Service processing method and apparatus

Abstract

The present application provides a service processing method and apparatus. The method comprises: an MA receives a first request, acquires first network information and / or first service information of a first network on the basis of the first request, and determines an OA deployment strategy on the basis of the first network information and / or the first service information, wherein the OA deployment strategy comprises an identifier of a second computing node and OA deployment resource information; and the MA sends an OA deployment request to the second computing node, wherein the OA deployment request comprises the OA deployment resource information, the OA deployment request is used for requesting deployment of an OA in the second computing node, and the OA is used for orchestrating any service to obtain S tasks and determining a scheduling agent (SA) for scheduling the S tasks. Since the process of OA service orchestration is autonomously completed by the OA and does not rely on experience of developers, it is conducive to improving the service orchestration efficiency, reducing the cost, and shortening the service orchestration cycle.

Description

Business processing method and device

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims priority to the Chinese patent application filed with the State Intellectual Property Office of the People's Republic of China on January 16, 2024, with application number 202410064562.6 and application name "A Business Processing Method and Device", the entire contents of which are incorporated by reference into this application. Technical Field

[0003] The present invention relates to the field of artificial intelligence technology, and in particular to a business processing method and device. Background Art

[0004] Due to the large variety of services and complex processing processes, to improve service processing efficiency, services need to be orchestrated and scheduled to ensure that one or more compute nodes process the services and output the corresponding results. Orchestration can be understood as splitting a service into S tasks (S is greater than or equal to 1), with certain dependencies between each of the S tasks. Scheduling can be understood as assigning the S tasks to one or more compute nodes, which then process the S tasks and determine the corresponding computational results for each of the S tasks.

[0005] Currently, developers typically analyze the business, determine the appropriate orchestration and scheduling strategies, then split the business into S tasks based on the orchestration strategy. These tasks are then assigned to corresponding compute nodes based on the scheduling strategy. The compute nodes then execute the S tasks and determine the corresponding computational results. This approach relies entirely on the developer's experience, resulting in high costs and long development cycles. Rapidly obtaining the appropriate orchestration strategy for the business is a pressing technical challenge.

[0006] Summary of the Invention

[0007] The present application provides a service processing method and apparatus for improving service orchestration efficiency.

[0008] In a first aspect, an embodiment of the present application provides a business processing method, the execution subject of the method is a management agent MA deployed on a first computing node in a first network. The method includes: MA receives a first request, wherein the first request is used to request the generation of an orchestration agent OA. Based on the first request, MA obtains first network information and / or first business information of the first network, and then determines the OA deployment strategy based on the first network information and / or the first business information, wherein the OA deployment strategy includes an identifier of the second computing node and OA deployment resource information. MA then sends an OA deployment request to the second computing node based on the identifier of the second computing node, wherein the OA deployment request includes OA deployment resource information, and the OA deployment request is used to request the deployment of OA in the second computing node based on the OA deployment resource information. OA is used to orchestrate any business to obtain S tasks, and to determine a scheduling agent SA for scheduling S tasks, where S is a positive integer.

[0009] In the above method, the MA can directly determine the OA deployment strategy based on the first network information and / or the first business information, and deploy the OA in the second computing node based on the OA deployment strategy. The above OA deployment process is completed autonomously by the MA without the involvement of developers, which can improve the efficiency of OA deployment. The OA deployed in the second computing node can orchestrate any business to obtain S tasks. The above OA business orchestration process is completed autonomously by the OA and does not rely on the experience of developers. It is beneficial to improve the efficiency of business orchestration, while also reducing costs and shortening the business orchestration cycle.

[0010] In one possible design, the OA deployment resource information may include an OA deployment template or an OA executable code, which is used to deploy the OA; the OA system information in the OA deployment template includes the type of OA to be deployed; and the OA executable code includes the type of OA to be deployed.

[0011] In one possible design, the MA determines the OA deployment strategy based on the first network information and / or the first business information, which may include: the MA inputs the first network information and / or the first business information into the first model to obtain the OA deployment strategy, wherein the first model has the ability to generate the OA deployment strategy.

