge from a complex model to a smaller, simpler model, allowing the simpler model to approximate or even surpass the performance of the complex model, thus achieving similar predictive results with less complexity. Model lightweighting techniques reduce computational overhead and consequently, energy consumption during model inference.

[0084] However, lightweighting techniques can only reduce computational energy consumption by decreasing the number of model parameters. To ensure model accuracy, the number of model parameters can only be reduced to a small extent. The actual factors affecting model inference energy consumption, besides the model itself, are directly related to the hardware capabilities running the model and the number of inference iterations. For example, experiments comparing the energy consumption of the GShard model and the GPT-3 model showed that GShard, with 619 billion parameters, consumed approximately 53 times less energy and had approximately 127 times less net carbon emissions than GPT-3, which had 175 billion parameters. This is mainly due to GShard's multiple optimizations in both algorithms and hardware. In other words, the energy reduction achieved by lightweighting techniques by decreasing the number of model parameters is very limited. Even with lightweighting techniques, the energy consumption of model inference remains high, and the energy efficiency of model inference is low.

[0085] To address the aforementioned technical problems, this application provides a communication method. Considering that model inference energy consumption can be estimated by multiplying the energy consumption of a single inference attempt by the number of inference attempts, this application improves model inference energy efficiency from two aspects: energy consumption per inference attempt and the number of inference attempts. 1. Energy efficiency is achieved by sacrificing transmission resources: by acquiring hardware energy efficiency information at different inference locations, the deployment location with the optimal energy efficiency is selected for inference execution. 2. Energy efficiency is achieved by sacrificing storage resources: historical inference results are stored, and when the input data deviation is less than a set threshold, historical inference results are directly used instead, reducing the number of inference execution attempts.

[0086] The communication method and communication system provided in the embodiments of this application will now be described with reference to the accompanying drawings.

[0087] The communication method provided in this application can be applied to various communication systems, such as: 6th generation (6G) mobile communication systems, Long Term Evolution (LTE) systems, 5th generation (5G) mobile communication systems, Wireless Fidelity (WiFi) systems, future communication systems, or systems integrating multiple communication systems, etc. This application does not limit the application to these systems. 5G can also be referred to as New Radio (NR).

[0088] The network architecture and business scenarios described in the embodiments of this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided in the embodiments of this application. As those skilled in the art will know, with the evolution of network architecture and the emergence of new business scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0089] Figure 2 shows a schematic diagram of the architecture of a communication system provided in an embodiment of this application. As shown in Figure 2, the communication system 20 includes a first node 210 and a second node 220. Optionally, as shown in Figure 2, the communication system also includes at least one third node 230.

[0090] The first node 210 is used to send first information to the second node 220, and the first information is used to query the processing capacity and / or power supply information of at least one third node 230.

[0091] The second node 220 is used to receive the first information and send the second information to the first node 210, the second information indicating the processing capacity and / or power supply information of the at least one third node 230.

[0092] The first node 210 is also used to receive second information and determine at least one target third node 230 for performing the task based on the second information.

[0093] Optionally, in the embodiments of this application, the first node can be, for example, a model loading consumer, the second node can be, for example, a model loading producer, and the third node can be, for example, a model inference device. Several possible application architectures will be given below in conjunction with different scenarios.

[0094] Figure 3 is a schematic diagram of a communication architecture provided in an embodiment of this application. As shown in Figure 3, the communication architecture includes a network management system (NMS), an element management system (EMS), and a radio access network (RAN) device.

[0095] Among them, NMS is responsible for the operation, management and maintenance functions of the network, and can also be called a cross-domain management system.

[0096] EMS is used to manage one or more network elements of a certain category; it can also be called a domain management system or a single-domain management system.

[0097] RAN equipment is a device in a mobile communication system that connects the fixed part and the radio part, and connects to the mobile terminal through an airborne wireless channel. Figure 3 illustrates the RAN equipment as a next-generation node B (gNB) as an example.

[0098] In this system, NMS acts as the model loading consumer, EMS acts as the model loading producer providing model loading services, and gNB has model inference capabilities (or simply inference capabilities), responsible for model inference, inputting data into the model to obtain the corresponding output. NMS can call the model loading service provided by EMS to load the model into a specified gNB.

[0099] Furthermore, the embodiments of this application are also applicable to O-RAN architecture or mobile intelligent function (MIF) architecture.

[0100] O-RAN aims to achieve intelligent, open access networks. Based on the concepts of RAN interoperability and standardization, O-RAN provides a unified interconnection standard for white-box hardware and software from different vendors.

[0101] As shown in Figure 4, in a possible O-RAN architecture, the NMS is embodied as the service management and orchestration framework (SMO), and the non-RT RIC belongs to the SMO and acts as the ML loading consumer. RAN equipment is divided into multiple logical network elements, including O-CU, O-DU, and O-RU modules. There is no EMS in O-RAN; the network element that implements EMS functionality is the Near-RT RIC. The Near-RT RIC simultaneously provides both ML loading producer and ML inference. Specifically, the Non-RT RIC acts as the ML loading consumer, the Near-RT RIC acts as the ML loading producer, and ML inference is deployed within the Near-RT RIC; multiple Near-RT RICs exist. The Non-RT RIC can select model deployment locations based on the inference hardware energy efficiency information of the Near-RT RIC, prioritizing maximum energy efficiency and minimum carbon emissions.

[0102] Alternatively, as shown in Figure 5, in another possible O-RAN architecture, the Near-RT RIC provides the ML loading producer, while any one of the O-CU, O-DU, or O-RU modules provides ML inference. Specifically, the Non-RT RIC acts as the ML loading consumer, the Near-RT RIC acts as the ML loading producer, and ML inference is deployed in the O-CU / O-DU / O-RU, with multiple O-CU / O-DU / O-RUs existing. The Non-RT RIC can select the model deployment location based on the inference hardware energy efficiency information of the O-CU / O-DU / O-RU, with the goal of maximizing energy efficiency and minimizing carbon emissions.

[0103] The MIF (Intelligent Service Logic Entity) is the AI-powered intelligent service processing logic entity in the RAN (Radio Area Network), used to implement intelligent data analysis functions in the RAN domain. The MIF connects to the gNB (Garden Network Node) via the G1 interface. The MIF deploys AI model-related functions such as model training and model evaluation.

[0104] As shown in Figure 6, in one possible MIF architecture, the MIF acts as the ML loading consumer, the CU / DU in the gNB acts as the ML loading producer, and ML inference is also deployed in the CU / DU. One MIF manages multiple gNBs. The MIF can select the model deployment location based on the inference hardware energy efficiency information of the CU / DU, with the goal of maximizing energy efficiency and minimizing carbon emissions.

[0105] Alternatively, as shown in Figure 7, in another possible MIF architecture, the MIF acts as the ML loading consumer, the CU acts as the ML loading producer, and ML inference is deployed in the DU. There are multiple DUs under one CU. The MIF can select the model deployment location based on the energy efficiency information of the DU's inference hardware, prioritizing maximum energy efficiency and minimum carbon emissions.

[0106] It should be understood that the network elements responsible for ML loading consumer, ML loading producer, and ML inference in the above examples are merely illustrative. In actual implementations, due to differences in network architecture, the network elements responsible for these roles may have other designs, which are not limited.

[0107] The communication method provided in the embodiments of this application will now be described with reference to the communication systems shown in Figures 2 to 7.

[0108] It should be noted that in the following embodiments of this application, the message names, parameter names, or information names between network elements are just examples. Other names may also be used in other embodiments. The communication method provided in this application does not specifically limit these names.

