Task scheduling method and communication device

Abstract

A task scheduling method and a communication device. The method comprises: acquiring the computing capability of at least one computing execution node; on the basis of the computing capability and service quality requirements for a computing task, determining the correspondence between each of first computing execution nodes and a task to be executed and the correspondence between a second computing execution node and the task to be executed, wherein the first computing execution nodes include nodes used for executing the computing task among the at least one computing execution node, the second computing execution node is any node among the first computing execution nodes, and the task to be executed is an entirety or part of the computing task; sending a first correspondence to a computing request node; and / or sending a second correspondence to the second computing execution node. By determining the correspondence between each of the computing execution nodes and the task to be executed on the basis of the computing capabilities of the computing execution nodes and the service quality requirements for the computing task, the computing task can be scheduled to a proper computing execution node for execution, thereby improving the utilization rate of computing resources.

Description

A task scheduling method and communication device

[0001] This application claims priority to the Chinese patent application filed with the State Intellectual Property Office of China on September 15, 2023, with application number 202311200893.X, and priority to the Chinese patent application entitled “A method and communication device for task scheduling”, all contents of which are incorporated by reference into this application. Technical Field

[0002] The present application relates to the field of communications, and more particularly, to a task scheduling method and a communication device. Background Art

[0003] As fifth-generation (5G) communication systems and future services become more immersive and intelligent, networks will need to support the scheduling of artificial intelligence (AI) resources and the processing of vast amounts of data to reduce terminal computing workloads and energy consumption. This requires the network to be aware of ubiquitous computing and services and dispatch computing tasks to the optimal execution nodes for processing, thereby improving the utilization of network and computing resources and enhancing the user experience. In this context, the question of how the network can properly schedule tasks is crucial.

[0004] Summary of the Invention

[0005] The present application provides a task scheduling method and a communication device, which can perform reasonable task scheduling, thereby improving computing resource utilization and enhancing user experience.

[0006] In a first aspect, a method for task scheduling is provided. The method can be performed by a computing management function network element, or by a module (e.g., a chip or circuit) of the computing management function network element, or by a logical node, logical module, or software that can implement all or part of the computing management function network element. This application is not limited to this.

[0007] The method includes: obtaining the computing capacity of each computing execution node in at least one computing execution node; determining a first corresponding relationship and / or a second corresponding relationship based on the computing capacity and the service quality requirement of the computing task, the first corresponding relationship including the corresponding relationship between each computing execution node in the first computing execution node and the task to be executed, the first computing execution node including one or more computing execution nodes in the at least one computing execution node for executing the computing task, the second corresponding relationship including the corresponding relationship between the second computing execution node and the task to be executed, the second computing execution node being any computing execution node in the first computing execution node, and the task to be executed being all or part of the tasks in the computing task; sending the first corresponding relationship to the computing request node; and / or sending the second corresponding relationship to the second computing execution node.

[0008] Based on the above scheme, by obtaining the computing power of the computing execution node and determining the correspondence between the computing execution node and the task to be executed based on the computing power and the service quality requirements of the computing task, the computing task can be scheduled to the appropriate computing execution node for execution, thereby improving the utilization rate of computing resources.

[0009] In combination with the first aspect, in certain implementations of the first aspect, the computing capability indicates first information, which includes at least one of the following: performance information of performing tests based on at least one test data set, performance information of performing tests based on at least one model, and performance information of performing tests based on at least one test tool or test plan.

[0010] Based on the above solution, the computing capability of the reasonable computing execution node can be reported by indicating at least one item of the first information through computing capability.

[0011] In combination with the first aspect, in certain implementations of the first aspect, the computing capability further indicates at least one moment, and the at least one moment corresponds to the first information.

[0012] Based on the above solution, by reporting the first information corresponding to at least one moment, the computing management function network element can obtain the current computing capacity based on the historical computing capacity of the computing execution node.

[0013] In combination with the first aspect, in certain implementations of the first aspect, the performance information of the test performed based on at least one test data set includes at least one of the following: identification information of the at least one test data set, the number of training samples per unit time based on the at least one test data set as a training data set, the time it takes to complete the AI ​​model training based on the at least one test data set as a training data set, or the accuracy of training the AI ​​model based on the at least one test data set as a training data set.

[0014] In combination with the first aspect, in certain implementations of the first aspect, the performance information of the test performed based on at least one model includes at least one of the following: identification information of the at least one model, the time required to obtain the inference result based on the at least one model, or the number of times the at least one model is executed in unit time to obtain the inference result.

[0015] In combination with the first aspect, in certain implementations of the first aspect, the performance information of the test performed according to at least one test tool or test scheme includes at least one of the following: identification information of the at least one test tool or test scheme, the rate of encoding performed according to the at least one test tool or test scheme, the duration of encoding performed according to the at least one test tool or test scheme, the rate of decoding performed according to the at least one test tool or test scheme, or the duration of decoding performed according to the at least one test tool or test scheme.

[0016] In combination with the first aspect, in certain implementations of the first aspect, the quality of service requirement includes at least one of the following information about the computing task: task type, computing type, input data rate or data volume, output data rate or data volume, latency requirement, computing power requirement, and memory requirement.

[0017] In combination with the first aspect, in some implementations of the first aspect, the method further includes: receiving a computing service request message from the computing request node, where the computing service request message includes the service quality requirement.

[0018] In combination with the first aspect, in some implementations of the first aspect, the computing task includes at least one subtask, and the computing service request message further includes association relationship information between the at least one subtask.

[0019] Based on the above solution, by carrying the association relationship between subtasks in the computing service request message, the computing management function network element can determine the computing execution node that executes at least one subtask based on the association relationship information, thereby reasonably allocating computing resources.

[0020] In combination with the first aspect, in certain implementations of the first aspect, the association relationship information includes at least one of the following: data flow direction, input or output data rate, input or output data throughput, and input-output dependency.

[0021] In combination with the first aspect, in certain implementations of the first aspect, the computing request node is a terminal device, and the computing service request message is carried in any of the following signaling: wireless resource control RRC signaling, signaling of the first protocol layer, computing resource control CRC signaling; wherein, the first protocol layer is located above the wireless resource control RRC protocol layer or in parallel with the RRC protocol layer.

[0022] In combination with the first aspect, in certain implementations of the first aspect, the first correspondence is carried in any one of the following messages: downlink control information DCI, media access control MAC control element CE, radio resource control RRC message, packet data convergence layer protocol PDCP control protocol data unit PDU, signaling of the first protocol layer, computing resource control CRC signaling; wherein, the first protocol layer is located above the radio resource control RRC protocol layer or in parallel with the RRC protocol layer.

[0023] In combination with the first aspect, in certain implementations of the first aspect, the computing execution node is an application server, and the second correspondence is carried in any one of the following messages: the tunnel protocol user plane part GTP-U header, the application programming interface API interface message of the network capability exposure function NEF.

[0024] In combination with the first aspect, in certain implementations of the first aspect, the method is applied to an access network device, an access network device distributed unit DU, or an access network device centralized unit CU.

[0025] In a second aspect, a method for task scheduling is provided. The method can be executed by a computing execution node, or by a module (e.g., a chip or circuit) of the computing execution node, or by a logic node, logic module, or software that can implement all or part of the computing execution node. This application is not limited to this.

[0026] The method includes: sending computing capacity to a computing management function network element, wherein the computing capacity is used to determine a second corresponding relationship, wherein the second corresponding relationship includes a corresponding relationship between a computing execution node and a task to be executed, wherein the task to be executed is all or part of the computing task; and receiving the second corresponding relationship from the computing management function network element.

[0027] Based on the above scheme, by obtaining the computing power of the computing execution node and determining the correspondence between the computing execution node and the task to be executed based on the computing power, the computing task can be scheduled to the appropriate computing execution node for execution, thereby improving the utilization rate of computing resources.

[0028] In combination with the second aspect, in certain implementations of the second aspect, the computing capability indicates first information, which includes at least one of the following: performance information of performing tests based on at least one test data set, performance information of performing tests based on at least one model, and performance information of performing tests based on at least one test tool or test plan.

[0029] In combination with the second aspect, in certain implementations of the second aspect, the computing capability further indicates at least one moment, and the at least one moment corresponds to the first information.

[0030] In combination with the second aspect, in certain implementations of the second aspect, the performance information of the test performed based on at least one test data set includes at least one of the following: identification information of the at least one test data set, the number of training samples per unit time based on the at least one test data set as a training data set, the time required to complete the AI ​​model training based on the at least one test data set as a training data set, or the accuracy of training the AI ​​model based on the at least one test data set as a training data set.

[0031] In combination with the second aspect, in certain implementations of the second aspect, the performance information of the test performed based on at least one model includes at least one of the following: identification information of the at least one model, the time required to obtain the inference result based on the at least one model, or the number of times the at least one model is executed in unit time to obtain the inference result.

[0032] In combination with the second aspect, in certain implementations of the second aspect, the performance information of the test performed according to at least one test tool or test scheme includes at least one of the following: identification information of the at least one test tool or test scheme, the rate of encoding performed according to the at least one test tool or test scheme, the duration of encoding performed according to the at least one test tool or test scheme, the rate of decoding performed according to the at least one test tool or test scheme, or the duration of decoding performed according to the at least one test tool or test scheme.

[0033] In combination with the second aspect, in certain implementations of the second aspect, the computing execution node is an application server, and the second correspondence is carried in any one of the following messages: the tunnel protocol user plane part GTP-U header, the application programming interface API interface message of the network capability exposure function NEF.

[0034] In a third aspect, a method for task scheduling is provided. The method can be executed by a computation requesting node, or by a module (e.g., a chip or circuit) of the computation requesting node, or by a logical node, logical module, or software that can implement all or part of a computation execution node. This application is not limited to this.

[0035] The method includes: sending a service quality requirement of a computing task to a computing management function network element, wherein the service quality requirement is used to determine a first corresponding relationship, wherein the first corresponding relationship includes a corresponding relationship between each computing execution node in a first computing execution node and a task to be executed, wherein the first computing execution node includes one or more computing execution nodes in at least one computing execution node for executing the computing task, and the task to be executed is all or part of the tasks in the computing task; and receiving the first corresponding relationship from the computing management function network element.

[0036] Based on the above scheme, by obtaining the service quality requirements of the computing task and determining the correspondence between the computing execution node and the task to be executed based on the service quality requirements, the computing task can be scheduled to the appropriate computing execution node for execution, thereby improving the utilization of computing resources.

[0037] In combination with the third aspect, in certain implementations of the third aspect, the first correspondence is determined based on the service quality requirement and the computing capability of the at least one computing execution node, and the computing capability indicates first information, which includes at least one item of the following: performance information of performing tests based on at least one test data set, performance information of performing tests based on at least one model, and performance information of performing tests based on at least one test tool or test plan.

[0038] In combination with the third aspect, in certain implementations of the third aspect, the computing capability further indicates at least one moment, and the at least one moment corresponds to the first information.

[0039] In combination with the third aspect, in certain implementations of the third aspect, the performance information of the test performed based on at least one test data set includes at least one of the following: identification information of the at least one test data set, the number of training samples per unit time based on the at least one test data set as a training data set, the time required to complete the AI ​​model training based on the at least one test data set as a training data set, or the accuracy of training the AI ​​model based on the at least one test data set as a training data set.