[0012] In one possible design, the above method may further include:

[0013] In the case where the OA deployment fails on the second computing node, the MA may receive an OA deployment failure message sent by the second computing node;

[0014] When the OA is successfully deployed on the second computing node, the MA can receive the OA deployment success message sent by the second computing node, wherein the OA deployment success message includes the identifier of the second computing node and the OA identifier of the OA deployed in the second computing node; the MA then adds the identifier of the second computing node and the OA identifier of the OA deployed in the second computing node to the first OA deployment relationship.

[0015] In one possible design, the method may further include: the MA receiving a second request, wherein the second request is for requesting adjustment of the OA; the MA then, based on the second request, obtaining second network information and / or second service information of the first network, wherein the second network information is the adjusted first network information, and the second service information is the adjusted first service information; the MA may determine an OA adjustment policy based on the second network information and / or second service information and the first OA deployment relationship, wherein the OA adjustment policy includes an identifier of the second compute node and the OA adjustment information. The MA may send an OA adjustment request to the second compute node based on the identifier of the second compute node, wherein the OA adjustment request includes the OA adjustment information, and the OA adjustment request is for requesting adjustment of the deployed OA in the second compute node based on the OA adjustment information. With this design, when the first network information and / or first service information changes, the MA may directly determine the OA adjustment policy based on the changed first network information and / or first service information and the first OA deployment relationship. The MA then adjusts the deployed OA in the second compute node based on the OA adjustment policy. The above OA adjustment process is performed autonomously by the MA without developer involvement, thereby improving OA adjustment efficiency. In addition, the OA adjustment strategy can be adjusted as the first network information and / or the first service information changes, which can effectively avoid the problem of an unreasonable number of deployed OAs.

[0016] In one possible design, the above method may further include:

[0017] In the case where the second computing node fails to adjust the deployed OA, the MA may receive an OA adjustment failure message sent by the second computing node;

[0018] When the second computing node successfully adjusts the deployed OA, the MA can receive an OA adjustment success message sent by the second computing node, wherein the OA adjustment success message includes the identifier of the second computing node and the OA identifier of the adjusted OA in the second computing node; the MA then modifies the first OA deployment relationship based on the identifier of the second computing node and the OA identifier of the adjusted OA.

[0019] In one possible design, the OA system information includes at least one of the following: role information, orchestration rules, first orchestration resource information, second orchestration resource information, orchestration memory information, and orchestration output format information; the orchestration role information is used to indicate the type to be deployed, the orchestration rules are the criteria used to orchestrate the business, the first orchestration resource information is used to indicate the resources used to orchestrate the business to obtain a first orchestration result, the second orchestration resource information is used to indicate the resources used to adjust the first orchestration result to obtain a second orchestration result, the orchestration memory information is used to indicate the memory used for the obtained orchestration result, and the orchestration output format information is used to indicate the output format of the orchestration result.

[0020] In one possible design, the first request is also used to request the generation of an SA. The above method may further include: the MA determining an SA deployment strategy based on the first network information and / or the first service information, wherein the SA deployment strategy includes an identifier of a third computing node and SA deployment resource information; the MA may send an SA deployment request to the third computing node based on the identifier of the third computing node, wherein the SA deployment request includes the SA deployment resource information, and the SA deployment request is used to request the deployment of the SA in the third computing node based on the SA deployment resource information. The SA is used to determine a fourth computing node that executes the received task.

[0021] Through this design, the MA can directly determine the SA deployment strategy based on the first network information and / or the first business information, and deploy the SA in the third computing node based on the SA deployment strategy. The above SA deployment process is completed autonomously by the MA without the involvement of developers, which can improve the deployment efficiency of SA. The SA deployed in the third computing node can schedule the received tasks and determine the computing node to execute the received tasks. The process of SA scheduling tasks is completed autonomously by the SA and does not rely on the experience of developers, which is conducive to improving task scheduling efficiency.