[0109] It is understood that in the embodiments of this application, each network element can execute some or all of the steps in the embodiments of this application. These steps or operations are merely examples, and the embodiments of this application can also perform other operations or variations of various operations. Furthermore, the steps can be executed in different orders as presented in the embodiments of this application, and it is not necessary to execute all the operations in the embodiments of this application.

[0110] Figure 8 illustrates an example of the communication method provided in this application. The method is described using the interaction between a first node and a second node as an example. Of course, the entity executing the action of the first node in this method can also be a device / module within the first node, such as a chip, processor, or processing unit within the first node; similarly, the entity executing the action of the second node in this method can also be a device / module within the second node, such as a chip, processor, or processing unit within the second node. This application does not specifically limit this. In this application, the processing performed by a single executing entity (e.g., the first node or the second node) can also be divided into multiple executing entities, which can be logically and / or physically separated. For example, as shown in Figure 8, the method may include the following steps:

[0111] S810, the first node sends the first message to the second node. Correspondingly, the second node receives the first message.

[0112] The first piece of information is used to query the processing capacity and / or power supply information of at least one third node. The processing capacity includes at least one of the following: hardware processing capacity, storage capacity, and hardware energy efficiency information.

[0113] In this embodiment, hardware processing power and storage capacity can be used, combined with model parameters, to determine whether the third node can meet the task requirements. Hardware processing power and hardware energy efficiency information can be used to determine the energy consumption required to execute the task. For example, hardware energy efficiency information includes at least one of the following: average processor power (the average power of the processor during operation, which can be measured) or power usage efficiency (the percentage of electrical energy consumed by the processor that is used for computation).

[0114] In this embodiment, the energy supply information includes an indication of the type of energy supplied to the third node. For example, the energy type can be divided into green energy and non-green energy. Due to factors such as natural conditions, green energy typically cannot be supplied around the clock, but rather has a limited availability time, while non-green energy can typically be supplied around the clock.

[0115] Optionally, in this embodiment, the energy supply information may also include the corresponding energy availability time, carbon emissions per unit of energy, etc., but this embodiment does not specifically limit this.

[0116] Optionally, the first information is also used to indicate the range information of the third node to be queried. The range information indicates the geographical range of the third node to be queried. By limiting the query to third nodes within a certain range, subsequent transmission overhead can be reduced.

[0117] S820: The second node sends the second information to the first node, and the first node receives the second information accordingly.

[0118] The second information indicates the processing capacity and / or power supply information of at least one third node.

[0119] Optionally, in this embodiment of the application, the second information includes the identification information of the third node and the corresponding processing capacity and / or power supply information.

[0120] Optionally, in this embodiment of the application, the second information may also indicate the range information of at least one third node.

[0121] Optionally, in this embodiment of the application, the second node can obtain the aforementioned second information from at least one third node.

[0122] S830, the first node determines at least one target third node for performing the task based on the second information.

[0123] Optionally, the task in this embodiment can be at least one of model training, model testing, model simulation, model loading, or model inference. After receiving the second information, the first node can select one or more third nodes with green energy supply, low energy consumption, and that meet the model task requirements as the target third node by comparing the processing capabilities and / or energy supply information of each third node. In other words, the energy consumption of the target third node is lower than that of at least one third node, and / or the energy type supplied by the target third node is green energy.

[0124] In one implementation, if there are multiple target third nodes, the target third nodes for task execution can be determined in time periods. For example, if the target third nodes include node A and node B, then node A can be determined to execute the task from 00:00 to 12:00, and node B can be determined to execute the task from 12:00 to 24:00. As another example, if target third node 1 has available green energy and no time limit, then target third node 1 is the actual target third node executing the task, and the execution time period may not be specified (the default is no time limit). In another scenario, if target third node 1 has available green energy from 10:00 AM to 4:00 PM, and target third node 2 also has available green energy from midnight to 10:00 AM and from 4:00 PM to midnight, then target third nodes 1 and 2 are determined to be the actual target third nodes executing the task, with target third node 1 executing from 10:00 AM to 24:00. At 4 PM, the execution period for target node 2 is from midnight to 10 AM and from 4 PM to midnight. In another scenario, target node 1 has green energy available from 10 AM to 4 PM, while other target nodes have no green energy available. In this case, the target nodes that actually perform the task are target node 1 and target node 2. Target node 2 is the target node with the lowest energy consumption for a single task. The execution period for target node 1 is from 10 AM to 4 PM, and the execution period for target node 2 is from midnight to 10 AM and from 4 PM to midnight.

[0125] It should be understood that the above description uses two target third nodes as an example. In actual implementation, the number of target third nodes can also be multiple, and the embodiments of this application do not limit this.

[0126] In one implementation, the first information is used to query the processing capacity of at least one third node. Correspondingly, the second information includes the identification information of the third node and its corresponding processing capacity, used to indicate the processing capacity of at least one third node. In this embodiment, the first node can estimate the energy consumption for executing a task. For example, single inference energy consumption = single inference execution time × average processor power × power efficiency. Wherein, single inference execution time = single inference computational load / processing speed. The processing speed depends on the hardware processing capacity; the stronger the hardware processing capacity, the faster the processing speed. Based on the above principle, after obtaining the hardware processing capacity of at least one third node, the first node can determine the single inference energy consumption of at least one third node based on its hardware processing capacity, and then select one or more third nodes with lower energy consumption as the target third node for executing the task.

[0127] In one implementation, the first information is used to query the energy supply information of at least one third node. Correspondingly, the second information includes the identification information of the third node and the corresponding energy supply information, used to indicate the energy supply information of at least one third node. In this embodiment, the first node can determine which energy source powers each third node based on the energy type in the energy supply information, and then select one or more third nodes powered by green energy as the target third node for performing the task.

[0128] Optionally, if the energy supply information also includes the energy availability time corresponding to the energy type, the target third node that is powered by which green energy source will perform the task in each time period can be determined based on the energy availability time. That is, the first node can determine the time when each target third node will perform the task based on the energy supply information in the second information. For example, if the target third nodes include node A and node B, and the energy availability time of node A is 00:00-12:00, and the energy availability time of node B is 12:00-24:00, then the first node can determine that node A will perform the task from 00:00-12:00, and node B will perform the task from 12:00-24:00.

[0129] In another implementation, the first information is used to query the processing capacity and power supply information of at least one third node. Correspondingly, the second information includes the identification information of the third node and its corresponding processing capacity and power supply information. In this case, the target third node for performing the task can be determined by comprehensively considering the power supply type and energy consumption.

[0130] For example, the first node can first identify candidate third nodes powered by green energy based on energy supply information, and then identify one or more candidate third nodes with lower energy consumption as the target third node to perform the task based on the processing capacity of the candidate third nodes.

[0131] For example, the first node can estimate the carbon emissions of a single inference attempt based on processing power and energy supply information, and select one or more third nodes with lower carbon emissions as target third nodes. Here, the carbon emissions per inference attempt (estimated) = execution time of a single inference attempt × average processor power × power efficiency × carbon emissions per unit of energy (indicated by energy supply information). The execution time of a single inference attempt = computational load per inference attempt / processing speed.