[0040] In combination with the third aspect, in certain implementations of the third aspect, the performance information of the test performed based on at least one model includes at least one of the following: identification information of the at least one model, the time required to obtain the inference result based on the at least one model, or the number of times the at least one model is executed in unit time to obtain the inference result.

[0041] In combination with the third aspect, in certain implementations of the third aspect, the performance information of the test performed according to at least one test tool or test scheme includes at least one of the following: identification information of the at least one test tool or test scheme, the rate of encoding performed according to the at least one test tool or test scheme, the duration of encoding performed according to the at least one test tool or test scheme, the rate of decoding performed according to the at least one test tool or test scheme, or the duration of decoding performed according to the at least one test tool or test scheme.

[0042] In combination with the third aspect, in certain implementations of the third aspect, the quality of service requirement includes at least one of the following information about the computing task: task type, computing type, input data rate or data volume, output data rate or data volume, latency requirement, computing power requirement, and memory requirement.

[0043] In combination with the third aspect, in certain implementations of the third aspect, the quality of service requirement is carried in a computing service request message, and the computing service request message further includes association relationship information between the at least one subtask.

[0044] In combination with the third aspect, in certain implementations of the third aspect, the association relationship information includes at least one of the following: data flow direction, input or output data rate, input or output data throughput, and input and output dependency.

[0045] In combination with the third aspect, in certain implementations of the third aspect, the computing request node is a terminal device, and the computing service request message is carried in any of the following signaling: wireless resource control RRC signaling, first protocol layer signaling, computing resource control CRC signaling; wherein, the first protocol layer is located above the wireless resource control RRC protocol layer or in parallel with the RRC protocol layer.

[0046] In combination with the third aspect, in certain implementations of the third aspect, the first correspondence is carried in any one of the following messages: downlink control information DCI, media access control MAC control element CE, radio resource control RRC message, packet data convergence layer protocol PDCP control protocol data unit PDU, signaling of the first protocol layer, computing resource control CRC signaling; wherein, the first protocol layer is located above the radio resource control RRC protocol layer or in parallel with the RRC protocol layer.

[0047] In a fourth aspect, a communication device is provided, which includes a transceiver unit and a processing unit. The transceiver unit is used to obtain the computing capacity of each computing execution node in at least one computing execution node; the processing unit is used to determine a first correspondence and / or a second correspondence based on the computing capacity and the service quality requirements of the computing task, the first correspondence including the correspondence between each computing execution node in the first computing execution node and the task to be executed, the first computing execution node including one or more computing execution nodes in the at least one computing execution node for executing the computing task, the second correspondence including the correspondence between the second computing execution node and the task to be executed, the second computing execution node being any computing execution node in the first computing execution node, and the task to be executed being all or part of the tasks in the computing task; the transceiver unit is also used to send the first correspondence to the computing request node; and / or send the second correspondence to the second computing execution node.

[0048] In combination with the fourth aspect, in certain implementations of the fourth aspect, the computing capability indicates first information, which includes at least one item of the following: performance information of performing tests based on at least one test data set, performance information of performing tests based on at least one model, and performance information of performing tests based on at least one test tool or test plan.

[0049] In combination with the fourth aspect, in certain implementations of the fourth aspect, the computing capability further indicates at least one moment, and the at least one moment corresponds to the first information.

[0050] In combination with the fourth aspect, in certain implementations of the fourth aspect, the performance information of the test performed based on at least one test data set includes at least one of the following: identification information of the at least one test data set, the number of training samples per unit time based on the at least one test data set as a training data set, the time required to complete the AI ​​model training based on the at least one test data set as a training data set, or the accuracy of training the AI ​​model based on the at least one test data set as a training data set.

[0051] In combination with the fourth aspect, in certain implementations of the fourth aspect, the performance information of the test performed based on at least one model includes at least one of the following: identification information of the at least one model, the time required to obtain the inference result based on the at least one model, or the number of times the at least one model is executed in unit time to obtain the inference result.

[0052] In combination with the fourth aspect, in certain implementations of the fourth aspect, the performance information of the test performed according to at least one test tool or test scheme includes at least one of the following: identification information of the at least one test tool or test scheme, the rate of encoding performed according to the at least one test tool or test scheme, the duration of encoding performed according to the at least one test tool or test scheme, the rate of decoding performed according to the at least one test tool or test scheme, or the duration of decoding performed according to the at least one test tool or test scheme.

[0053] In combination with the fourth aspect, in certain implementations of the fourth aspect, the quality of service requirement includes at least one of the following information about the computing task: task type, computing type, input data rate or data volume, output data rate or data volume, latency requirement, computing power requirement, and memory requirement.

[0054] In combination with the fourth aspect, in certain implementations of the fourth aspect, the transceiver unit is further configured to receive a computing service request message from the computing request node, where the computing service request message includes the quality of service requirement.

[0055] In combination with the fourth aspect, in certain implementations of the fourth aspect, the computing task includes at least one subtask, and the computing service request message further includes association relationship information between the at least one subtask.

[0056] In combination with the fourth aspect, in certain implementations of the fourth aspect, the association relationship information includes at least one of the following: data flow direction, input or output data rate, input or output data throughput, and input and output dependency.

[0057] In combination with the fourth aspect, in certain implementations of the fourth aspect, the computing request node is a terminal device, and the computing service request message is carried in any of the following signaling: wireless resource control RRC signaling, first protocol layer signaling, computing resource control CRC signaling; wherein, the first protocol layer is located above the wireless resource control RRC protocol layer or in parallel with the RRC protocol layer.

[0058] In combination with the fourth aspect, in certain implementations of the fourth aspect, the first correspondence is carried in any one of the following messages: downlink control information DCI, media access control MAC control element CE, radio resource control RRC message, packet data convergence layer protocol PDCP control protocol data unit PDU, signaling of the first protocol layer, computing resource control CRC signaling; wherein, the first protocol layer is located above the radio resource control RRC protocol layer or in parallel with the RRC protocol layer.

[0059] In combination with the fourth aspect, in certain implementations of the fourth aspect, the computing execution node is an application server, and the second correspondence is carried in any one of the following messages: the tunnel protocol user plane part GTP-U header, the application programming interface API interface message of the network capability exposure function NEF.

[0060] In combination with the fourth aspect, in certain implementations of the fourth aspect, the communication device is applied to an access network device, an access network device distributed unit DU, or an access network device centralized unit CU.

[0061] In the fifth aspect, a communication device is provided, which includes a transceiver unit, which is used to send computing power to a computing management function network element, and the computing power is used to determine a second correspondence relationship, and the second correspondence relationship includes a correspondence between a computing execution node and a task to be executed, and the task to be executed is all or part of the computing task; the transceiver unit is also used to receive the second correspondence relationship from the computing management function network element.

[0062] In combination with the fifth aspect, in certain implementations of the fifth aspect, the computing capability indicates first information, which includes at least one item of the following: performance information of performing tests based on at least one test data set, performance information of performing tests based on at least one model, and performance information of performing tests based on at least one test tool or test plan.

[0063] In combination with the fifth aspect, in certain implementations of the fifth aspect, the computing capability further indicates at least one moment, and the at least one moment corresponds to the first information.

[0064] In combination with the fifth aspect, in certain implementations of the fifth aspect, the performance information of the test performed based on at least one test data set includes at least one of the following: identification information of the at least one test data set, the number of training samples per unit time based on the at least one test data set as a training data set, the time it takes to complete the AI ​​model training based on the at least one test data set as a training data set, or the accuracy of training the AI ​​model based on the at least one test data set as a training data set.

[0065] In combination with the fifth aspect, in certain implementations of the fifth aspect, the performance information of the test performed based on at least one model includes at least one of the following: identification information of the at least one model, the time required to obtain the inference result based on the at least one model, or the number of times the at least one model is executed in unit time to obtain the inference result.

[0066] In combination with the fifth aspect, in certain implementations of the fifth aspect, the performance information of the test performed according to at least one test tool or test scheme includes at least one of the following: identification information of the at least one test tool or test scheme, the rate of encoding performed according to the at least one test tool or test scheme, the duration of encoding performed according to the at least one test tool or test scheme, the rate of decoding performed according to the at least one test tool or test scheme, or the duration of decoding performed according to the at least one test tool or test scheme.

[0067] In combination with the fifth aspect, in certain implementations of the fifth aspect, the communication device is an application server, and the second correspondence is carried in any one of the following messages: the tunnel protocol user plane part GTP-U header, the application programming interface API interface message of the network capability exposure function NEF.

[0068] In the sixth aspect, a communication device is provided, which includes a transceiver unit, which is used to send the service quality requirements of the computing task to the computing management function network element, and the service quality requirements are used to determine a first corresponding relationship, and the first corresponding relationship includes the corresponding relationship between each computing execution node in the first computing execution node and the task to be executed, and the first computing execution node includes one or more computing execution nodes in at least one computing execution node for executing the computing task, and the task to be executed is all or part of the task in the computing task; the transceiver unit is also used to receive the first corresponding relationship from the computing management function network element.

[0069] In combination with the sixth aspect, in certain implementations of the sixth aspect, the first correspondence is determined based on the service quality requirement and the computing capability of the at least one computing execution node, and the computing capability indicates first information, which includes at least one item of the following: performance information of performing tests based on at least one test data set, performance information of performing tests based on at least one model, and performance information of performing tests based on at least one test tool or test plan.

[0070] In combination with the sixth aspect, in certain implementations of the sixth aspect, the computing capability further indicates at least one moment, and the at least one moment corresponds to the first information.

[0071] In combination with the sixth aspect, in certain implementations of the sixth aspect, the performance information of the test performed based on at least one test data set includes at least one of the following: identification information of the at least one test data set, the number of training samples per unit time based on the at least one test data set as a training data set, the time it takes to complete the AI ​​model training based on the at least one test data set as a training data set, or the accuracy of training the AI ​​model based on the at least one test data set as a training data set.

[0072] In combination with the sixth aspect, in certain implementations of the sixth aspect, the performance information of the test performed based on at least one model includes at least one of the following: identification information of the at least one model, the time required to obtain the inference result based on the at least one model, or the number of times the at least one model is executed in unit time to obtain the inference result.

[0073] In combination with the sixth aspect, in certain implementations of the sixth aspect, the performance information of the test performed according to at least one test tool or test scheme includes at least one of the following: identification information of the at least one test tool or test scheme, the rate of encoding performed according to the at least one test tool or test scheme, the duration of encoding performed according to the at least one test tool or test scheme, the rate of decoding performed according to the at least one test tool or test scheme, or the duration of decoding performed according to the at least one test tool or test scheme.

[0074] In combination with the sixth aspect, in certain implementations of the sixth aspect, the quality of service requirement includes at least one of the following information about the computing task: task type, computing type, input data rate or data volume, output data rate or data volume, latency requirement, computing power requirement, and memory requirement.

[0075] In combination with the sixth aspect, in certain implementations of the sixth aspect, the service quality requirement is carried in a computing service request message, and the computing service request message also includes association relationship information between the at least one subtask.

[0076] In combination with the sixth aspect, in certain implementations of the sixth aspect, the association relationship information includes at least one of the following: data flow direction, input or output data rate, input or output data throughput, and input and output dependency.