[0022] In one possible design, the SA deployment resource information includes an SA deployment template or SA executable code, which is used to deploy SA; the SA system information in the SA deployment template includes the type of SA to be deployed; and the SA executable code includes the type of SA to be deployed.

[0023] In one possible design, the MA determines the SA deployment strategy based on the first network information and / or the first business information, which may include: the MA inputs the first network information and / or the first business information into the first model to obtain the SA deployment strategy, wherein the first model also has the ability to generate the SA deployment strategy.

[0024] In one possible design, the above method may further include:

[0025] In the case that the deployment of the SA by the third computing node fails, the MA may receive an SA deployment failure message sent by the third computing node;

[0026] If the SA is successfully deployed on the third computing node, the MA can receive the SA deployment success message sent by the third computing node, where the SA deployment success message includes the identifier of the third computing node and the SA identifier of the SA deployed by the third computing node; the MA then adds the identifier of the third computing node and the SA identifier of the SA deployed in the third computing node to the first SA deployment relationship.

[0027] In one possible design, the method may further include: the MA receiving a second request, wherein the second request is for requesting SA adjustment; the MA then, based on the second request, obtaining second network information and / or second service information of the first network, wherein the second network information is the adjusted first network information, and the second service information is the adjusted first service information; the MA may determine an SA adjustment policy based on the second network information and / or second service information and the first SA deployment relationship, wherein the SA adjustment policy includes an identifier of a third computing node and the SA adjustment information; and the MA may send an SA adjustment request to the third computing node based on the identifier of the third computing node, wherein the SA adjustment request includes the SA adjustment information, and the SA adjustment request is for requesting adjustment of a deployed SA in the third computing node based on the SA adjustment information. With this design, when the first network information and / or first service information changes, the MA may directly determine an SA adjustment policy based on the changed first network information and / or first service information and the first SA deployment relationship. The MA then adjusts the deployed SA in the third computing node based on the SA deployment policy. The above SA adjustment process is performed autonomously by the MA without developer involvement, thereby improving SA adjustment efficiency. In addition, the SA adjustment strategy can be adjusted as the first network information and / or the first service information changes, which can effectively avoid the problem of an unreasonable number of deployed SAs.

[0028] In one possible design, the above method may further include:

[0029] In the case that the third computing node fails to adjust the deployed SA, the MA may receive an SA adjustment failure message sent by the third computing node;

[0030] When the third computing node successfully adjusts the deployed SA, the MA can receive an SA adjustment success message sent by the third computing node, wherein the SA adjustment success message includes the identifier of the third computing node and the SA identifier of the adjusted SA in the third computing node; the MA then modifies the first SA deployment relationship based on the identifier of the third computing node and the SA identifier of the adjusted SA.

[0031] In one possible design, the SA system information includes at least one of the following: role information, scheduling rules, first scheduling resource information, second scheduling resource information, scheduling memory information, and scheduling output format information; the scheduling role information is used to indicate the type to be deployed, the scheduling rules are the criteria used for scheduling the received tasks, the first scheduling resource information is used to indicate the resources used to schedule the received tasks to obtain the first scheduling result, the second scheduling resource information is used to indicate the resources used to adjust the first scheduling result to obtain the second scheduling result, the scheduling memory information is used to indicate the memory used for the obtained scheduling result, and the scheduling output format information is used to indicate the output format of the scheduling result.

[0032] In a second aspect, embodiments of the present application provide a service processing method, performed by a second computing node in a first network. The method comprises: the second computing node receiving an OA deployment request sent by an MA in the first computing node, wherein the OA deployment request includes OA deployment resource information; the second computing node may deploy the OA on the second computing node based on the OA deployment resource information; wherein the OA is used to orchestrate any service to obtain S tasks and determine an SA for scheduling the S tasks, where S is a positive integer.