[0132] In another example, the first node can score at least one third node based on the energy type determined by energy supply information and the energy consumption determined by processing capacity. The node with the higher sum of scores is selected as the target third node for the task. For example, green energy scores +8 points, and non-green energy scores +4 points. The energy consumption score is calculated by taking the reciprocal of the energy consumption value and multiplying it by a common multiplier. For instance, if the energy consumption of a third node in a single inference operation is 25 (this is a general explanation and does not restrict the unit of energy consumption), then the reciprocal of 25 is 0.04, and the common multiplier is 100, resulting in a score of +4. Assuming third node A has an energy score of +8 and an energy consumption score of +1, its final score is +9. Similarly, assuming third node B has an energy score of +4 and an energy consumption score of +4, its final score is +8. Between third nodes A and B, third node A is preferentially selected as the target third node.

[0133] In another implementation, the first information, in addition to querying the processing capacity and / or power supply information of the third node, is also used to specify the range information of the third node to be queried. Correspondingly, the second information includes the identification information of the third node, range information, and at least one of the following: processing capacity and power supply information. In this case, the first node can preferentially determine third nodes within a certain range as candidate third nodes based on the range information, and then select the target third node from the candidate third nodes based on the processing capacity and / or power supply information. The rule for determining this "certain range" can be flexibly set. For example, locations less than a preset distance threshold from the second node are all considered "a certain range," or locations less than a preset distance threshold from the third node from which the execution result of the task needs to be obtained are all considered "a certain range," etc. This application embodiment does not specifically limit this. Determining the target third node based on the range information of the third node can ensure lower transmission overhead.

[0134] In this embodiment, the first node first queries the processing capacity and / or power supply information of the third node, and determines the target third node for performing the task based on the processing capacity and / or power supply information of the third node. This allows the selected target third node to perform the task with lower energy consumption (and / or lower carbon emissions), thereby improving energy efficiency during the task execution process.

[0135] In one embodiment, as shown in FIG9, the communication method provided in this application embodiment further includes the following steps S840-S890:

[0136] S840: The first node sends the third information to the second node, and the second node receives the third information accordingly.

[0137] The third information is used to indicate the third node (hereinafter referred to as the demand third node) that needs to obtain the execution result of the task and at least one target third node.

[0138] Optionally, third-party information can also be used to indicate the tasks that need to be performed.

[0139] For example, the third information includes the identification information of the demand third node and the identification information of at least one target third node. Optionally, the third information may also include the identification information or description information of the task to be performed.

[0140] Optionally, when there are multiple target third nodes, the third information can also be used to indicate the time for each target third node to perform its task.

[0141] S850, the second node sends the fourth information to the target third node, and the target third node receives the fourth information accordingly.

[0142] The fourth piece of information is used to indicate the third node of the requirement.

[0143] Optionally, the fourth piece of information can also be used to indicate the tasks that need to be performed.

[0144] In this embodiment of the application, the target third node can determine which third node the required third node is based on the fourth information.

[0145] Optionally, in this embodiment of the application, the target third node may also determine the model for performing the task based on the indicated task.

[0146] S860: The second node sends the fifth message to the third node that requires it, and the third node receives the fifth message accordingly.

[0147] The fifth piece of information is used to indicate the third node of the target.

[0148] Optionally, the fifth piece of information can also be used to indicate the tasks that need to be performed.

[0149] Optionally, when there are multiple target third nodes, the fifth information can also be used to indicate the time for each target third node to perform its task.

[0150] S870: The requesting third node sends the task parameters required to execute the task to the target third node, and the target third node executes the task parameters accordingly.

[0151] Among them, the third node of the demand determines the third node (i.e. the target third node) to execute the task based on the fifth information in S860, and sends the task parameters required to execute the task to it based on this.

[0152] S880, the target third node executes the task according to the task parameters and obtains the task execution result.

[0153] Once the target third node receives the task parameters, it can execute the task and obtain the task execution result.

[0154] S890, the target third node sends the task execution result to the demand third node, and correspondingly, the demand third node receives the task execution result.

[0155] When the third node of the requirement receives the execution result of the task, the task requirement of the third node of the requirement has been fulfilled.

[0156] Optionally, in this embodiment, when there are multiple target third nodes (i.e., the fifth information also indicates the execution time of each target third node), the requesting third node can send task parameters to the corresponding target third nodes at different execution times. For example, if the target third nodes include target third node A and target third node B, and the execution times for target third node A and target third node B are time period 1 and time period 2 respectively, then the requesting third node sends task parameters to target third node A in time period 1 and sends task parameters to target third node B in time period 2. Accordingly, target third node A executes the task according to the received task parameters in time period 1, obtains the execution result of the task, and feeds it back to the requesting third node in time period 1; target third node B executes the task according to the received task parameters in time period 2, obtains the execution result of the task, and feeds it back to the requesting third node in time period 2.

[0157] In this embodiment of the application, the target third node executes the task according to the task parameters and sends the execution result of the task to the requesting third node. Since the target third node has low energy consumption (and / or low carbon emissions) when executing the task, the energy efficiency during the execution of the task is improved.

[0158] In one embodiment, as shown in FIG10, the communication method provided in this application embodiment further includes the following steps S1001-S1003:

[0159] S1001, the first node sends the sixth message to the second node, and the second node receives the sixth message accordingly.

[0160] The sixth piece of information indicates the historical processing results of the task being invoked.

[0161] Optionally, the sixth piece of information may also indicate a data deviation threshold used to determine whether historical processing results for a task are available.

[0162] In this embodiment, the first node can flexibly design scenarios to trigger the sending of the sixth message. For example, if the first node detects a high degree of data similarity between historical tasks and tasks within the current region / time period, it triggers the sending of the sixth message. Optionally, the data deviation threshold can be determined based on the similarity of historical data.

[0163] S1002, the second node sends the sixth message to the target third node, and the target third node receives the sixth message accordingly.

[0164] Among them, the target third node (i.e. the third node that executes the task) stores historical task parameters and corresponding historical processing results.

[0165] S1003, the third node of the target determines the historical processing results of the target as the execution result of the task.

[0166] In this embodiment of the application, the historical task parameters corresponding to the target historical processing result match the current task parameters. In other words, the data difference between the historical task parameters corresponding to the target historical processing result and the current task parameters is less than the data deviation threshold.

[0167] In this embodiment of the application, using historical processing results to replace the actual task execution process can reduce the computing power required to execute the task, thereby improving energy efficiency during the task execution process.

[0168] In one embodiment, optionally, as shown in FIG11, the communication method provided in this application embodiment further includes the following steps S1004-S1006:

[0169] S1004, the first node sends the seventh message to the second node, and the second node receives the seventh message accordingly.

[0170] The seventh piece of information is used to indicate the storage of historical processing results.

[0171] Optionally, the seventh piece of information is also used to indicate the time and / or region where the results of the historical processing are stored.

[0172] S1005, the second node sends the seventh message to the target third node, and the target third node receives the seventh message accordingly.

[0173] S1006, the target third node stores the task parameters and corresponding historical processing results of historical tasks according to the seventh information.

[0174] The task parameters and corresponding historical processing results of the historical tasks stored in the target third node can be used to determine the execution result in step S1003.

[0175] As can be seen from the above embodiments, the embodiments of this application improve energy efficiency during task execution by selecting a third node with lower energy consumption and / or lower carbon emissions to perform the task. Furthermore, the embodiments of this application design a mechanism that uses historical processing results to replace the processing results of the current task, further reducing the necessary number of executions and thus further improving the energy efficiency of task execution.

[0176] The following uses the communication system shown in Figure 2 as an example of the communication architecture shown in Figure 3, and in conjunction with the embodiments described in Figures 8 to 11, to give several communication methods provided by the embodiments of this application, as shown in Figures 12 to 14 respectively.