[0077] In combination with the sixth aspect, in certain implementations of the sixth aspect, the device is a terminal device, and the computing service request message is carried in any of the following signaling: wireless resource control RRC signaling, first protocol layer signaling, computing resource control CRC signaling; wherein, the first protocol layer is located above the wireless resource control RRC protocol layer or in parallel with the RRC protocol layer.

[0078] In combination with the sixth aspect, in certain implementations of the sixth aspect, the first correspondence is carried in any one of the following messages: downlink control information DCI, media access control MAC control element CE, radio resource control RRC message, packet data convergence layer protocol PDCP control protocol data unit PDU, signaling of the first protocol layer, computing resource control CRC signaling; wherein, the first protocol layer is located above the radio resource control RRC protocol layer or in parallel with the RRC protocol layer.

[0079] In a seventh aspect, a communication device is provided, comprising a processor coupled to a memory and configured to execute instructions in the memory to implement any of the aforementioned aspects of the first aspect, as well as the method of any possible implementation of any of the aspects of the first aspect. Optionally, the device further comprises a memory, which may be deployed separately from the processor or centrally. Optionally, the device further comprises a communication interface, the processor coupled to the communication interface.

[0080] In one implementation, the communication interface may be a transceiver, or an input / output interface.

[0081] In another implementation, the device is a computing and management function network element, or a chip configured in the computing and management function network element. It can also be a logic module or software that can implement all or part of the computing and management function network element functions. When the device is a chip, the communication interface can be an input / output interface, interface circuit, output circuit, input circuit, pin, or related circuit on the chip or chip system. The processor can also be embodied as a processing circuit or a logic circuit.

[0082] Optionally, the transceiver may be a transceiver circuit. Optionally, the input / output interface may be an input / output circuit.

[0083] In a specific implementation, the processor may be one or more chips, the input circuit may be an input pin, the output circuit may be an output pin, and the processing circuit may be a transistor, a gate circuit, a trigger, or various logic circuits. The input signal received by the input circuit may be, but is not limited to, received and input by a receiver, and the signal output by the output circuit may be, but is not limited to, output to and transmitted by a transmitter. The input circuit and the output circuit may be the same circuit, which functions as an input circuit and an output circuit at different times. The embodiments of the present application do not limit the specific implementation of the processor and various circuits.

[0084] In an eighth aspect, a communication device is provided, which includes a logic circuit and an input / output interface, wherein the logic circuit is used to couple with the input / output interface and transmit data through the input / output interface to execute any aspect of the above-mentioned second aspect and any possible implementation method of any aspect of the second aspect.

[0085] In one implementation, the device is a computational execution node, or a chip configured in a computational execution node. It may also be a logic module or software that implements all or part of the computational execution node's functionality. When the device is a chip, the communication interface may be an input / output interface, interface circuit, output circuit, input circuit, pin, or related circuit on the chip or chip system. The processor may also be embodied as a processing circuit or a logic circuit.

[0086] In a ninth aspect, a communication device is provided, which includes a logic circuit and an input / output interface, wherein the logic circuit is used to couple with the input / output interface and transmit data through the input / output interface to execute any aspect of the third aspect above and any possible implementation method of any aspect of the third aspect.

[0087] In one implementation, the device is a computation request node, or a chip configured in a computation request node. It may also be a logic module or software that implements all or part of the computation request node's functionality. When the device is a chip, the communication interface may be an input / output interface, interface circuit, output circuit, input circuit, pin, or related circuit on the chip or chip system. The processor may also be embodied as a processing circuit or a logic circuit.

[0088] In the tenth aspect, the present application provides a chip system, comprising: a processor, which is used to execute the computer program or instructions in the memory, so that the chip system implements the method in the above-mentioned first to third aspects and any possible implementation method of the first to third aspects.

[0089] In the eleventh aspect, the present application provides a communication system, including: a computing management function network element, a computing execution node and a computing request node, the computing management function network element is used to execute the method in the above-mentioned first aspect and any possible implementation of the first aspect; the computing execution node is used to execute the method in the above-mentioned second aspect and any possible implementation of the second aspect; the computing request node is used to execute the method in the above-mentioned third aspect and any possible implementation of the third aspect.

[0090] In the twelfth aspect, a computer-readable storage medium is provided, which stores a computer program (also referred to as code, or instructions). When the computer program is run on a computer, the computer executes any one of the above-mentioned aspects 1 to 3, and any possible implementation of the methods in the first to third aspects.

[0091] In the thirteenth aspect, a computer program product is provided, which includes: a computer program (also referred to as code, or instructions), which, when executed, enables a computer to execute any one of the above-mentioned first to third aspects, and any possible implementation method of the first to third aspects.

[0092] The beneficial effects brought about by the above-mentioned fourth to thirteenth aspects can be referred to the description of the beneficial effects in the first to third aspects, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0093] FIG1 is a schematic diagram of a network architecture applicable to an embodiment of the present application;

[0094] FIG2 is a schematic diagram of a computing power network architecture applicable to an embodiment of the present application;

[0095] FIG3 is a schematic flow chart of a task scheduling method provided in an embodiment of the present application;

[0096] FIG4 is a schematic structural diagram of a gNB applicable to an embodiment of the present application;

[0097] FIG5 is a schematic diagram of the structure of an O-RAN applicable to an embodiment of the present application;

[0098] FIG6 is a schematic structural diagram of a communication device provided in an embodiment of the present application;

[0099] FIG7 is a schematic structural diagram of another communication device provided in an embodiment of the present application;

[0100] FIG8 is a schematic structural diagram of another communication device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0101] The technical solutions in the embodiments of the present application will be described below with reference to the accompanying drawings.

[0102] The technical solutions of the embodiments of the present application can be applied to various communication systems, for example, long term evolution (LTE), fifth generation (5G), new radio (NR), Internet of Things (IoT), wireless fidelity (WiFi), wireless communications related to the 3rd Generation Partnership Project (3GPP), or other wireless communications that may appear in the future, such as the sixth generation (6G) communication system, etc., which is not limited in this application.

[0103] The technical solution provided in this application can also be applied to machine type communication (MTC), device-to-device (D2D) network, machine-to-machine (M2M) network, Internet of Things (IoT) network or other networks. Among them, IoT network can include, for example, Internet of Vehicles. Among them, the communication mode in the Internet of Vehicles system is collectively referred to as vehicle to other devices (vehicle to X, V2X, X can represent anything), for example, the V2X can include: vehicle to vehicle (V2V) communication, vehicle to infrastructure (V2I) communication, vehicle to pedestrian (V2P) communication or vehicle to network (V2N) communication, etc.

[0104] Figure 1 is a schematic diagram of a network architecture applicable to an embodiment of the present application. As shown in Figure 1, the network architecture takes the 5G system (5th generation system, 5GS) as an example.

[0105] The network architecture can be divided into two parts: access network and core network. The network architecture can include but is not limited to: unified data management (UDM), network exposure function (NEF), network storage function (NF repository function, NRF), policy control function (PCF), application function (AF), access and mobility management function (AMF), session management function (SMF), user equipment (UE), wireless access network equipment, user plane function (UPF), data network (DN). Among them, DN can be the Internet; UDM, NEF, NRF, PCF, AF, AMF, SMF, UPF are network elements in the core network. Since Figure 1 takes the 5G system as an example, the core network can be called 5G core network (5GC or 5GCN).

[0106] The following is a brief introduction to each network element shown in FIG1 .

[0107] 1. User equipment (UE): can be called terminal equipment, access terminal, user unit, user station, mobile station, mobile station, remote station, remote terminal, mobile device, user terminal, terminal, wireless communication device, user agent or user device.

[0108] The terminal device can be a device that provides voice / data to users, for example, a handheld device or vehicle-mounted device with a wireless connection function. At present, some examples of terminals include: mobile phones, tablet computers, laptop computers, PDAs, mobile internet devices (MIDs), virtual reality (VR) devices, augmented reality (AR) devices, wireless terminals in industrial control, wireless terminals in self-driving, wireless terminals in remote medical surgery, wireless terminals in smart grids, wireless terminals in transportation safety, wireless terminals in smart cities, wireless terminals in smart homes, cellular phones, cordless phones, session initiation protocol (SIP) phones, wireless local loop (WLL) stations, personal digital assistants (PDAs), handheld devices with wireless communication capabilities, computing devices or other processing devices connected to wireless modems, terminal devices in 5G networks or terminal devices in future evolved public land mobile networks (PLMNs), etc. The embodiments of the present application are not limited to this.

[0109] As an example and not a limitation, in the embodiment of the present application, the terminal device may also be a wearable device. Wearable devices may also be called wearable smart devices, which are a general term for wearable devices that are intelligently designed and developed using wearable technology for daily wear, such as glasses, gloves, watches, clothing, and shoes. A wearable device is a portable device that is worn directly on the body or integrated into the user's clothes or accessories. Wearable devices are not only hardware devices, but also achieve powerful functions through software support, data interaction, and cloud interaction. Broadly speaking, wearable smart devices include those that are fully functional, large in size, and can achieve complete or partial functions without relying on smartphones, such as smart watches or smart glasses, as well as those that only focus on a certain type of application function and need to be used in conjunction with other devices such as smartphones, such as various smart bracelets and smart jewelry for vital sign monitoring.

[0110] In addition, in the embodiments of the present application, the terminal device can also be a terminal device in an IoT system. IoT is an important part of the future development of information technology. Its main technical feature is to connect objects to the network through communication technology, thereby realizing an intelligent network that interconnects people and machines and things.

[0111] It should be noted that the terminal device and the access network device can communicate with each other using some air interface technology (such as new radio (NR) or LTE technology). The terminal devices can also communicate with each other using some air interface technology (such as NR or LTE technology).

[0112] In the embodiments of the present application, the device for implementing the function of the terminal device can be the terminal device, or it can be a device that can support the terminal device to implement the function, such as a chip system or chip, which can be installed in the terminal device. In the embodiments of the present application, the chip system can be composed of a chip, or it can include a chip and other discrete devices.

[0113] 2. (Radio) Access Network (R)AN) equipment: This equipment provides access to the communications network for authorized users in a specific area. Specifically, it can include wireless network equipment in 3rd Generation Partnership Project (3GPP) networks and access points in non-3GPP networks. For ease of description, the term "RAN" is used below.

[0114] RAN equipment can adopt different wireless access technologies. There are currently two types of wireless access technologies: 3GPP access technology (for example, the wireless access technology used in the third generation (3rd generation, 3G), fourth generation (4G) or 5G systems) and non-3GPP (non-3GPP) access technology. 3GPP access technology refers to access technology that complies with 3GPP standard specifications. For example, the access network equipment in the 5G system is called the next generation Node Base station (gNB) or RAN equipment. Non-3GPP access technologies may include air interface technologies represented by access points (APs) in wireless fidelity (WiFi), worldwide interoperability for microwave access (WiMAX), code division multiple access (CDMA), etc. AN equipment can allow terminal devices and the 3GPP core network to interconnect and communicate using non-3GPP technologies.

[0115] RAN equipment is responsible for radio resource management, quality of service (QoS) management, data compression and encryption, etc. on the air interface side. AN equipment provides access services to terminal devices and forwards control signals and user data between terminal devices and the core network.