[0033] In the above method, the OA deployed in the second computing node can orchestrate any business to obtain S tasks. The above OA orchestration process is completed by OA independently and does not rely on the experience of developers, which is conducive to improving business orchestration efficiency. At the same time, it can also reduce costs and shorten the business orchestration cycle.

[0034] In one possible design, the OA deployment resource information includes an OA deployment template or an OA executable code, which is used to deploy the OA; the OA system information in the OA deployment template includes the type of OA to be deployed; and the OA executable code includes the type of OA to be deployed.

[0035] In one possible design, the OA deployment resource information includes an OA deployment template;

[0036] The second computing node deploys OA on the second computing node based on the OA deployment resource information, which may include: the second computing node obtains OA mirror resources based on the OA deployment template, wherein the OA mirror resources are used to deploy OA; the second computing node then uses the OA system information in the OA deployment template to update the mirror parameters in the OA mirror resources to obtain the updated OA mirror resources; finally, the second computing node can use the updated OA mirror resources to deploy OA in the second computing node.

[0037] In one possible design, the OA deployment resource information includes OA executable code;

[0038] The second computing node deploys the OA on the second computing node based on the OA deployment resource information, which may include: the second computing node may use the OA executable code to deploy the OA in the second computing node.

[0039] In one possible design, the above method may further include: the second computing node receives an OA adjustment request sent by the MA, wherein the OA adjustment request includes OA adjustment information; the second computing node may adjust the OA deployed in the second computing node based on the OA adjustment information.

[0040] In a third aspect, embodiments of the present application provide a service processing method, performed by a third computing node in a first network. The method includes: receiving, by the third computing node, an SA deployment request, wherein the SA deployment request includes SA deployment resource information; and deploying, by the third computing node, an SA based on the SA deployment resource information, wherein the SA is used to determine a fourth computing node to execute the received task.

[0041] In the above method, the SA deployed in the third computing node can schedule the received tasks and determine the computing node to execute the received tasks. The process of SA scheduling tasks is completed independently by SA and does not rely on the experience of developers, which is conducive to improving task scheduling efficiency.

[0042] In one possible design, the SA deployment request comes from the MA in the first computing node, or the OA in the second computing node.

[0043] In one possible design, the SA deployment resource information includes an SA deployment template or SA executable code, which is used to deploy SA; the SA system information in the SA deployment template includes the type of SA to be deployed; and the SA executable code includes the type of SA to be deployed.

[0044] In one possible design, the SA deployment resource information includes an SA deployment template; the third computing node deploys SA on the third computing node based on the SA deployment resource information, which may include: the third computing node obtains SA mirror resources based on the SA deployment template in the SA deployment template, wherein the SA mirror resources are used to deploy SA; the third computing node then uses the SA system information in the SA deployment template to update the mirror parameters in the SA mirror resources to obtain updated SA mirror resources; finally, the third computing node uses the updated SA mirror resources to deploy SA in the third computing node.

[0045] In one possible design, the SA deployment resource information includes SA executable code; the third computing node deploys SA on the third computing node based on the SA deployment resource information, which may include: the third computing node may use the SA executable code to deploy SA in the third computing node.

[0046] In one possible design, the above method may further include: the third computing node receives an SA adjustment request, wherein the SA adjustment request includes SA adjustment information; and the third computing node then adjusts the SA deployed in the third computing node based on the SA adjustment information.

[0047] In a fourth aspect, the present application further provides a service processing device capable of implementing any of the methods provided in any of the first to third aspects above. The service processing device may be implemented in hardware or by hardware executing corresponding software. The hardware or software includes one or more units or modules corresponding to the above-mentioned functions.

[0048] In one possible design, the service processing device includes a processor configured to support the service processing device in executing the corresponding functions performed by the MA in the first computing node, the second computing node, or the third computing node in the method described above. The service processing device may also include a memory, which may be coupled to the processor and stores program instructions and data necessary for the service processing device.

[0049] In one possible design, the service processing device includes corresponding functional modules for implementing the steps in the above method. The functions can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above functions.