[0177] In one possible implementation, this scheme trades transmission resources for energy efficiency. By acquiring hardware energy efficiency information at different inference locations, it selects the deployment location with the optimal energy efficiency to perform inference, as shown in Figure 12. This communication method includes the following steps:

[0178] S1210, NMS sends a query request message to EMS. Correspondingly, EMS receives the query request message.

[0179] The query request message includes first information, which is used to query the processing capabilities of at least one gNB.

[0180] In this embodiment, the processing capability includes at least one of the following: hardware processing capability, storage capability, hardware energy efficiency information, or range information. For a detailed explanation of the processing capability, please refer to the relevant content in step S810 of the embodiment shown in Figure 8, which will not be repeated here.

[0181] In this embodiment of the application, the first information may include at least one of the following: information for indicating the query of the hardware processing capability of the gNB, information for indicating the query of the storage capability of the gNB, information for indicating the query of the hardware energy efficiency information of the gNB, and information for indicating the query of the range information of the gNB.

[0182] S1220, EMS sends a query response message to NMS, and NMS receives the query response message accordingly.

[0183] The query response message includes second information, which indicates the processing capability of at least one gNB.

[0184] In this embodiment of the application, the second information may include, for example, the identification information of at least one gNB and at least one of the following: the hardware processing capability of at least one gNB, the storage capability of at least one gNB, or the hardware energy efficiency information of at least one gNB. It is understood that the type of information included in the second information depends on the query content of the first information. For example, if the first information is used to query the hardware processing capability information of a gNB, then the second information includes the identification information of the gNB and the corresponding hardware processing capability information. As another example, if the first information is used to query the hardware processing capability information and hardware energy efficiency information of a gNB, then the second information includes the identification information of the gNB, the corresponding hardware processing capability information, and the corresponding hardware energy efficiency information.

[0185] For details regarding the second information, please refer to the relevant content in step S820 of the embodiment shown in Figure 8, which will not be repeated here.

[0186] In one possible implementation, as described in this embodiment, the EMS may store the processing capacity of at least one gNB. In this case, after receiving a query request message, the EMS can send a query response message to the NMS.

[0187] In another possible implementation, in this embodiment, the EMS does not store the processing capabilities of at least one gNB. In this case, after receiving a query request message, the EMS needs to query the processing capabilities of at least one gNB. For example, as shown in Figure 12, the process includes: S1201, the EMS sends a query request message to at least one gNB (e.g., including the target gNB, the requesting gNB, and other gNBs). Correspondingly, at least one gNB receives the query request message from the EMS. S1202, at least one gNB sends a query response message to the EMS. Correspondingly, the EMS receives a query response message from at least one gNB. The relevant descriptions of the query request message and query response message can be found in steps S1210 and S1220, and will not be repeated here.

[0188] It should be noted that the query response message in step S1202 indicates the processing capability of the gNB that responded to the query response message. The processing capability of one or more gNBs indicated by the query response message in step S1220 will be uniformly explained here and will not be repeated below. For example, the second information in the query response message replied by the target gNB to the EMS indicates the processing capability of the target gNB; the second information in the query response message replied by the requesting gNB to the EMS indicates the processing capability of the requesting gNB; the second information in the query response messages replied by other gNBs to the EMS indicates the processing capability of other gNBs; the second information in the query response message replied by the EMS to the NMS indicates the processing capability of the target gNB, the processing capability of the requesting gNB, and the processing capability of other gNBs.

[0189] It should be understood that the target gNB is the gNB used to execute the task, the demand gNB is the gNB that needs to obtain the execution result of the task, and the other gNBs are gNBs other than the target gNB and the demand gNB, and are the gNBs queried by the query request message in S1210.

[0190] S1230, the NMS determines at least one target gNB for performing the task based on the query response message.

[0191] The explanation of how the NMS determines at least one target gNB for performing the task based on the query response message can be found in step S830 of the embodiment shown in Figure 8, and will not be repeated here.

[0192] S1240, NMS sends Task Request Message 1 to EMS, and EMS receives Task Request Message 1 from NMS accordingly.

[0193] The task request message 1 includes third information, which indicates the requesting gNB and at least one target gNB (Figure 12 illustrates this using a target gNB as an example). Optionally, the third information may also indicate the task. In this embodiment, gNBs other than the requesting gNB and the target gNB are referred to as other gNBs.

[0194] In this embodiment, the third information may include, for example, the task's identification information, the requirement gNB's identification information, and the target gNB's identification information. For instance, when the task is model loading, the task's identification information is the model's identification information, and the aforementioned task request message 1 can be replaced with model loading request 1. The explanation of the third information can be found in step S840 of the embodiment shown in FIG9, and will not be repeated here.

[0195] S1250, EMS sends task request message 2 to the target gNB, and the target gNB receives task request message 2 from EMS.

[0196] The task request message 2 includes fourth information, which includes the identifier information of the requesting gNB. Optionally, the fourth information may also include the identifier information of the task. For example, when the task is model loading, the identifier information of the task is the identifier information of the model, and the task request message 2 can be replaced with model loading request 2. The explanation of the fourth information can be found in step S850 of the embodiment shown in Figure 9, and will not be repeated here.

[0197] S1260, EMS sends a task notification message to the requesting gNB, and the requesting gNB receives the task notification message from EMS accordingly.

[0198] The task notification message includes fifth information, which includes the identifier information of the target gNB. Optionally, the fifth information may also include the identifier information of the task. For example, when the task is model loading, the identifier information of the task is the identifier information of the model, and the above-mentioned task notification message can also be replaced by a model loading notification. The explanation of the fifth information can be found in the relevant content of step S860 in the embodiment shown in Figure 9, and will not be repeated here.

[0199] S1270, the requesting gNB sends a task request message 3 to the target gNB, and the target gNB receives the task request message 3 from the requesting gNB.

[0200] The task request message 3 includes the task parameters required to execute the task. For example, when the task is model inference, the task request message 3 can be replaced with a model inference request. The description of the task parameters required to execute the task can be found in step S870 of the embodiment shown in Figure 9, and will not be repeated here.

[0201] S1280, the target gNB executes the task according to task request message 3 and obtains the task execution result.

[0202] Upon receiving the task parameters in task request message 3, the target gNB can execute the task and obtain the execution result. A related description can be found in step S880 of the embodiment shown in Figure 9, and will not be repeated here.

[0203] S1290, the target gNB sends the task execution result to the requesting gNB, and correspondingly, the requesting gNB receives the task execution result.

[0204] For example, when the task is model inference, the execution result of the above task can also be replaced by the model inference result. The explanation of the task execution result can be found in the relevant content of step S890 in the embodiment shown in Figure 9, and will not be repeated here.

[0205] In another possible implementation, similar to the embodiment shown in Figure 12, this scheme trades transmission resources for energy efficiency. By acquiring hardware energy efficiency information at different inference locations, it selects the deployment location with the best energy efficiency to perform inference. The difference from the embodiment shown in Figure 12 is, for example, that in this embodiment, in scenarios where green energy is used, the deployment location with the lowest carbon emissions can be selected to perform inference by acquiring hardware processing capabilities and green energy resource information at different inference locations. As shown in Figure 13, taking a target third node including target gNB1 and target gNB2, where target gNB1 is powered by green energy and target gNB2 is powered by non-green energy, the communication method includes the following steps:

[0206] S1310, NMS sends a query request message to EMS. Correspondingly, EMS receives the query request message.

[0207] The query request message includes first information, which is used to query the processing capacity and power supply information of at least one gNB.