[0116] RAN equipment may include, but is not limited to, macro base stations, micro base stations (also known as small stations), radio network controllers (RNCs), node Bs (NBs), base station controllers (BSCs), base transceiver stations (BTSs), home base stations (e.g., home evolved NodeBs, or home NodeBs, HNBs), baseband units (BBUs), APs in WiFi systems, wireless relay nodes, wireless backhaul nodes, transmission points (TPs) or transmission and reception points (TRPs), etc. It may also be a gNB or transmission point (TRP or TP) in a 5G (e.g., NR) system, one or a group of antenna panels (including multiple antenna panels) of a base station in a 5G system, or a network node constituting a gNB or transmission point, such as a distributed unit (DU), or a base station under different distributed radio access networks (D-RANs) in a large area, or a cloud radio access network (C-RAN). The RAN device may be a base station under a D-RAN (C-RAN) architecture, or a mobile edge computing (MEC) platform that is deployed to a base station under both D-RAN and C-RAN access network architectures, or a base station in a 6G communication system. The embodiments of the present application do not limit the specific technology and device form used by the RAN device.

[0117] In different systems, CU (or CU-CP and CU-UP), DU or RU may also have different names, but those skilled in the art can understand their meanings. For example, in an open radio access network (O-RAN) system, CU may also be referred to as an open-central unit (O-CU); DU may also be referred to as an open-distributed unit (O-DU); CU-CP may also be referred to as O-CU-CP, CU-UP may also be referred to as O-CU-UP, and RU may also be referred to as O-RU. For the sake of convenience of description, this application uses CU, CU-CP, CU-UP, DU and RU as examples for description. Any of the CU (or CU-CP, CU-UP), DU and RU in this application may be implemented by a software module, a hardware module, or a combination of a software module and a hardware module.

[0118] 3. User plane function (UPF): This element can be used for packet routing and forwarding, or for quality of service (QoS) processing of user plane data. User data can be connected to the data network (DN) through this element. In the embodiments of the present application, this element can be used to implement the functions of the user plane element.

[0119] 4. Access and Mobility Management Function (AMF): This is primarily used for mobility management and access management. It can be used to implement other functions of the Mobility Management Entity (MME) in addition to session management, such as lawful interception or access authorization (or authentication). In the embodiment of the present application, it can be used to implement the functions of the access and mobility management function element.

[0120] 5. Session Management Function (SMF): This is primarily used for session management, IP address allocation and management of terminal devices, selection and management of user plane functions, policy control, or termination points of charging function interfaces, and downlink data notification. In the embodiments of the present application, it can be used to implement the functions of the session management network element.

[0121] 6. Policy control function (PCF): A unified policy framework used to guide network behavior and provide policy rule information to control plane functional network elements (such as AMF, SMF network elements, etc.).

[0122] 7. Network exposure function (NEF): used to securely expose service and capability information (such as terminal location and network congestion status) provided by 3GPP network function elements to the outside world.

[0123] 8. Network repository function (NRF): used to store descriptions of network function entities and the services they provide, and to support service discovery, network element entity discovery, and other functions;

[0124] 9. Authentication server function (AUSF) network element: mainly used for user authentication, etc.

[0125] 10. Unified data management (UDM): used for unified data management, 5G user data management, processing user identification, access authentication, registration, or mobility management.

[0126] 11. Digital network (DN): A network used to transmit data. Examples include carrier networks, the Internet, and third-party business networks. For example, a DN may include applications (Apps), mobile edge computing platforms (MEPs), and so on.

[0127] 12. Application Function (AF): used for data routing affected by applications, accessing network open functional elements, or interacting with the policy framework for policy control, etc. In this application, AF can also be understood as an application server.

[0128] 13. MEC: MEC can be seen as a cloud server running at the edge of the mobile network and performing specific tasks. The European Telecommunications Standards Institute (ETSI) defines MEC as a platform that provides users with IT architecture and cloud computing capabilities in the RAN network close to mobile users. The actual deployment location of MEC in the wireless network generally corresponds to the edge UPF network element of the 5G core network. Among them, the edge UPF implements local offloading and diversion of services based on the N6 interface between the local data network (local-DN), thereby realizing local processing of services and achieving an acceleration effect. MEC can also be called multi-access edge computing.

[0129] In the network architecture shown in Figure 1, network elements can communicate with each other through the interfaces shown in the figure. As shown in the figure, the N1 interface is the reference point between the terminal device and the AMF; the N2 interface is the reference point between the RAN and the AMF, used for sending non-access stratum (NAS) messages, etc.; the N3 interface is the reference point between the RAN and the UPF, used for transmitting user plane data, etc.; the N4 interface is the reference point between the SMF and the UPF, used for transmitting information such as tunnel identification information of the N3 connection, data buffer indication information, and downlink data notification messages; the N5 interface is the reference point between the PCF and the AF; the N6 interface is the reference point between the UPF and the DN, used for transmitting user plane data, etc.; the N7 interface is the reference point between the SMF and the PCF; and the N9 interface is the interface between UPFs, such as the interface between the visited-policy control function (V-PCF) and the home-policy control function (H-PCF), or the interface between a UPF connected to the DN and a UPF connected to the RAN, used to transmit user plane data between UPFs. The N33 interface is the reference point between the NEF and AF. For the sake of brevity, the relationship between other interfaces and various network elements is not described in detail here.

[0130] It should be understood that the network architecture shown in Figure 1 above is only an exemplary illustration, and the network architecture applicable to the embodiments of the present application is not limited to this. Any network architecture that can realize the functions of the above-mentioned network elements is applicable to the embodiments of the present application.

[0131] It should also be understood that the functions or network elements such as AMF, SMF, UPF, PCF, UDM shown in Figure 1 can be understood as network elements for implementing different functions. For example, they can be combined into network slices as needed. These network elements can be independent devices, or they can be integrated into the same device to implement different functions, or they can be network elements in hardware devices, or they can be software functions running on dedicated hardware, or they can be virtualized functions instantiated on a platform (for example, a cloud platform). This application does not limit the specific form of the above network elements.

[0132] It should also be understood that the above naming is defined only to facilitate the distinction between different functions and should not constitute any limitation to this application. This application does not exclude the possibility of adopting other naming in 6G networks and other future networks. For example, in a 6G network, some or all of the above network elements may continue to use the terminology used in 5G, or may adopt other names.

[0133] To facilitate understanding of the embodiments of the present application, a brief explanation of the terms involved in the present application is first given.

[0134] 1. Computing power

[0135] Hashrate is computing power. More specifically, it's the ability to process information data to achieve a desired output. Hashrate can be measured in different units in different application scenarios, such as hash operations per second (H / S) for Bitcoin and floating-point operations per second (FLOP / S) for artificial intelligence (AI) and graphics processing. The computing power demanded by the intelligent society is primarily floating-point computing power. The Bitcoin network must perform intensive mathematical and cryptographic operations for security purposes. For example, when the network reaches a hash rate of 10 Th / s, it means it can perform 10 trillion calculations per second.

[0136] 2. Computing Network (CN)

[0137] The computing power network can perceive ubiquitous computing and services through the network, and dispatch computing tasks to the optimal computing execution nodes for processing, thereby improving the utilization of network resources and computing resources, while providing users with low-latency, highly reliable business services and improving user experience.

[0138] The computing network is a new type of information infrastructure that allocates and flexibly schedules computing, storage, and network resources across the cloud, network, and edge based on business needs. This network infrastructure senses, schedules, and orchestrates computing resources, providing new information and communications technology (ICT) converged services across networking, computing, and storage at the network layer. It represents a new generation of network-based technology architecture and business operations system.

[0139] Figure 2 is a schematic diagram of the architecture of a computing network applicable to an embodiment of the present application. A computing management function (CMF) network element is introduced into the architecture to perform computing power, computing services, and perception of network status, as well as scheduling of computing tasks. CMF can be deployed independently or in conjunction with other network functions. For example, as shown in Figure 2, CMF and access network equipment (for example, base stations) collaborate to support convergence scheduling (CS) of computing power and communication; alternatively, CMF can be deployed independently in the core network and connected to the access network equipment through the NG_AP interface. When CMF is deployed independently in the core network, the information transmitted between CMF and RAN can be transferred through AMF, or CMF can be directly connected to RAN; alternatively, CMF can also be co-located with access network equipment or core network elements (for example, AMF, SMF).

[0140] In this application, the CMF network element has functions such as computing resource perception, computing task perception, computing task scheduling, and network status perception, and can schedule computing power for access network equipment, terminal equipment, and mobile edge computing (MEC) platforms.

[0141] Services in future communication systems will become increasingly immersive and intelligent. Therefore, computing networks may be required to support the scheduling of artificial intelligence (AI) resources and the processing of extremely large amounts of data, reducing terminal computing workloads and energy consumption. This requires a shift in network architecture from a "connection-centric" to a "task-centric" approach. A "connection-centric" network architecture aims to provide the necessary communication channels between terminals and between terminals and application servers, along with corresponding lifecycle management mechanisms and quality of service (QoS) guarantees. A "task-centric" network architecture, on the other hand, coordinates and schedules the computing power, connections, algorithms, and data resources of multiple nodes to achieve a specific goal. Within this "task-centric" computing network architecture, the proper scheduling of computing tasks is crucial.

[0142] In view of this, the present application provides a task scheduling method and communication device to schedule computing tasks to appropriate computing execution nodes for execution, thereby improving resource utilization.

[0143] In order to facilitate understanding of the embodiments of the present application, the following explanations are made.

[0144] First, in the various embodiments of the present application, unless otherwise specified or provided for, the terms and / or descriptions between the different embodiments are consistent and can be referenced by each other. The technical features in the different embodiments can be combined to form new embodiments according to their inherent logical relationships.

[0145] Second, the first, second, and various numerical numbers (e.g., "#1," "#2," etc.) shown in this application are merely for convenience of description and are used to distinguish between objects. They are not intended to limit the scope of the embodiments of this application. For example, they are used to distinguish between different terminal devices, etc. They are not intended to describe a specific order or precedence. It should be understood that the objects described in this manner can be interchanged where appropriate to describe solutions other than the embodiments of this application.

[0146] Third, in the embodiments of the present application, "of", "corresponding, relevant", "corresponding" and "associate" can sometimes be used interchangeably. It should be pointed out that when the distinction between them is not emphasized, the meanings they intend to express are consistent.

[0147] Figure 3 is a schematic flow chart of a task scheduling method provided in an embodiment of the present application. The method may include at least the following steps.

[0148] S310: The computing management function network element obtains the computing capability of each computing execution node in at least one computing execution node.

[0149] The computing management function network element may also be referred to as a task management function network element or a management function network element, and this application does not limit this. The functions of the computing management function network element in this application include the functions of the CMF network element described in FIG2 .

[0150] The at least one computing execution node can be understood as at least one computing execution node served (managed) by the computing management function network element.

[0151] The computing capability may indicate the first information, or in other words, the computing capability may be represented by the first information.

[0152] The first information may include at least one of the following:

[0153] Performance information of tests performed according to at least one test data set, performance information of tests performed according to at least one model, and performance information of tests performed according to at least one test tool or test plan.