[0050] In one possible design, the structure of the business processing device includes a processing unit and a communication unit, which can perform the corresponding functions in the above method examples. For details, please refer to the description of the method provided in any one of the first to third aspects, which will not be repeated here.

[0051] In a fifth aspect, embodiments of the present application further provide a service processing device, comprising modules / units for executing any of the methods provided in any of the first to third aspects. These modules / units may be implemented in hardware, or in hardware executing corresponding software implementations.

[0052] In the sixth aspect, an embodiment of the present application provides a business processing device, comprising a memory and a processor; wherein the processor is used to execute a computer program or instruction stored in the memory, so that the business processing device implements any method provided in any one of the first to third aspects above.

[0053] In the seventh aspect, an embodiment of the present application also provides a computer-readable storage medium, which includes a computer program. When the computer program runs on a business processing device, the computer-readable storage medium implements any method provided in any one of the first to third aspects above.

[0054] In an eighth aspect, an embodiment of the present application further provides a computer program product, which, when running on a business processing device, enables the business processing device to implement the above-mentioned first to third aspects, as well as any possible design method of the first to third aspects.

[0055] In a ninth aspect, a chip is provided, comprising a processor and possibly a memory, for implementing the method of any of the first through third aspects, and any possible implementation of any of the aspects. The chip may be comprised of a single chip or may include a chip and other discrete components.

[0056] In a tenth aspect, a business processing system is provided, comprising: a MA of a first computing node, a second computing node, and a third computing node;

[0057] The MA of the first computing node is used to implement the method in the aforementioned first aspect and any possible implementation of the first aspect; the second computing node is used to implement the method in the aforementioned second aspect and any possible implementation of the second aspect; the third computing node is used to implement the method in the aforementioned second aspect and any possible implementation of the second aspect.

[0058] These and other aspects of the present application will become more readily apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] FIG1 is a schematic diagram of the structure of a device-edge cloud provided in an embodiment of the present application;

[0060] FIG2 is a flow chart of a method for orchestrating intelligent agents and scheduling intelligent agent deployment provided in an embodiment of the present application;

[0061] FIG3 is a schematic structural diagram of an arrangement method provided in an embodiment of the present application;

[0062] FIG4( a ) is a schematic diagram of a structure of a dependency relationship between a first adaptation model and a second model provided in an embodiment of the present application;

[0063] FIG4( b ) is a schematic structural diagram of a dependency relationship between a first adaptation model and a second model provided in an embodiment of the present application;

[0064] FIG4( c ) is a schematic diagram of a structure of a dependency relationship between a first adaptation model and a second model provided in an embodiment of the present application;

[0065] FIG4( d ) is a schematic diagram of a structure of a dependency relationship between a first adaptation model and a second model provided in an embodiment of the present application;

[0066] FIG5( a ) is a schematic diagram of a structure of a dependency relationship between tasks provided in an embodiment of the present application;

[0067] FIG5( b ) is a schematic diagram of a structure of a dependency relationship between tasks provided in an embodiment of the present application;

[0068] FIG6 is a flow chart of a method for orchestrating intelligent agents and scheduling intelligent agent deployment provided in an embodiment of the present application;

[0069] FIG7 is a flow chart of a method for orchestrating intelligent agents and scheduling intelligent agent deployment provided in an embodiment of the present application;

[0070] FIG8 is a flow chart of a method for orchestrating intelligent agents and scheduling intelligent agent deployment provided in an embodiment of the present application;

[0071] FIG9 is a flow chart of a method for arranging an agent and adjusting a scheduling agent provided in an embodiment of the present application;

[0072] FIG10 is a flow chart of a method for arranging an agent and adjusting a scheduling agent provided in an embodiment of the present application;

[0073] FIG11 is a flow chart of a method for arranging an agent and adjusting a scheduling agent provided in an embodiment of the present application;

[0074] FIG12 is a flow chart of a method for arranging an agent and adjusting a scheduling agent provided in an embodiment of the present application;

[0075] FIG13 is a schematic diagram of the structure of a service processing device provided in an embodiment of the present application;

[0076] FIG14 is a schematic structural diagram of a business processing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0077] In order to make the purpose, technical solutions and advantages of this application more clear, the application will be further described in detail below with reference to the accompanying drawings. The specific operation methods and functional descriptions in the method embodiments can also be applied to the device embodiments or system embodiments.