[0208] In this embodiment, processing capability includes at least one of the following: hardware processing capability, storage capability, hardware energy efficiency information, or range information. Energy supply information includes energy type; optionally, it also includes energy availability time and / or carbon emissions per unit of energy. For further explanation of processing capability and energy supply information, please refer to step S810 in the embodiment shown in Figure 8; details will not be repeated here.

[0209] In this embodiment of the application, the first information may include at least one of the following: information for indicating the query of the hardware processing capability of the gNB, information for indicating the query of the storage capability of the gNB, information for indicating the query of the hardware energy efficiency information of the gNB, or information for indicating the query of the range information of the gNB; and the first information includes information for querying the energy type of the gNB.

[0210] S1320, EMS sends a query response message to NMS, and NMS receives the query response message accordingly.

[0211] The query response message includes second information, which indicates the processing capacity and power supply information of at least one gNB.

[0212] In this embodiment of the application, the second information may include, for example, the identification information of at least one gNB and at least one of the following: the power supply type of at least one gNB, the energy availability time of at least one gNB, the hardware processing capability of at least one gNB, the storage capability of at least one gNB, or the hardware energy efficiency information of at least one gNB. It is understood that the type of information included in the second information depends on the query content of the first information. For example, if the first information is used to query the hardware processing capability information of a gNB, then the second information includes the identification information of the gNB and the corresponding hardware processing capability information. As another example, if the first information is used to query the hardware processing capability information and hardware energy efficiency information of a gNB, then the second information includes the identification information of the gNB, the corresponding hardware processing capability information, and the corresponding hardware energy efficiency information. As yet another example, if the first information is used to query the hardware processing capability information, power supply type, and hardware energy efficiency information of a gNB, then the second information includes the identification information of the gNB, the corresponding hardware processing capability information, the corresponding power supply type, and the corresponding hardware energy efficiency information.

[0213] For details regarding the second information, please refer to the relevant content in step S820 of the embodiment shown in Figure 8, which will not be repeated here.

[0214] In one possible implementation, as described in this embodiment, the EMS may store the processing capacity and power supply information of at least one gNB. In this case, after receiving a query request message, the EMS can send a query response message to the NMS.

[0215] In another possible implementation, in this embodiment, the EMS does not store the processing capacity and power supply information of at least one gNB. In this case, after receiving a query request message, the EMS needs to query the processing capacity and power supply information of at least one gNB. For example, as shown in Figure 13, the process includes: S1301, the EMS sends a query request message to at least one gNB (e.g., including the target gNB, the demand gNB, and other gNBs). Correspondingly, at least one gNB receives the query request message from the EMS. S1302, at least one gNB sends a query response message to the EMS. Correspondingly, the EMS receives a query response message from at least one gNB. The relevant descriptions of the query request message and query response message can be found in steps S1310 and S1320, and will not be repeated here.

[0216] It should be noted that the query response message in step S1302 indicates the processing capacity and power supply information of the gNB that responded to the query response message. The query response message in step S1320 indicates the processing capacity and power supply information of one or more gNBs. For example, the second information in the query response message from the target gNB to the EMS indicates the processing capacity and power supply information of the target gNB; the second information in the query response message from the requesting gNB to the EMS indicates the processing capacity and power supply information of the requesting gNB; the second information in the query response messages from other gNBs to the EMS indicates the processing capacity and power supply information of other gNBs; the second information in the query response message from the EMS to the NMS indicates the processing capacity and power supply information of the target gNB, the requesting gNB, and other gNBs.

[0217] S1330, NMS determines at least one target gNB for performing the task based on the query response message.

[0218] The NMS selects at least one gNB with the lowest carbon emissions as the target gNB based on the processing capacity and energy supply information of each gNB in ​​the query response message. If there are multiple target gNBs and the energy supply information also includes the energy availability time corresponding to the energy type, the target gNB powered by green energy can be determined to perform the task in each time period based on the energy availability time.

[0219] For example, in this embodiment, the target gNB includes target gNB1 and target gNB2. Target gNB1 is powered by green energy, while target gNB2 is powered by non-green energy. The execution time of the task corresponding to target gNB1 is time period 1, and the execution time of the task corresponding to target gNB2 is a time period outside of time period 1 (e.g., time period 2).

[0220] The explanation of how NMS determines at least one target gNB for performing the task based on the query response message can be found in step S830 of the embodiment shown in Figure 8, and will not be repeated here.

[0221] S1340, NMS sends Task Request Message 1 to EMS, and EMS receives Task Request Message 1 from NMS accordingly.

[0222] The task request message 1 includes third information, which indicates the requesting gNB and at least one target gNB (e.g., target gNB1 and target gNB2 in Figure 13). Optionally, the third information may also indicate the task. In this embodiment, gNBs other than the requesting gNB and the target gNB are referred to as other gNBs. The explanation of the third information can be found in step S1240 of the embodiment shown in Figure 12, and will not be repeated here.

[0223] S1350, EMS sends task request message 2 to the target gNB, and the target gNB receives task request message 2 from EMS.

[0224] The task request message 2 includes fourth information, which includes the identifier of the requesting gNB. Optionally, the fourth information may also include the task identifier. For a description of the task request message 2, please refer to the relevant content in step S1250 of the embodiment shown in Figure 12, which will not be repeated here.

[0225] As described above, there may be multiple target gNBs. In this embodiment of the application, the target gNBs include target gNB1 and target gNB2. Then, EMS sends task request message 2 to target gNB1 and target gNB2 respectively. Correspondingly, target gNB1 and target gNB2 receive task request message 2 from EMS.

[0226] S1360, EMS sends a task notification message to the requesting gNB, and the requesting gNB receives the task notification message from EMS accordingly.

[0227] The task notification message includes fifth information, which includes the identification information of the target gNB (e.g., target gNB1 and target gNB2 in Figure 13). Optionally, the fifth information may also include the task's identification information. For a description of the task notification message, please refer to the relevant content in step S1260 of the embodiment shown in Figure 12, which will not be repeated here.

[0228] In time period 1, the embodiments of this application further include the following steps S1370a-S1390a:

[0229] S1370a, the requesting gNB sends a task request message 3 to the target gNB1 during time period 1, and correspondingly, the target gNB1 receives the task request message 3 from the requesting gNB during time period 1.

[0230] The task request message 3 includes the task parameters required to execute the task. A description of the task request message 3 can be found in step S1270 of the embodiment shown in Figure 12, and will not be repeated here.

[0231] S1380a, target gNB1 executes the task according to task request message 3 in time period 1 and obtains the task execution result.

[0232] Upon receiving the task parameters carried in the task request message 3, the target gNB1 can execute the task and obtain the execution result. A related description can be found in step S880 of the embodiment shown in Figure 8, and will not be repeated here.

[0233] S1390a, the target gNB1 sends the task execution result to the requesting gNB during time period 1, and correspondingly, the requesting gNB receives the task execution result from the target gNB1 during time period 1.

[0234] The relevant description of S1390 can be found in step S1290 of the embodiment shown in Figure 12, and will not be repeated here.

[0235] Similarly, in time period 2, the embodiments of this application also include the following steps S1370b-S1390b:

[0236] S1370b, the requesting gNB sends a task request message 3 to the target gNB2 during time period 2, and correspondingly, the target gNB2 receives the task request message 3 from the requesting gNB during time period 2.

[0237] S1380b, target gNB2 executes the task according to task request message 3 in time period 2 and obtains the task execution result.

[0238] S1390b, the target gNB2 sends the task execution result to the requesting gNB in ​​time period 2, and correspondingly, the requesting gNB receives the task execution result from the target gNB2 in time period 2.