[0154] Specifically, the performance information of executing a test based on at least one test data set may include at least one of the following:

[0155] The identification information of the at least one test data set, the number of training samples per unit time based on the at least one test data set as a training data set (also known as the learning throughput), the time it takes to complete the AI ​​model training based on the at least one test data set as a training data set, or the accuracy of training the AI ​​model based on the at least one test data set as a training data set.

[0156] Taking the test dataset as a test dataset for AI business (for example, a visual test dataset or a natural language processing test dataset), the test performance information includes the learning throughput based on the at least one test dataset as a training dataset. The unit of measurement of the test performance information can be the number of images processed per second (images / s) (for example, using a visual test dataset), or the number of sentences processed per second (sentences / s) (for example, using a natural language processing test dataset).

[0157] Among them, the test data set for AI business can refer to the description in the "Information Technology Artificial Intelligence Server System Performance Test Specification" (protocol specification number T / CESA 1169-2021). For example, the test data set includes the Imagenet2012 data set for image recognition, the cn-wiki or en-wiki data set for natural language processing, etc.

[0158] It should be understood that the above-mentioned AI business-oriented data set is only an example of a test data set, and this application does not limit the type of test data set. For example, the test data set can also be a test data set for general scenarios. The test data set for general scenarios can refer to the description in the "Technical Requirements for General Computing CPU Performance Testing and Evaluation". For example, standard test data sets for general scenarios include online transaction processing (OLTP), online analytical processing (OLAP), Spark (distributed parallel data processing framework) database, etc.

[0159] For example, performance information for executing a test based on at least one test dataset can be shown in Table 1. The at least one test dataset includes the MNIST dataset, the CIFAR-10 dataset, and the Imagenet2012 dataset; the performance information includes the throughput of executing the test on the MNIST dataset, the CIFAR-10 dataset, and the Imagenet2012 dataset. Taking the MNIST dataset as an example, the performance information includes the identifier of the MNIST dataset and the throughput of executing the test on the MNIST dataset: 100 images / s, meaning that the computing execution node can train on 100 samples from the MNIST dataset per second.

[0160] Table 1

[0161] Alternatively, the performance information of the test performed based on at least one test dataset may further include relative performance information. Specifically, the relative performance information indicates a ratio of the performance information of the test performed based on the test dataset to the performance information of the test performed based on a reference test dataset. The reference test dataset may be a designated test dataset. For example, the reference test dataset may be pre-agreed or executed by a computing management function network element.

[0162] Alternatively, the performance information may be comprehensive performance information represented by a first value v. The first value v is a weighted calculation result of the ratio of the performance information of the computing execution node when performing the test based on at least one test data set to the baseline performance of the benchmark system when performing the test based on the at least one test data set. For example,

[0163] Among them, P si represents the performance information of the benchmark system based on the test data set i; P i represents the performance information of the execution node according to the test data set i; w i represents the weight of the performance information of the test performed based on test dataset i; α is an adjustment coefficient, α is a positive integer, for example, α can be 100. Test dataset i represents the i-th test dataset in the at least one test dataset, i = 1, 2, ..., N, where N is the number of the at least one test dataset.

[0164] The test dataset in this application can also be called a standard dataset. That is, for multiple computing execution nodes, if a specific test dataset is specified, the data samples included in the test dataset are consistent across the multiple computing execution nodes. If the data samples in the standard dataset change, the data samples in all computing execution nodes using the standard dataset must be updated.

[0165] Performance information for executing tests according to at least one model includes at least one of the following:

[0166] The identification information of the at least one model, the time length for obtaining the inference result of the at least one model (inference calculation delay), the number of times the model is executed per unit time to obtain the inference result, or the frequency of executing the model to obtain the inference result.

[0167] Taking the model as an AI business-oriented model, and the performance information including the duration of obtaining an inference result according to the at least one model as an example, the first information may be as shown in Table 2. The at least one model includes an AlexNet model, a VGG-16 model, and a ResNet model, and the performance information of the test performed according to the at least one model includes the duration of obtaining an inference result according to the AlexNet model, the VGG-16 model, and the ResNet model, for example, 100ms, 50ms, and 30ms, respectively.

[0168] Among them, AI business-oriented models can refer to the description in the "Information Technology Artificial Intelligence Server System Performance Test Specification" (Specification No.: T / CESA 1169-2021). Exemplarily, AI business-oriented models can also include the resnet101_v1 model for image recognition, the resnet50_v1.5 model, etc.

[0169] Table 2

[0170] It should be understood that the above-mentioned AI business-oriented model is only an example of a test data set, and this application does not limit the type of model.

[0171] Similarly, the performance information of the test performed according to at least one model may also be relative performance information. For details, reference may be made to the relative performance information of the test performed according to at least one test data set.

[0172] Performance information of tests performed according to at least one test tool or test plan includes at least one of the following:

[0173] Identification information of the at least one test tool or test scheme, the rate at which encoding is performed by the at least one test tool or test scheme, the duration for which encoding is performed by the at least one test tool or test scheme, the rate at which decoding is performed by the at least one test tool or test scheme, or the duration for which decoding is performed by the at least one test tool or test scheme.

[0174] For example, the performance information of performing tests according to at least one test tool or test scheme includes performance information of performing encoding or decoding according to at least one test tool or test scheme, and the at least one test tool or test scheme can be a gCC codec, an x264 codec scheme, or a gzip compression scheme; the performance information includes the rate or duration of performing encoding or decoding using the at least one test tool or test scheme.

[0175] Optionally, the computing capability may further indicate at least one moment, the at least one moment corresponding to the first information. In other words, the computing capability is represented by the first information corresponding to at least one moment. The at least one moment includes a moment before the current moment.

[0176] The first information corresponding to the at least one moment includes at least one of the following: performance information of a test performed based on at least one test data set at at least one moment, performance information of a test performed based on at least one model at at least one moment, and performance information of a test performed based on at least one test tool or test plan at at least one moment.

[0177] In other words, when the computing management function network element obtains the current computing capacity of each computing execution node, each computing execution node can return the first information corresponding to at least one moment before the current moment, or in other words, each computing execution node returns historical performance information.

[0178] When the computing capability is represented by first information corresponding to at least one moment, the method may further include: the computing management function network element determining first information at the current moment based on the first information corresponding to the at least one moment. For example, the computing management function network element fits the first information at the current moment using the first information corresponding to the at least one moment as input to a model.

[0179] Specifically, the computing management function network element may obtain the computing capability of each computing execution node in the manner shown in the following example.

[0180] In Example 1, the computing management function network element obtains the computing capacity of each computing execution node during the computing capacity registration process of the computing execution node. For example, the computing management function network element receives computing capacity registration information from the computing node, and the computing capacity registration information may include the computing capacity of each computing execution node.

[0181] In Example 2, the computing management function network element queries the computing capacity of each computing execution node. For example, the computing management function network element sends a computing capacity query message to each computing execution node; in response to the computing capacity query message, the computing execution node returns the computing capacity of the computing execution node to the computing management function network element.

[0182] Optionally, before the computing management function network element obtains the computing capacity of each computing execution node, the computing management function network element may also send configuration information to the computing execution node. The configuration information may include at least one of the following information:

[0183] Test data set and / or identification information of the test data set, model and / or identification information of the model, test tool (scheme) and / or identification information of the test tool (scheme).

[0184] That is, the computing management function network element configures a test data set, a model, and at least one of a test tool or a test solution for testing the performance information of the computing execution node; the computing execution node executes the test according to the configuration information and reports the computing capacity.

[0185] For example, the configuration information includes three test data sets as shown in Table 1, or the configuration information includes identification information of the three test data sets; the computing execution node reports the test performance information as shown in Table 1 according to the configuration information.

[0186] S320: The computing management function network element determines a first corresponding relationship and / or a second corresponding relationship according to the computing capability of each computing execution node and the quality of service (QoS) requirement of the computing task.

[0187] The computing task may be a computing task to be executed in the computing request node. The computing task may also be referred to as a computing service, an AI reasoning task, an AI training task, or an AI service, etc. This application does not limit the name of the computing task.

[0188] The QoS requirement of the computing task may come from the computing requesting node or be determined by configuration information.

[0189] For example, before S310, the computing management function network element receives a computing service request message from a computing request node. The computing service request message may carry the QoS requirement of the computing task.

[0190] Alternatively, the computing management function network element may be configured to include at least one computing task and the QoS requirements for each of the at least one computing task; the computing management function network element may receive indication information indicating the type of computing task from the computing request service, and determine the QoS requirements for the computing task based on the indication information and the configuration information.

[0191] For example, the quality of service requirement of a computing task may include at least one of the following:

[0192] The computing task includes the task type or computing type, input data rate or data volume, output data rate or data volume, latency requirements, computing power requirements, and memory requirements.

[0193] Among them, task types may include video rendering, AI training and / or reasoning, image processing, compression tasks, perception processing, high-performance computing, big data offline analysis, etc.

[0194] The computing type can also be referred to as the processor type. For example, the processor type can be a central processing unit (CPU), a graphics processing unit (GPU), a neural-network processing unit (NPU), a tensor processing unit (TPU), etc.

[0195] The input or output data rate or data volume can represent the communication service quality (QoS) requirements of the computing task. The communication QoS requirements can also include communication QoS parameters defined by 5G NR, such as the 5G QoS identifier (5QI). For example, the 5QI can indicate the uplink / downlink packet delay budget (PDB), guaranteed bit rate (GBR), and maximum burst data volume (MDBV).

[0196] The computing power requirement can refer to the computing power required to perform the computing task, or in other words, the computing power requirement represents the required processor computing power. For example, the number of operations per second (OPS) that the processor needs to perform or the number of floating-point operations per second (FLOPS). Optionally, the computing power requirement can also include information such as the model and main frequency of the hardware required to perform the computing task, for example, the main frequency required to reach hertz (Hz) or GHz.

[0197] Memory requirements include the storage size and input / output (I / O) speed required to perform the computing task. The I / O speed is measured in read / write operations per second (IOPS), and the storage size represents the number of bytes of memory required.

[0198] Exemplarily, the first correspondence includes a correspondence between a first computing execution node and a task to be executed, where the first computing execution node includes a computing execution node among the at least one computing execution node for executing the computing task, and the first computing execution node may be one or more computing execution nodes. The second correspondence includes a correspondence between a second computing execution node and a task to be executed, where the second computing execution node is any computing execution node among the first computing execution nodes. The task to be executed is all or part of the computing task.

[0199] That is, the computing management function network element determines the computing execution node (i.e., the first computing execution node) in the at least one computing execution node used to execute the computing task, and the tasks to be executed corresponding to each computing execution node in the first computing execution node (the second computing execution node) based on the computing capacity of each computing execution node in the at least one computing execution node and the QoS requirements of the computing task.

[0200] Optionally, the computing task may be decomposed into at least one subtask. In other words, the computing task may include at least one subtask. In this case, the QoS requirement of the computing task may be the QoS requirement of each subtask.

[0201] Furthermore, the computing management function network element determines the first correspondence and / or the second correspondence based on the computing capability and the quality of service QoS requirements of the computing task, which may include: the computing management function network element determines the first correspondence and / or the second correspondence based on the computing capability, the quality of service QoS requirements of each subtask and the association relationship between the subtasks.