[0078] Below, some terms used in the embodiments of the present application are explained to facilitate understanding by those skilled in the art.

[0079] Large language models (LLMs) are deep learning models trained using large amounts of text data. They can generate natural language text or understand the meaning of text. Large language models can handle a variety of natural language tasks, such as text classification, question-answering, and conversation, and are a key path to artificial intelligence.

[0080] Prompt information, also known as guidance information, is a piece of text used to guide the large language model to generate specific types of output. By designing and optimizing a set of prompt information, the large language model can perform specific tasks.

[0081] The end-edge-cloud distributed network consists of the cloud, edge, and endpoint. The cloud refers to the nodes that provide cloud computing, cloud storage, and cloud applications. Edges refer to edge nodes such as the network edge, mobile edge, and IoT edge. They connect the cloud and endpoints, enabling high-speed data transmission and collaborative processing. Endpoints, or terminals, can be various devices, sensors, and intelligent hardware, such as mobile phones, computers, tablets, drones, robots, and cars. They generate large amounts of data and provide data support for the cloud.

[0082] Figure 1 shows a schematic diagram of the structure of an end-edge-cloud distributed network. Ends and edges can be connected directly or indirectly via wired or wireless communication, without limitation. Edges and clouds can be connected directly or indirectly via wired or wireless communication, without limitation. Furthermore, different edges connected to the same cloud can be connected directly or indirectly via wired or wireless communication, without limitation. Different clouds can be connected directly or indirectly via wired or wireless communication, without limitation.

[0083] Next, the business processing method provided by the embodiment of the present application is introduced with reference to the accompanying drawings. The business processing method provided by the embodiment of the present application can be applied to a computing node in a first network. The first network is composed of at least one computing node, and the first network can be any end-edge-cloud distributed network. If the first network is an end-edge-cloud distributed network, the at least one computing node includes a cloud node, an edge node, and an end node.

[0084] Figure 2 shows a flow chart of the orchestration agent (OA) and scheduling agent (SA) deployment method provided by an embodiment of the present application, which is interactively executed by the management agent (MA) in the first computing node, the second computing node, and the third computing node, wherein the first computing node can be any node in the first network, the second computing node can be any node in the first network, the third computing node can be any node in the first network, and the first computing node, the second computing node, and the third computing node can be the same computing node or different computing nodes. The OA and SA deployment methods provided by the embodiment of the present application may include the following steps.

[0085] S201: The MA receives a first request, wherein the first request is used to request the generation of an OA and a SA.

[0086] In the embodiments of the present application, the first request can be sent by a developer to the MA via any computing node in the first network, or via any computing node in another network. The first request can be sent to the MA by a network device in the first network, or by a network device in another network, without limitation. The network device includes at least one of the following: a core network element, an access network device, etc.

[0087] S202: The MA obtains first network information and / or first service information of the first network based on the first request.

[0088] In an embodiment of the present application, the first network information may include at least one of the following: computing power resource information of each computing node in the first network, network relationships of each computing node in the first network, data collection capabilities of each computing node in the first network, and capabilities of algorithm resources provided by each computing node in the first network.

[0089] In one possible implementation, the computing resource information of a computing node includes at least one of the following: the computing resource type of the computing node, the number of computing resources of the computing node, the memory capacity of the computing node, and the storage capacity of the computing node. The computing resource type of the computing node may include the graphics processing unit (GPU) resources, central processing unit (CPU) resources, embedded neural network processor (NPU) resources, tensor processing unit (TPU) resources, etc., of the computing node, without limitation. The number of computing resources of a computing node may include the total number of computing resources of the computing node, the available amount of computing resources of the computing node, the used amount of computing resources of the computing node, etc.