[0239] It should be understood that, compared to S1370a-S1390a executed in time period 1, although the target gNB for S1370b-S1390b ​​in time period 2 changes from target gNB1 to target gNB2, the implementation principle of each step is the same. The relevant explanations of each step in time period 2 can be found in the explanations of the corresponding steps in time period 1, and will not be repeated here.

[0240] In another possible implementation, this scheme trades storage resources for energy efficiency by storing historical inference results. When the input data deviation is less than a set threshold, the historical inference results are directly used instead, reducing the number of inference executions. As shown in Figure 14, this communication method may include:

[0241] S1410, the NMS sends a query request message to the EMS. Correspondingly, the EMS receives the query request message from the NMS.

[0242] The description of the query request message can be found in step S1210 in Figure 12 or step S1310 in Figure 13, and will not be repeated here.

[0243] S1420, EMS sends a query response message to NMS, and correspondingly, NMS receives the query response message from EMS.

[0244] The explanation of the query response message can be found in step S1220 in Figure 12 or step S1320 in Figure 13, and will not be repeated here.

[0245] Optionally, the communication method may further include: S1401, the EMS sends a query request message to at least one gNB (e.g., including a target gNB, a demand gNB, and other gNBs). Correspondingly, at least one gNB receives the query request message from the EMS. S1402, at least one gNB sends a query response message to the EMS. Correspondingly, the EMS receives a query response message from at least one gNB. The descriptions of S1401 and S1402 can be found in the descriptions of steps S1201 and S1202 in Figure 12 or steps S1301 and S1302 in Figure 13, and will not be repeated here.

[0246] S1430, the NMS determines at least one target gNB for performing the task based on the query response message.

[0247] The implementation of step S1430 can be found in the description of step S1230 in Figure 12 or step S1330 in Figure 13, and will not be repeated here.

[0248] S1440, NMS sends Task Request Message 1 to EMS, and EMS receives Task Request Message 1 from NMS accordingly.

[0249] The task request message 1 includes third information and sixth information. The explanation of the third information can be found in step S1240 of the embodiment described in Figure 12, and will not be repeated here. The sixth information indicates the historical processing results of the task. Optionally, the sixth information may also indicate a data deviation threshold used to determine whether the historical processing results of the task are usable. For example, the sixth information may include a data deviation threshold and a historical call indication message. The explanation of the sixth information can be found in step S1001 of Figure 10, and will not be repeated here.

[0250] The description of task request message 1 can be found in step S1240 in Figure 12 or step S1340 in Figure 13, and will not be repeated here.

[0251] S1450, EMS sends task request message 2 to the target gNB, and the target gNB receives task request message 2 from EMS.

[0252] Task request message 2 includes sixth information and fourth information. The explanation of the sixth information can be found in the relevant explanation of S1440, and will not be repeated here. The explanation of the fourth information and task request message 2 can be found in step S1250 in Figure 12 or step S1350 in Figure 13, and will not be repeated here.

[0253] S1460, EMS sends a task notification message to the requesting gNB, and the requesting gNB receives the task notification message from EMS accordingly.

[0254] The description of the task notification message can be found in step S1260 in Figure 12 or step S1360 in Figure 13, and will not be repeated here.

[0255] S1470, the requesting gNB sends a task execution request message 3 to the target gNB, and the target gNB receives the task request message 3 from the requesting gNB.

[0256] The description of task request message 3 can be found in step S1270 in Figure 12 or step S1370 in Figure 13, and will not be repeated here.

[0257] S1480, the target gNB determines the historical processing results of the target as the execution result of the task.

[0258] Wherein, the data difference between the historical task parameters corresponding to the target historical processing result and the task parameters in task request message 3 in S1470 is less than the data deviation threshold, the process of determining the execution result of the task can be referred to the explanation of step S1003 in Figure 10, and will not be repeated here.

[0259] S1490, the target gNB sends the execution result of the task to the requesting gNB, and correspondingly, the requesting gNB receives the execution result of the task from the target gNB.

[0260] It should be understood that the message names in each step of Figures 12 to 14 are exemplary descriptions, and in actual implementation, they can be other names without restriction.

[0261] Optionally, the communication system shown in FIG2 can also be used in the communication architecture shown in FIG4-7 in this embodiment. The interaction process related to the communication method of this embodiment can be referred to the embodiments described in FIG12 to FIG14 above. Compared with the embodiments shown in FIG12, FIG13, or FIG14, the difference is that the execution subject of each step in the embodiments described in FIG12 to FIG14 is different in this embodiment. The following is an explanation with reference to examples.

[0262] For example, when the communication system shown in Figure 2 is used in the communication architecture shown in Figure 4, in the embodiments described in Figures 8 to 11, the first node can be a Non-RT RIC, and the Non-RT RIC performs the steps performed by the first node. The second and third nodes are both Near-RT RICs, and the Near-RT RIC performs the steps performed by the second and third nodes.

[0263] In another example, when the communication system shown in Figure 2 is used in the communication architecture shown in Figure 5, the first node in the embodiments described in Figures 8 to 11 can be a Non-RT RIC, which executes the steps performed by the first node; the second node is a Near-RT RIC, which executes the steps performed by the second node; and the third node is an O-CU / O-DU / O-RU, which executes the steps performed by the third node.

[0264] In another example, when the communication system shown in Figure 2 is used in the communication architecture shown in Figure 6, the first node in the embodiments described in Figures 8 to 11 can be a MIF, which executes the steps performed by the first node. The second and third nodes can both be CUs / DUs in the gNB, which execute the steps performed by the second and third nodes.

[0265] In another example, when the communication system shown in Figure 2 is used in the communication architecture shown in Figure 7, the first node in the embodiments described in Figures 8 to 11 can be a MIF, which executes the steps executed by the first node; the second node is a CU in the gNB, which executes the steps executed by the second node; and the third node is a DU in the gNB, which executes the steps executed by the third node.

[0266] As can be seen from the implementation methods described above, this application improves energy efficiency during task execution by selecting model inference nodes with lower energy consumption and / or lower carbon emissions to perform tasks. Furthermore, this application design incorporates a mechanism that uses historical processing results to replace the current task's processing result, further reducing the necessary number of executions and thus further enhancing the energy efficiency of task execution.

[0267] The foregoing mainly describes the solution provided by the embodiments of this application from the perspective of the execution logic of each step. It is understood that each node, such as the first node, includes the corresponding hardware structure and / or software module for executing each function in order to achieve the above-mentioned functions. Those skilled in the art should readily recognize that, in conjunction with the algorithm steps of the examples described in the embodiments disclosed herein, the method of the embodiments of this application can be implemented in hardware, software, or a combination of hardware and computer software. Whether a function is executed in a hardware or computer software-driven hardware manner depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0268] This application embodiment can divide the first node into functional modules according to the above method example. For example, each function can be divided into its own functional module, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware or as a software functional module. It should be noted that the module division in this application embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods.

[0269] In specific implementations, each node shown in this application may adopt the composition structure shown in Figure 15 or include the components shown in Figure 15. Figure 15 is a schematic diagram of the structure of a communication device provided in an embodiment of this application. When the communication device has the function of the first node described in the embodiment of this application, the communication device may be the first node or a chip or system-on-a-chip in the first node. When the communication device has the function of the second node described in the embodiment of this application, the communication device may be the second node or a chip or system-on-a-chip in the second node.

[0270] As shown in Figure 15, the communication device may include a processor 1501, a communication line 1502, a transceiver 1503, and a memory 1504. The processor 1501, memory 1504, and transceiver 1503 can be connected via the communication line 1502. In one example, the processor 1501 may include one or more CPUs, such as CPU0 and CPU1 in Figure 15.