[0202] The association relationship between subtasks may include at least one of the following:

[0203] Data flow direction, input or output data rate, input or output data throughput, input and output dependencies.

[0204] Taking a computing task including two subtasks as an example, the data flow direction can be data flowing from one subtask to another; the input-output dependency relationship can be that the output of one of the two subtasks serves as the input of the other subtask. In other words, the association relationship between the subtasks can include the association relationship between each subtask in the computing task and the previous subtask and / or the subsequent subtask, the input data of each subtask in the computing task can be the output data of the previous subtask corresponding to each subtask, and the output data of each subtask can be the input data of the subsequent subtask corresponding to each subtask. The input or output data rate can be the input and / or output data rate of each subtask in the computing task, or the output data rate of the previous subtask corresponding to each subtask in the computing task and / or the input data rate of the subsequent subtask corresponding to each subtask in the computing task; the input or output data throughput can be the input and / or output data throughput of each subtask in the computing task, or the output data throughput of the previous subtask corresponding to each subtask in the computing task and / or the input data throughput of the subsequent subtask corresponding to each subtask in the computing task.

[0205] The following uses specific subtasks, QoS requirements of subtasks, and computing capabilities of computing execution nodes as examples to illustrate how the computing management function network element determines the first correspondence and / or second correspondence based on the computing capabilities and QoS requirements of the computing tasks.

[0206] Exemplarily, the computing task includes subtask #1 and subtask #2, and the QoS requirements of subtask #1 and subtask #2 include the task type, latency requirement, and computing power requirement of the subtasks. The task type of subtask #1 is image detection, the latency requirement of subtask #1 is 10ms, and the computing power requirement is 0.1 tera floating-point operations per second (TFLOPS); the task type of subtask #2 is image recognition, the latency requirement is 10ms, the computing power is 1TFLOPS, and the output data of subtask #1 is the input data of subtask #2; the at least one computing execution node includes computing execution node x and computing execution node y, wherein the GPU of computing execution node x uses model #1 (at least one example of a model) to obtain an inference result in 10ms and a computing power of 2TFLOPS; the GPU of computing execution node y uses model #1 to obtain an inference result in 5ms and a computing power of 4TFLOPS.

[0207] At this time, due to the input-output dependency between the subtasks, the computing management function network element can select one of the computing execution nodes to execute subtask #1 and subtask #2. Furthermore, the computing management function network element can select a computing execution node (for example, computing execution node x) whose time to obtain the inference result in the computing execution node meets the latency requirements of the two subtasks and whose computing power is closer to the computing power requirements of the two subtasks to execute subtask #1 and subtask #2, thereby avoiding waste of computing power resources. That is, the first correspondence is the same as the second correspondence, and the first correspondence and the second correspondence include the correspondence between the computing execution node x and the tasks (subtask #1 and subtask #2) to be executed by the computing execution node x.

[0208] Alternatively, the computing management function network element may select a computing execution node (e.g., computing execution node y) whose computing power meets the computing power requirements of the two subtasks and takes less time to obtain the inference results to execute subtask #1 and subtask #2, thereby reducing the execution delay of the computing tasks. That is, the first correspondence and the second correspondence are the same, and the first correspondence and the second correspondence include the correspondence between computing execution node y and the tasks (subtask #1 and subtask #2) that computing execution node y needs to execute.

[0209] In addition, when the computing management function network element selects one of computing execution node x and computing execution node y to execute subtask #1 and subtask #2, the computing management function network element can also instruct the other computing execution node to enter low power consumption mode to save energy.

[0210] S330: The computing management function network element sends the first corresponding relationship to the computing requesting node.

[0211] That is, the computing management function network element sends information about the computing execution node that executes the computing task and the task that needs to be executed corresponding to the computing execution node to the computing request node.

[0212] For example, the computing task includes subtask #1 and subtask #2. Subtask #1 is executed by computing execution node x; subtask #2 is executed by computing execution node y. The first correspondence may include a correspondence between computing execution node x and subtask #1, and a correspondence between computing execution node y and subtask #2.

[0213] In the case where there is an association relationship between the subtasks, the first correspondence relationship may further include location information of the computing execution nodes where other computing tasks that have an association relationship with the computing task executed by the computing execution node are located.

[0214] For example, the computing task includes subtask #1 and subtask #2, and the output of subtask #1 can be used as the input of subtask #2. Subtask #1 is executed by computing execution node x; subtask #2 is executed by computing execution node y. The first correspondence may include the correspondence between computing execution node x and subtask #1, as well as the location information (e.g., IP address information) of computing execution node y; the first correspondence also includes the correspondence between computing execution node y and subtask #2. Optionally, the first correspondence may also include the QoS requirements of the subtasks. The first correspondence may be as shown in Table 3.

[0215] Table 3

[0216] Optionally, the method further includes:

[0217] S340: The computing management function network element sends the second corresponding relationship to the second computing execution node.

[0218] That is, the computing management function network element sends the task to be executed corresponding to each computing execution node that executes the computing task.

[0219] For example, the computing task includes subtask #1 and subtask #2. Subtask #1 is executed by computing execution node x; subtask #2 is executed by computing execution node y. The second correspondence is correspondence #1 between computing execution node x and subtask #1, or the second correspondence is correspondence #2 between computing execution node y and subtask #2.

[0220] Similarly, in the case where there is an association relationship between subtasks, the second correspondence relationship may also include the location information of the computing execution node that has an association relationship with the computing execution node that executes the computing task, or the location information of the computing nodes where other computing tasks are located that have an association relationship with the computing task executed by the second computing execution node.

[0221] For example, the computing task includes subtask #1 and subtask #2, and the output of subtask #1 can be used as the input of subtask #2. Subtask #1 is executed by computing execution node x, and subtask #2 is executed by computing execution node y. The second correspondence includes the correspondence between computing execution node x and subtask #1, as well as the location information (e.g., IP address information) of computing execution node y.

[0222] Optionally, the method further includes: the computation requesting node sending data of the computation task to the first computation executing node. Correspondingly, the first computation executing node receives the data from the computation requesting node.

[0223] The data of the computing task can be understood as the data required to perform the computing task, or the input data of the computing task. For example, when the computing task is image detection or recognition, the data of the computing task can include the image to be detected or recognized.

[0224] Optionally, the first computing execution node sends the processing result of the computing task to the computing request node. Correspondingly, the computing request node receives the processing result of the computing task from the first computing execution node.

[0225] The processing result of a computing task can be understood as the result obtained by performing a calculation on the data of the computing task. For example, when the computing task is image detection or recognition, the processing result of the computing task can include the detection or recognition result obtained by performing image detection or recognition on the image to be detected or recognized.

[0226] In the steps described above, when the computing request node is a node of different types and the computing management function network element is deployed at different locations, the message transmitted between the computing request node and the computing management function network element (for example, computing service request message, first correspondence) can be carried on different signaling.

[0227] Exemplarily, when the computing requesting node is a terminal device and the CMF is deployed on a RAN node (e.g., a gNB in ​​an NR system), the computing service request message may be carried in radio resource control (RRC) signaling, signaling of a first protocol layer, or computing resource control CRC signaling; wherein the first protocol layer is located above the radio resource control RRC protocol layer or in parallel with the RRC protocol layer. The first correspondence may be carried in downlink control information (DCI) or a media access control (MAC) control element (CE), an RRC message, a packet data convergence protocol (PDCP) control protocol data unit (PDU), signaling of the first protocol layer, or computing resource control CRC signaling.

[0228] Optionally, if the CMF is integrated into the gNB-CU, information transmitted between the CMF and the computation requesting node may be transmitted via the F1 interface between the gNB-CU and the gNB-DU. Specifically, the computation requesting node transmits a computation service request message to the gNB-DU; the gNB-DU transmits the computation service request message to the gNB-CU via the F1 interface. The gNB-CU transmits the first correspondence to the gNB-DU via the F1 interface, and the gNB-DU transmits the first correspondence to the computation requesting node via the aforementioned DCI or MAC CE signaling.

[0229] If the CMF is integrated into the gNB-DU, information transmitted between the CMF and the computation requesting node may be transmitted via the F1 interface between the gNB-CU and the gNB-DU. Specifically, the computation requesting node transmits a computation service request message to the gNB-CU via the aforementioned RRC signaling, Layer 1 protocol signaling, or CRC signaling. The gNB-CU transmits the computation service request message to the gNB-DU via the F1 interface. The gNB-DU transmits the first correspondence to the gNB-CU via the F1 interface. The gNB-CU transmits the first correspondence to the computation requesting node via the aforementioned RRC message, PDCP control PDU, Layer 1 protocol signaling, or CRC signaling.

[0230] Furthermore, when the CMF is integrated into the gNB-DU and transmits the first correspondence to the computation requesting node via RRC messages, PDCP control PDUs, first protocol layer signaling, or CRC signaling, the gNB-DU needs to transmit the first correspondence to the gNB-CU via the F1 interface. If the gNB-CU is further separated into the gNB-CU-CP and the gNB-CU-UP, and the PDCP control PDU is used to carry the first correspondence, the gNB-CU-CP further transmits the first correspondence to the gNB-CU-UP.

[0231] Figure 4 shows a schematic diagram of the air interface protocol stack for a scenario where the gNB-CU and gNB-DU are separated, and the gNB-CU is further separated into the gNB-CU-CP and gNB-CU-UP. As shown in Figure 4, the gNB-CU-CP and gNB-CU-UP logical functional units may include the PDCP protocol layer. The RRC protocol layer resides above the PDCP protocol layer of the gNB-CU-CP. Specifically, RRC messages transmitted to the PDCP protocol layer are encapsulated as PDCP data PDUs. PDCP data PDUs can be sent by the gNB-CU-CP and gNB-CU-UP, respectively. The PDCP control PDU is generated by the PDCP layer itself and is used to convey robust header compression (ROHC) context information and is typically sent by the PDCP protocol layer of the gNB-CU-UP. The gNB-DU is responsible for processing physical layer protocols and real-time services, implementing the functions of the radio link control (RLC), media access control (MAC), and physical (PHY) layers.

[0232] Similarly, when the computing execution nodes are different types of nodes and the computing management function network elements are deployed at different locations, the messages transmitted between the computing execution nodes and the computing management function network elements (for example, the first information, the second corresponding relationship) can be carried on different signaling.

[0233] Exemplarily, when the computing execution node is an application server deployed in a data network (DN) and the CMF is deployed in a RAN node (for example, a gNB in ​​an NR system), the second correspondence may be carried in a user plane part of the GPRS tunneling protocol (GTP-U) header or an application programming interface (API) interface message of a network capability exposure function.

[0234] For example, the gNB sends the second correspondence to the UPF via one or more GTP-Us of uplink data; the UPF detects GTP-U header information of the uplink data and obtains the second correspondence; the UPF sends the second correspondence to the application server via the NEF or the local NEF (Local-NEF) on the RAN side in an API manner. Alternatively, the UPF modifies the header of a transmitted data packet to add the second correspondence and sends the second correspondence to the server. For example, the UPF adds the second correspondence to the header of a data packet encapsulated using the User Datagram Protocol (UDP), Quick UDP Internet Connections (QUIC), or Real-time Transport Protocol (RTP).