[0090] The network relationship of computing nodes includes the network topology relationship of computing nodes, the network transmission performance of computing nodes, etc., which are not limited here. The network transmission performance of computing nodes includes transmission rate, bandwidth, throughput, latency, etc.

[0091] The data collection capability of a computing node includes whether the computing node has the ability to collect data through a data collector, where the data collector includes a camera, a voice collector, a locator, etc., which are not limited here.

[0092] The capabilities of the algorithm resources provided by the computing node include the algorithm resources that the computing node can provide externally. For example, the computing node can provide face recognition algorithms, object detection algorithms, etc.

[0093] In the embodiment of the present application, the first service information may be service information of any service and is not limited here.

[0094] S203: The MA determines an OA deployment strategy and an SA deployment strategy based on the first network information and / or the first service information.

[0095] In a possible implementation, the MA inputs the first network information and / or the first business information into the first model to obtain the OA deployment strategy and the SA deployment strategy, wherein the first model has the ability to generate the OA deployment strategy and the SA deployment strategy.

[0096] It should be understood that the first model can be a large language model, etc., which is not limited here. The first model can run on the first computing node, or on other computing nodes in the first network, or on computing nodes in other networks, which is not limited here.

[0097] When the first model runs on the first computing node, the MA may directly input the first network information and / or the first service information into the first model to obtain the OA deployment strategy and the SA deployment strategy.

[0098] When the first model is running on other computing nodes in the first network, the MA can first obtain the first model from the other computing nodes in the first network, run the first model on the first computing node, and then input the first network information and / or the first business information into the first model to obtain the OA deployment policy and the SA deployment policy. Alternatively, the MA can also send the first network information and / or the first business information to other computing nodes in the first network, and the other computing nodes in the first network input the first network information and / or the first business information into the first model to obtain the OA deployment policy and the SA deployment policy. The other computing nodes in the first network then send the OA deployment policy and the SA deployment policy to the MA in the first computing node.

[0099] When the first model is running on a computing node in another network, the MA can first obtain the first model from the computing node in the other network, run the first model on the first computing node, and then input the first network information and / or the first business information into the first model to obtain the OA deployment policy and the SA deployment policy. Alternatively, the MA can also send the first network information and / or the first business information to the computing node in the other network, and the computing node in the other network inputs the first network information and / or the first business information into the first model to obtain the OA deployment policy and the SA deployment policy. The computing node in the other network then sends the OA deployment policy and the SA deployment policy to the MA in the first computing node.

[0100] In one possible implementation, the OA deployment strategy may include the number of OAs to be deployed, the identifier of the computing node on which each OA to be deployed is to be deployed, the computing power resource information required to deploy each OA to be deployed, and the OA deployment resource information corresponding to each OA to be deployed, wherein the OA deployment resource information is used to deploy the OA.

[0101] In one possible implementation, the OA deployment strategy may include the number of OAs to be deployed (M), the identifiers of M second computing nodes, the computing power resource information required to deploy the M OAs to be deployed, and the M OA deployment resource information, wherein the second computing node is a computing node on which the OA can be deployed, the identifiers of the M second computing nodes correspond one-to-one to the M OA deployment resource information, and M is a positive integer.

[0102] For ease of description, the steps in Figure 2 are discussed using the OA deployment strategy as an example, including the number of OAs to be deployed (1), the identifier of the second computing node (1), the computing power resource information required to deploy the OA (1), and the OA deployment resource information (1).

[0103] In one possible implementation, the computing resource information required for deploying OA includes at least one of the following: the type of computing resources required for deploying OA, the amount of computing resources required for deploying OA, the memory capacity required for deploying OA, and the storage capacity required for deploying OA, etc.

[0104] In a possible implementation, the OA deployment resource information includes an OA deployment template or an OA executable code, and the OA deployment template or the OA executable code is used to deploy the OA.