[0271] As an alternative implementation, the communication device may include multiple processors, for example, in addition to processor 1501 in FIG15, it may also include processor 1507.

[0272] The processor 1501 can be a central processing unit (CPU), a network processor (NP), a digital signal processor (DSP), a microprocessor, a microcontroller, a programmable logic device (PLD), or any combination thereof. The processor 1501 can also be other devices with processing capabilities, such as circuits, devices, or software modules.

[0273] Communication line 1502 is used to transmit information between the components included in the communication device.

[0274] Transceiver 1503 is used to communicate with other devices or other communication networks. These other communication networks can be Ethernet, radio access network (RAN), wireless local area network (WLAN), etc. Transceiver 1503 can be an interface circuit, pins, RF module, transceiver, or any device capable of enabling communication.

[0275] Furthermore, the communication device may also include a memory 1504. The memory 1504 is used to store instructions. These instructions may be computer programs.

[0276] The memory 1504 can be a read-only memory (ROM) or other type of static storage device that can store static information and / or instructions; it can also be a random access memory (RAM) or other type of dynamic storage device that can store information and / or instructions; it can also be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage, magnetic disk storage medium or other magnetic storage device; optical disc storage includes compressed optical discs, laser discs, optical discs, digital universal optical discs, or Blu-ray discs, etc.

[0277] It should be noted that the memory 1504 can exist independently of the processor 1501, or it can be integrated with the processor 1501. The memory 1504 can be used to store instructions, program code, or some data, etc. The memory 1504 can be located inside or outside the communication device, without limitation. When the processor 1501 executes the instructions stored in the memory 1504, it can implement the method provided in the embodiments of this application.

[0278] As an optional implementation, the communication device also includes an output device 1505 and an input device 1506. For example, the input device 1506 is a device such as a keyboard, mouse, microphone, or joystick, and the output device 1505 is a device such as a display screen or speaker.

[0279] It should be noted that the communication device can be a desktop computer, a portable computer, a web server, a mobile phone, a tablet computer, a wireless terminal, an embedded device, a chip system, or a device with a similar structure to that shown in Figure 15. Furthermore, the composition shown in Figure 15 does not constitute a limitation on the communication device. In addition to the components shown in Figure 15, the communication device may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0280] In this embodiment of the application, the chip system may be composed of chips or may include chips and other discrete devices.

[0281] Figure 16 shows a structural diagram of a communication device 160 applied to a first node. Each module in the device shown in Figure 16 has the function of implementing the execution steps of the first node in Figures 8-14, and can achieve its corresponding technical effects. The beneficial effects of each module's execution steps can be referred to the descriptions of the corresponding steps in Figures 8-14, and will not be repeated here. The functions can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above functions. For example, the communication device includes:

[0282] The transceiver module 1601 is used to send first information to the second node, the first information being used to query the processing capacity and / or power supply information of at least one third node; the transceiver module 1601 is also used to receive second information, the second information indicating the processing capacity and / or power supply information of at least one third node; the processing module 1602 is used to determine at least one target third node for performing the task based on the second information.

[0283] Figure 17 shows a structural diagram of a communication device 170 applied to a second node. Each module in the device shown in Figure 17 has the function of implementing the execution steps of the second node in Figures 8-14, and can achieve its corresponding technical effects. The beneficial effects of each module's execution steps can be referred to the descriptions of the corresponding steps in Figures 8-14, and will not be repeated here. The functions can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above functions. For example, the communication device includes:

[0284] The transceiver module 1701 is used to receive first information, which is used to query the processing capacity and / or power supply information of at least one third node; the processing module 1702 is used to determine second information based on the first information, which indicates the processing capacity and / or power supply information of at least one third node; the transceiver module 1701 is also used to send the second information.

[0285] Figure 18 shows a structural diagram of a communication device 180 applied to a second node. Each module in the device shown in Figure 18 has the function of implementing the execution steps of the second node in Figures 8-14, and can achieve its corresponding technical effects. The beneficial effects of each module's execution steps can be referred to the descriptions of the corresponding steps in Figures 8-14, and will not be repeated here. The functions can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above functions. For example, the communication device includes:

[0286] The transceiver module 1801 is used to receive fourth information, wherein the fourth information indicates a third node that needs to obtain the execution result of the task; the transceiver module 1801 is also used to receive task parameters required to execute the task; the processing module 1802 is used to execute the task according to the task parameters and obtain the execution result of the task; the transceiver module 1801 is also used to send the execution result to the third node that needs to obtain the execution result of the task.

[0287] This application also provides a communication system, which includes a first node and a second node. The first node may have the functions of the aforementioned communication device 160, and the second node may have the functions of the aforementioned communication device 170.

[0288] In one embodiment, the communication system may further include a third node, which may have the functions of the communication device 180 described above.

[0289] This application also provides a computer-readable storage medium. All or part of the processes in the above method embodiments can be implemented by a computer program instructing related hardware. This program can be stored in the computer-readable storage medium, and when executed, it can include the processes of the above method embodiments. The computer-readable storage medium can be a terminal device of any of the foregoing embodiments, such as an internal storage unit including a data sending end and / or a data receiving end, such as a hard disk or memory of the terminal device. The computer-readable storage medium can also be an external storage device of the terminal device, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the terminal device. Further, the computer-readable storage medium can include both the internal storage unit and the external storage device of the terminal device. The computer-readable storage medium is used to store the computer program and other programs and data required by the terminal device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0290] This application also provides computer instructions. All or part of the processes in the above method embodiments can be executed by computer instructions to instruct related hardware (such as computers, processors, network devices, and terminals). The program can be stored in the aforementioned computer-readable storage medium.

[0291] This application also provides a chip system. The chip system can be composed of chips or may include chips and other discrete devices, without limitation. The chip system includes a processor and a transceiver. All or part of the processes in the above method embodiments can be completed by this chip system. For example, the chip system can be used to implement the function performed by the first node in the above method embodiments, or to implement the function performed by the second node in the above method embodiments.

[0292] In one possible design, the chip system further includes a memory for storing program instructions and / or data. When the chip system is running, the processor executes the program instructions stored in the memory to cause the chip system to perform the function performed by the first node in the above method embodiment or to perform the function performed by the second node in the above method embodiment.

[0293] In the embodiments of this application, the processor may be a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components, capable of implementing or executing the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly manifested as being executed by a hardware processor, or executed by a combination of hardware and software modules within the processor.

[0294] In the embodiments of this application, the memory can be non-volatile memory, such as a hard disk drive (HDD) or a solid-state drive (SSD), or it can be volatile memory, such as random-access memory (RAM). Memory is any other medium capable of carrying or storing desired program code in the form of instructions or data structures, and accessible by a computer, but is not limited thereto. The memory in the embodiments of this application can also be a circuit or any other device capable of implementing storage functions, used to store instructions and / or data.

[0295] It should be noted that the terms "first" and "second," etc., in the specification, claims, and drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0296] It should be understood that in the embodiments of this application, "at least one (item)" refers to one or more, "more than one" refers to two or more, "at least two (items)" refers to two or three or more, and "and / or" is used to describe the association relationship of related objects, indicating that there can be three relationships. For example, "A and / or B" can represent: only A exists, only B exists, and A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the related objects before and after are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple. It should be understood that in the embodiments of this application, "B corresponding to A" means that B is associated with A. For example, B can be determined based on A. It should also be understood that determining B based on A does not mean determining B solely based on A; B can also be determined based on A and / or other information. Furthermore, the term "connection" in the embodiments of this application refers to various connection methods, such as direct or indirect connections, to achieve communication between devices; the embodiments of this application do not impose any limitations on this.