[0235] For the gNB architecture with separate gNB-CU and gNB-DU, if the CMF is integrated into the gNB-DU and the second correspondence is determined by the gNB-DU, the gNB-DU needs to inform the gNB-CU of the second correspondence via the F1 interface. If the gNB-CU is further separated into the gNB-CU-CP and the gNB-CU-UP, the gNB-CU-CP needs to further inform the gNB-CU-UP of this information, or the gNB-DU can send the second correspondence to the gNB-CU-UP via the F1-u interface, for example, by carrying the second correspondence in the GTP-u header of the uplink data on the F1-u interface.

[0236] Optionally, the CMF sends the second correspondence to the application server through the UE. For example, the CMF may send the second correspondence to the terminal device; the terminal device carries the second correspondence in the header of the data packet sent to the computing execution node, or the application of the terminal device sends the second correspondence to the application server through metadata. Metadata can be a description and context of the data, and metadata can be used to organize, find, and understand the data.

[0237] In addition, the solution of the present application can also be applied to the computing network architecture based on Open-RAN (O-RAN). Figure 6 shows a schematic diagram of the structure of an O-RAN, which may include a gNB-CU, a gNB-DU, and a RAN Intelligent Controller (RIC). Furthermore, the RIC may include near real-time RIC (NRT-RIC) and non-real-time RIC (Non-Real-Time RIC, Non-RT RIC) functional modules.

[0238] In an O-RAN-based computing network architecture, CMF can be integrated with NRT-RIC, or NRT-RIC can be integrated with CMF, as shown in Figure 5(a) and Figure 5(b) respectively.

[0239] For the above-mentioned O-RAN architecture with integrated CMF, when the computation execution node is a gNB, the gNB needs to send the computation capability to the NRT-RIC through the E2 interface, or to the non-NRT RIC through the O1 interface.

[0240] When the calculation requesting node is a terminal device, NRT-RIC may send the first corresponding relationship to the terminal device in the following manner.

[0241] If the first correspondence is determined by NRT-RIC and DCI or MAC CE is used to carry the first correspondence, NRT-RIC needs to send the first correspondence to the gNB-DU. If the first correspondence is determined by NRT-RIC and RRC message or PDCP control PDU is used to carry the first correspondence, NRT-RIC needs to send the first correspondence to the gNB-CU.

[0242] If the gNB-CU is further separated into gNB-CU-CP and gNB-CU-UP, the NRT-RIC needs to send the first correspondence relationship to the gNB-CU-CP or gNB-CU-UP, respectively. The NRT-RIC can also use a similar method to send the second correspondence relationship to the computation execution node.

[0243] When the computing execution node is an application server deployed in the data network, the gNB-CU or gNB-CU-UP carries the second correspondence relationship through one or more GTP-U headers of uplink data and sends it to the UPF on the NG3 interface; the UPF detects the GTP-U header information of the uplink data and obtains the second correspondence relationship; the UPF sends the second correspondence relationship to the application server through the NEF or Local-NEF in an API manner.

[0244] The order of the steps in the above flowchart is determined according to the internal logic of the method. The serial numbers shown in the above flowchart are only examples and do not limit the order of the steps of this application.

[0245] The task scheduling method provided in the embodiment of the present application is described in detail above in conjunction with Figures 3 to 5. Below, the communication device provided in the embodiment of the present application is described in detail in conjunction with Figures 6 to 8. It should be understood that the description of the device embodiment corresponds to the description of the method embodiment. Therefore, for matters not described in detail, please refer to the method embodiment above. For the sake of brevity, it will not be repeated here.

[0246] 6 shows a schematic diagram of a communication device 600 provided in an embodiment of the present application. The device 600 includes an interface unit 610, which can be used to implement corresponding communication functions and can also be referred to as a communication interface or a communication unit.

[0247] Optionally, the apparatus 600 may further include a processing unit 620 , which may be configured to perform data processing.

[0248] Optionally, the device 600 also includes a storage unit, which can be used to store instructions and / or data. The processing unit 620 can read the instructions and / or data in the storage unit so that the device can implement the actions of different devices in the aforementioned method embodiments.

[0249] In one possible design, the device 600 may be the computing management function network element in the aforementioned embodiment, or a component (such as a chip) of the computing management function network element. The device 600 may implement the steps or processes executed by the computing management function network element in the above method embodiment. The interface unit 610 may be used to perform operations related to the transmission and reception of the computing management function network element in the above method embodiment; and the processing unit 620 may be used to perform operations related to the processing of the computing management function network element in the above method embodiment.

[0250] In another possible design, the device 600 can be the computing execution node in the aforementioned embodiment, or a component (such as a chip) of the computing execution node. The device 600 can implement the steps or processes corresponding to those performed by the computing execution node in the above method embodiment. The interface unit 610 can be used to perform the operations related to the transmission and reception of the computing execution node in the above method embodiment; and the processing unit 620 can be used to perform the operations related to the processing of the computing execution node in the above method embodiment.

[0251] In another possible design, the device 600 can be the computation request node in the aforementioned embodiment, or a component (e.g., a chip) of the computation request node. The device 600 can implement the steps or processes corresponding to the computation request node requests in the above method embodiments. The interface unit 610 can be used to request the computation request node's transceiver-related operations in the above method embodiments; and the processing unit 620 can be used to request the computation request node's processing-related operations in the above method embodiments.

[0252] FIG7 is a schematic block diagram of a communication device 700 provided in an embodiment of the present application. The device 700 includes a processor 710, which is coupled to a memory 720. Optionally, the memory 720 is further included. The memory 720 is configured to store computer programs or instructions and / or data. The processor 710 is configured to execute the computer programs or instructions stored in the memory 720, or read data stored in the memory 720, to perform the methods described in the above method embodiments.

[0253] Optionally, there are one or more processors 710 .

[0254] Optionally, there are one or more memories 720 .

[0255] Optionally, the memory 720 is integrated with the processor 710 or provided separately.

[0256] Optionally, as shown in FIG7 , the apparatus 700 further includes a transceiver 730, which is used to receive and / or transmit signals. For example, the processor 710 is used to control the transceiver 730 to receive and / or transmit signals. The transceiver 730 may also be referred to as an interface.

[0257] As a solution, the device 700 is used to implement the operations performed by the computing management function network element in each of the above method embodiments.

[0258] For example, the processor 710 is configured to execute computer programs or instructions stored in the memory 720 to implement the relevant operations of the computing management function network element in the above various method embodiments.

[0259] As another solution, the device 700 is used to implement the operations performed by the computing execution node in each of the above method embodiments.

[0260] For example, the processor 710 is configured to execute computer programs or instructions stored in the memory 720 to implement related operations of the computing execution node in the above various method embodiments.

[0261] As another solution, the device 700 is used to implement the operations performed by the computing request node in each of the above method embodiments.

[0262] For example, the processor 710 is configured to execute computer programs or instructions stored in the memory 720 to implement operations related to computing request nodes in the above various method embodiments.

[0263] During implementation, each step of the above method can be completed by an integrated logic circuit of the hardware in the processor 710 or by instructions in the form of software. The method disclosed in conjunction with the embodiments of the present application can be directly embodied as being executed by a hardware processor, or can be executed by a combination of hardware and software modules in the processor. The software module can be located in a storage medium mature in the art, such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 720, and the processor 710 reads the information in the memory 720 and completes the steps of the above method in combination with its hardware. To avoid repetition, it will not be described in detail here.

[0264] It should be understood that in the embodiments of the present application, the processor may be one or more integrated circuits for executing relevant programs to execute the method embodiments of the present application.

[0265] A processor (e.g., processor 710) may include one or more processors and be implemented as a combination of computing devices. The processor may include one or more of the following: a microprocessor, a microcontroller, a digital signal processor (DSP), a digital signal processing device (DSPD), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), a programmable logic device (PLD), gating logic, transistor logic, discrete hardware circuits, processing circuits, or other suitable hardware, firmware, and / or a combination of hardware and software to perform the various functions described in this disclosure. The processor may be a general-purpose processor or a special-purpose processor. For example, processor 710 may be a baseband processor or a central processing unit. A baseband processor may be used to process communication protocols and communication data. A central processing unit may be used to enable the device to execute software programs and process data in the software programs. In addition, a portion of the processor may also include non-volatile random access memory. For example, the processor may also store information about the device type.

[0266] In this application, the term "program" is used broadly to refer to software. Non-limiting examples of software include program code, program, subroutine, instruction, instruction set, code, code segment, software module, application, or software application. The program can be executed in a processor and / or computer to cause the device to perform the various functions and / or processes described in this application.

[0267] The memory (e.g., memory 720) can store data required by the processor (e.g., processor 710) when executing software. The memory can be implemented using any suitable storage technology. For example, the memory can be any available storage medium that can be accessed by the processor and / or computer. Non-limiting examples of storage media include random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM), static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct rambus RAM (DR RAM), removable media, optical disk storage, magnetic disk storage media, magnetic storage devices, flash memory, registers, state memory, remotely mounted storage, local or remote memory components, or any other medium capable of carrying or storing software, data, or information and accessible by a processor / computer. It should be noted that the memory described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0268] The memory (e.g., memory 720) and the processor (e.g., processor 710) may be provided separately or integrated together. The memory may be used to connect to the processor so that the processor can read information from the memory and store and / or write information in the memory. The memory may be integrated into the processor. The memory and the processor may be provided in an integrated circuit (e.g., the integrated circuit may be provided in a UE or other network node).

[0269] 8 is a schematic block diagram of a chip system 800 provided in an embodiment of the present application. The chip system 800 (or also referred to as a processing system) includes a logic circuit 810 and an input / output interface 820.

[0270] The logic circuit 810 may be a processing circuit in the chip system 800. The logic circuit 810 may be coupled to a storage unit and call instructions in the storage unit so that the chip system 800 can implement the methods and functions of the various embodiments of the present application. The input / output interface 820 may be an input / output circuit in the chip system 800, outputting information processed by the chip system 800 or inputting data or signaling information to be processed into the chip system 800 for processing.

[0271] As a solution, the chip system 800 is used to implement the operations performed by the computing management function network element in the above various method embodiments.

[0272] For example, the logic circuit 810 is used to implement the processing-related operations performed by the computing management function network element in the above method embodiment; the input / output interface 820 is used to implement the sending and / or receiving-related operations performed by the computing management function network element in the above method embodiment.

[0273] As another solution, the chip system 800 is used to implement the operations performed by the computing execution nodes in the above various method embodiments.

[0274] For example, the logic circuit 810 is used to implement the processing-related operations performed by the computing execution node in the above method embodiment; the input / output interface 820 is used to implement the sending and / or receiving-related operations performed by the computing execution node in the above method embodiment.

[0275] As another solution, the chip system 800 is used to implement the operations requested by the computing request node in each of the above method embodiments.

[0276] For example, the logic circuit 810 is used to implement the processing-related operations requested by the computing request node in the above method embodiment; the input / output interface 820 is used to implement the sending and / or receiving-related operations requested by the computing request node in the above method embodiment.

[0277] An embodiment of the present application further provides a computer-readable storage medium on which computer instructions are stored for implementing the methods executed by a communication device (such as a terminal device, a network device) in the above-mentioned method embodiments.