[0297] Unless otherwise specified, the term "transmission" in the embodiments of this application refers to bidirectional transmission, encompassing the actions of sending and / or receiving. Specifically, "transmission" in the embodiments of this application includes sending data, receiving data, or both sending and receiving data. In other words, data transmission here includes uplink and / or downlink data transmission. Data may include channels and / or signals; uplink data transmission refers to uplink channel and / or uplink signal transmission, and downlink data transmission refers to downlink channel and / or downlink signal transmission. The terms "network" and "system" in the embodiments of this application refer to the same concept; a communication system is a communication network.

[0298] Through the above description of the embodiments, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0299] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0300] The units described as separate components may or may not be physically separate. A component shown as a unit can be one or more physical units; that is, it can be located in one place or distributed in multiple different locations. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0301] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a device, such as a microcontroller, chip, or processor, to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, mobile hard drives, ROM, RAM, magnetic disks, or optical disks. The above descriptions are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of protection of the claims.

Claims

1. A communication method, characterized in that, Applied to the first node, including: Send first information to the second node, the first information being used to query the processing capacity and / or power supply information of at least one third node; Receive second information, the second information indicating the processing capacity and / or power supply information of the at least one third node; Based on the second information, at least one target third node is determined for performing the task.

2. The communication method according to claim 1, characterized in that, The method further includes: Send a third message to the second node, wherein the third message is used to indicate the third node that needs to obtain the execution result of the task and the at least one target third node.

3. The communication method according to claim 1 or 2, characterized in that, The second information includes the identification information of the at least one third node and the corresponding power supply information; the method further includes: The time for each of the target third nodes to perform the task is determined based on the energy supply information in the second information.

4. The communication method according to any one of claims 1-3, characterized in that, The first information is also used to indicate the range information of the third node to be queried.

5. The communication method according to any one of claims 1-4, characterized in that, The method further includes: A sixth message is sent to the second node, wherein the sixth message indicates the invocation of the historical processing results of the task.

6. The communication method according to any one of claims 1-5, characterized in that, The task includes at least one of the following: model training, model testing, model simulation, model loading, or model inference.

7. The communication method according to any one of claims 1-6, characterized in that, The processing capability includes at least one of the following: Hardware processing power, storage capacity, or hardware energy efficiency information; The power supply information includes at least one of the following: The type of energy supplied, the time when the energy is available, or the carbon emissions per unit of energy.

8. A communication method, characterized in that, Applied to the second node, including: Receive first information, which is used to query the processing capacity and / or power supply information of at least one third node; The second information is determined based on the first information, and the second information indicates the processing capacity and / or power supply information of the at least one third node; Send the second message.

9. The communication method according to claim 8, characterized in that, The method further includes: Receive third information, the third information being used to indicate a third node that needs to obtain the execution result of the task and a target third node for executing the task; A fourth message is sent to the target third node that performs the task, wherein the fourth message indicates the third node that needs to obtain the execution result of the task.

10. The communication method according to claim 9, characterized in that, The method further includes: Receive a sixth message, wherein the sixth message indicates the retrieval of historical processing results; The sixth message is sent to the target third node.

11. The communication method according to any one of claims 8-10, characterized in that, The first information is also used to indicate the range information of the third node to be queried.

12. The communication method according to any one of claims 8-11, characterized in that, The method further includes: Obtain the processing capacity and / or power supply information of the at least one third node.

13. A communication method, characterized in that, The third node applied to the execution of tasks includes: Receive fourth information, wherein the fourth information indicates a third node that needs to obtain the execution result of the task; Receive the task parameters required to execute the task; The task is executed according to the task parameters to obtain the execution result of the task; Send the execution result to the third node of the task described in the requirements.

14. The communication method according to claim 13, characterized in that, The third node executing the task stores historical task parameters and corresponding historical processing results; The step of executing the task according to the task parameters and obtaining the execution result of the task includes: The target historical processing result is determined as the execution result of the task, wherein the historical task parameters corresponding to the target historical processing result match the task parameters.

15. A communication device, characterized in that, Applied to the first node, including: The transceiver module is used to send first information to the second node, wherein the first information is used to query the processing capacity and / or power supply information of at least one third node; The transceiver module is also used to receive second information, the second information indicating the processing capacity and / or power supply information of the at least one third node; The processing module is configured to determine at least one target third node for performing the task based on the second information.

16. The communication device according to claim 15, characterized in that, The device further includes: The transceiver module is used to send third information to the second node, wherein the third information is used to indicate the third node that needs to obtain the execution result of the task and the at least one target third node.

17. The communication device according to claim 15 or 16, characterized in that, The second information includes the identification information of the at least one third node and the corresponding power supply information; the processing module is further configured to: The time for each of the target third nodes to perform the task is determined based on the energy supply information in the second information.

18. The communication device according to any one of claims 15-17, characterized in that, The first information is also used to indicate the range information of the third node to be queried.

19. The communication device according to any one of claims 15-18, characterized in that, The transceiver module is also used for: A sixth message is sent to the second node, wherein the sixth message indicates the invocation of the historical processing results of the task.

20. The communication device according to any one of claims 15-19, characterized in that, The task includes at least one of the following: model training, model testing, model simulation, model loading, or model inference.

21. The communication device according to any one of claims 15-20, characterized in that, The processing capability includes at least one of the following: Hardware processing power, storage capacity, or hardware energy efficiency information; The power supply information includes at least one of the following: The type of energy supplied, the time when the energy is available, or the carbon emissions per unit of energy.

22. A communication device, characterized in that, Applied to the second node, including: The transceiver module is used to receive first information, which is used to query the processing capacity and / or power supply information of at least one third node. A processing module is configured to determine second information based on the first information, wherein the second information indicates the processing capacity and / or power supply information of the at least one third node; The transceiver module is also used to send the second information.

23. The communication device according to claim 22, characterized in that, The transceiver module is also used to receive third information, which is used to indicate a third node that needs to obtain the execution result of the task and a target third node for executing the task. The transceiver module is also used to send a fourth message to the target third node that performs the task, wherein the fourth message indicates the third node that requests the task.

24. The communication device according to claim 23, characterized in that, The transceiver module is also used for: Receive a sixth message, wherein the sixth message indicates the retrieval of historical processing results; The sixth message is sent to the target third node.

25. The communication device according to any one of claims 22-24, characterized in that, The first information is also used to indicate the range information of the third node to be queried.

26. The communication device according to any one of claims 22-25, characterized in that, The processing module is also used for: The transceiver module obtains the processing capacity and / or power supply information of the at least one third node.

27. A communication device, characterized in that, The third node applied to the execution of tasks includes: A transceiver module is used to receive fourth information, wherein the fourth information indicates a third node that requests the task; The transceiver module is also used to receive task parameters required for executing the task; The processing module is used to execute the task according to the task parameters and obtain the execution result of the task; The transceiver module is also used to send the execution result to the third node that requests the task.

28. A communication device, characterized in that, The communication device includes a processor and a transceiver, the processor and the transceiver being configured to support the communication device in performing the method as described in any one of claims 1-14.

29. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed, perform the method as described in any one of claims 1-14.

30. A communication system, characterized in that, The communication system includes a first node and a second node, or the communication system includes a first node, a second node and a third node, wherein the first node is used to perform the method as described in any one of claims 1-7, the second node is used to perform the method as described in any one of claims 8-12, and the third node is used to perform the method as described in claim 13 or 14.

Images