[0278] An embodiment of the present application also provides a computer program product comprising instructions, which, when executed by a computer, implement the methods performed by a communication device (such as a computing management function network element, a computing execution node, or a computing request node) in the above-mentioned method embodiments.

[0279] An embodiment of the present application also provides a communication system, which includes at least one of the computing management function network element, computing execution node and computing request node in the above embodiments.

[0280] The explanation of the relevant contents and beneficial effects of any of the above-mentioned devices can be referred to the corresponding method embodiments provided above, which will not be repeated here.

[0281] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above-mentioned units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0282] The units described above as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected according to actual needs to implement the solutions provided in this application.

[0283] In addition, each functional unit in each embodiment of the present application may be integrated into one unit, each unit may exist physically separately, or two or more units may be integrated into one unit.

[0284] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel 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.

[0285] When software is used for implementation, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable devices. For example, the computer can be a personal computer, a server, or a network device, etc. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website, computer, server or data center to another website, computer, server or data center by wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.). About computer-readable storage media, reference can be made to the above description.

[0286] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A task scheduling method, characterized in that: include: Obtaining the computing capacity of each computing execution node in at least one computing execution node; Determine a first correspondence and / or a second correspondence according to the computing capacity and the service quality requirement of the computing task, the first correspondence includes a correspondence between each computing execution node in the first computing execution node and the task to be executed, the first computing execution node includes one or more computing execution nodes in the at least one computing execution node for executing the computing task, the second correspondence includes a correspondence between a second computing execution node and the task to be executed, the second computing execution node is any computing execution node in the first computing execution node, and the task to be executed is all or part of the tasks in the computing task; Sending the first corresponding relationship to the computing requesting node; and / or, The second corresponding relationship is sent to the second computing execution node.

2. The method according to claim 1, characterized in that The computing capability indicates first information, wherein the first information includes at least one of the following: Performance information of tests performed according to at least one test data set, performance information of tests performed according to at least one model, and performance information of tests performed according to at least one test tool or test solution.

3. The method according to claim 2, characterized in that The computing capability further indicates at least one time, and the at least one time corresponds to the first information.

4. The method according to claim 2 or 3, characterized in that: The performance information of the test performed according to at least one test data set includes at least one of the following: The identification information of the at least one test data set, the number of training samples per unit time based on the at least one test data set as a training data set, the time it takes to complete the AI ​​model training based on the at least one test data set as a training data set, or the accuracy of training the AI ​​model based on the at least one test data set as a training data set.

5. The method according to claim 2 or 3, characterized in that: The performance information of the test performed according to at least one model includes at least one of the following: The identification information of the at least one model is based on the time taken for the at least one model to obtain the inference result, or the number of times the at least one model is executed within a unit time to obtain the inference result.

6. The method according to claim 2 or 3, characterized in that: The performance information of the test performed according to at least one test tool or test scheme includes at least one of the following: The identification information of the at least one test tool or test scheme, the rate at which encoding is performed according to the at least one test tool or test scheme, the duration for which encoding is performed according to the at least one test tool or test scheme, the rate at which decoding is performed according to the at least one test tool or test scheme, or the duration for which decoding is performed according to the at least one test tool or test scheme.

7. The method according to any one of claims 1 to 6, characterized in that The quality of service requirement includes at least one of the following information of the computing task: Task type, computing type, input data rate or data volume, output data rate or data volume, latency requirements, computing power requirements, and memory requirements.

8. The method according to any one of claims 1 to 7, characterized in that The method further comprises: A computing service request message is received from the computing request node, where the computing service request message includes the service quality requirement.

9. The method according to claim 8, characterized in that The computing task includes at least one subtask, and the computing service request message further includes association relationship information between the at least one subtask.

10. The method according to claim 9, characterized in that The association relationship information includes at least one of the following: Data flow direction, input or output data rate, input or output data throughput, input and output dependencies.

11. The method according to any one of claims 1 to 10, characterized in that The computing request node is a terminal device, and the computing service request message is carried in any of the following signaling: Radio resource control RRC signaling, signaling of the first protocol layer, computing resource control CRC signaling; The first protocol layer is located above the radio resource control RRC protocol layer or in parallel with the RRC protocol layer.

12. The method according to any one of claims 1 to 11, characterized in that The first corresponding relationship is carried in any of the following messages: Downlink control information DCI, media access control MAC control element CE, radio resource control RRC message, packet data convergence layer protocol PDCP control protocol data unit PDU, signaling of the first protocol layer, computing resource control CRC signaling; The first protocol layer is located above the radio resource control RRC protocol layer or in parallel with the RRC protocol layer.

13. The method according to any one of claims 1 to 10, characterized in that The computing execution node is an application server, and the second corresponding relationship is carried in any one of the following messages: Tunnel protocol user plane part GTP-U header, network capability exposure function NEF application programming interface API interface message.

14. The method according to any one of claims 1 to 13, characterized in that The method is applied to access network equipment, access network equipment distributed unit DU, or access network equipment centralized unit CU.

15. A method for task scheduling, characterized in that: include: Sending computing capacity to a computing management function network element, where the computing capacity is used to determine a second corresponding relationship, where the second corresponding relationship includes a corresponding relationship between a computing execution node and a task to be executed, where the task to be executed is all or part of the computing task; The second corresponding relationship is received from the computing management function network element.

16. The method according to claim 15, characterized in that The computing capability indicates first information, wherein the first information includes at least one of the following: Performance information of tests performed according to at least one test data set, performance information of tests performed according to at least one model, and performance information of tests performed according to at least one test tool or test solution.

17. The method according to claim 16, characterized in that The computing capability further indicates at least one time, and the at least one time corresponds to the first information.

18. The method according to claim 15 or 16, characterized in that The performance information of the test performed according to at least one test data set includes at least one of the following: The identification information of the at least one test data set, the number of training samples per unit time based on the at least one test data set as a training data set, the time it takes to complete the AI ​​model training based on the at least one test data set as a training data set, or the accuracy of training the AI ​​model based on the at least one test data set as a training data set.

19. The method according to claim 15 or 16, characterized in that The performance information of the test performed according to at least one model includes at least one of the following: The identification information of the at least one model is based on the time taken for the at least one model to obtain the inference result, or the number of times the at least one model is executed within a unit time to obtain the inference result.

20. The method according to claim 15 or 16, characterized in that The performance information of the test performed according to at least one test tool or test scheme includes at least one of the following: The identification information of the at least one test tool or test scheme, the rate at which encoding is performed according to the at least one test tool or test scheme, the duration for which encoding is performed according to the at least one test tool or test scheme, the rate at which decoding is performed according to the at least one test tool or test scheme, or the duration for which decoding is performed according to the at least one test tool or test scheme.

21. The method according to any one of claims 15 to 20, characterized in that The computing execution node is an application server, and the second corresponding relationship is carried in any one of the following messages: Tunnel protocol user plane part GTP-U header, network capability exposure function NEF application programming interface API interface message.

22. A method for task scheduling, characterized in that: include: Sending a service quality requirement of a computing task to a computing management function network element, where the service quality requirement is used to determine a first corresponding relationship, where the first corresponding relationship includes a corresponding relationship between each computing execution node in a first computing execution node and a task to be executed, where the first computing execution node includes one or more computing execution nodes in at least one computing execution node for executing the computing task, and the task to be executed is all or part of the computing task; Receive the first corresponding relationship from the computing management function network element.

23. The method according to claim 22, characterized in that The first corresponding relationship is determined by the quality of service requirement and the computing capability of the at least one computing execution node, the computing capability indicates first information, and the first information includes at least one of the following: Performance information of tests performed according to at least one test data set, performance information of tests performed according to at least one model, and performance information of tests performed according to at least one test tool or test solution.

24. The method according to claim 23, characterized in that The computing capability further indicates at least one time, and the at least one time corresponds to the first information.

25. The method according to claim 23 or 24, characterized in that The performance information of the test performed according to at least one test data set includes at least one of the following: The identification information of the at least one test data set, the number of training samples per unit time based on the at least one test data set as a training data set, the time it takes to complete the AI ​​model training based on the at least one test data set as a training data set, or the accuracy of training the AI ​​model based on the at least one test data set as a training data set.

26. The method according to claim 23 or 24, characterized in that The performance information of the test performed according to at least one model includes at least one of the following: The identification information of the at least one model is based on the time taken for the at least one model to obtain the inference result, or the number of times the at least one model is executed within a unit time to obtain the inference result.

27. The method according to claim 23 or 24, characterized in that The performance information of the test performed according to at least one test tool or test scheme includes at least one of the following: The identification information of the at least one test tool or test scheme, the rate at which encoding is performed according to the at least one test tool or test scheme, the duration for which encoding is performed according to the at least one test tool or test scheme, the rate at which decoding is performed according to the at least one test tool or test scheme, or the duration for which decoding is performed according to the at least one test tool or test scheme.

28. The method according to any one of claims 22 to 27, characterized in that The quality of service requirement includes at least one of the following information of the computing task: Task type, computing type, input data rate or data volume, output data rate or data volume, latency requirements, computing power requirements, and memory requirements.

29. The method according to any one of claims 22 to 27, characterized in that The service quality requirement is carried in the computing service request message.

30. The method according to claim 29, characterized in that The computing task includes at least one subtask, and the computing service request message further includes association relationship information between the at least one subtask.

31. The method according to claim 30, characterized in that The association relationship information includes at least one of the following: Data flow direction, input or output data rate, input or output data throughput, input and output dependencies.

32. The method according to any one of claims 22 to 31, characterized in that The computing request node is a terminal device, and the computing service request message is carried in any of the following signaling: Radio resource control RRC signaling, signaling of the first protocol layer, computing resource control CRC signaling; The first protocol layer is located above the radio resource control RRC protocol layer or in parallel with the RRC protocol layer.

33. The method according to any one of claims 22 to 32, characterized in that The first corresponding relationship is carried in any of the following messages: Downlink control information DCI, media access control MAC control element CE, radio resource control RRC message, packet data convergence layer protocol PDCP control protocol data unit PDU, signaling of the first protocol layer, computing resource control CRC signaling; The first protocol layer is located above the radio resource control RRC protocol layer or in parallel with the RRC protocol layer.

34. A communication device, characterized in that: The apparatus comprises a unit for executing the method according to any one of claims 1 to 14, or a unit for executing the method according to any one of claims 15 to 21, or a unit for executing the method according to any one of claims 22 to 33.

35. A communication device, characterized in that: include: A processor, the processor is coupled to a memory, the memory stores instructions, and when the instructions are executed by the processor, the device executes the method as claimed in any one of claims 1 to 14, or executes the method as claimed in any one of claims 15 to 21, or executes the method as claimed in any one of claims 22 to 33.

36. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store a computer program, and when the computer program is run on a computer, the device executes the method as claimed in any one of claims 1 to 14, or executes the method as claimed in any one of claims 15 to 21, or executes the method as claimed in any one of claims 22 to 33.

37. A computer program product, characterized in that The computer program product comprises a computer program or instructions, which, when executed on a communication device, causes the communication device to execute the method according to any one of claims 1 to 14, or the method according to any one of claims 15 to 21, or the method according to any one of claims 22 to 33.

Images