[0105] In the case where the OA deployment resource information includes an OA executable code, the OA executable code includes the type of OA to be deployed. Different OA executable codes result in different types of OA to be deployed.

[0106] When the OA deployment resource information includes an OA deployment template, the OA deployment template includes OA system information and OA user information. The OA system information includes the type of OA to be deployed. Different OA system information results in different OAs to be deployed. The OA user information includes the user's orchestration request description information.

[0107] In a possible implementation, the OA system information includes at least one of the following: role information, orchestration rules, first orchestration resource information, second orchestration resource information, orchestration memory information, and orchestration output format information.

[0108] The following is a detailed explanation of the contents of the OA system information:

[0109] First, the role information is used to indicate the type of the agent to be deployed. For example, if the role information is OA, it indicates that the agent to be deployed is an orchestration agent.

[0110] Second, the orchestration rules are the criteria used to orchestrate the service. The orchestration rules include at least one of the following: quality of service (QoS), a first upper limit, and a second upper limit. The quality of service is the upper limit of the execution time of the tasks obtained by orchestrating the service. The quality of service can be the upper limit of the total execution time corresponding to the service or the upper limit of the execution time corresponding to each task. The first upper limit is the upper limit of the number of tasks obtained by orchestrating the service, and the second upper limit is the upper limit of the number of each type of tasks obtained by orchestrating the service.

[0111] As an example, let's take the quality of service as an example, where the upper limit of the execution time for each task is used. The orchestration rules are as follows:

[0112] <Service Quality: 1s>

[0113] <First Upper Limit: 10>

[0114] <Second upper limit: 3>

[0115] After OA deployment is complete, OA can orchestrate any business according to the orchestration rules to obtain S tasks. For example, if OA receives a face recognition business, it can orchestrate the face recognition business based on the above orchestration rules to obtain S tasks, where the S tasks include s1 data acquisition type tasks, s2 reasoning type tasks, and s3 post-processing type tasks (S = s1 + s2 + s3). The execution time of each of the S tasks is capped at 1 second, S is an integer less than or equal to 10, and s1, s2, and s3 are all integers less than or equal to 3.

[0116] Third, the first orchestration resource information is used to indicate the resources used to orchestrate the service and obtain the first orchestration result. The first orchestration resource information includes at least one of the following: version information and location information of the second model, or version information and location information of the orchestration tool. The second model or the orchestration tool is used to orchestrate the service and obtain the first orchestration result. Different second models with different version information and location information are also different second models. Different orchestration tools with different version information and location information are also different orchestration tools.

[0117] It should be understood that the second model can run on the second computing node, other computing nodes in the first network, or computing nodes in other networks, without limitation. The orchestration tool can run on the second computing node, other computing nodes in the first network, or computing nodes in other networks, without limitation.

[0118] In the case where the first orchestration resource information includes version information and location information of the orchestration tool, after the OA deployment is completed, the OA receives the first business information. The OA can input the orchestration rules and the first business information into the orchestration tool. The orchestration tool orchestrates the first business and obtains an orchestration result, wherein the orchestration result includes S tasks and the identifier of the SA that schedules the S tasks.

[0119] In the case where the first orchestration resource information includes the version information and location information of the second model, after the OA deployment is completed, the OA receives the first business information. The OA can input the orchestration rules and the first business information into the second model. The second model orchestrates the first business to obtain an orchestration result, wherein the orchestration result includes S tasks and the identifier of the SA that schedules the S tasks.

[0120] For example, after the OA deployment is completed, if the OA receives the face recognition service, the OA can input the orchestration rules and the business information of the face recognition service into the second model. The second model orchestrates the face recognition service and outputs the orchestration results, where the orchestration results include S tasks, and the S tasks include s1 data acquisition type tasks, s2 reasoning type tasks, and s3 post-processing type tasks, S = s1 + s2 + s3.

[0121] In one possible implementation, the OA can input the first guidance information, orchestration rules, and first business information into the second model, which then orchestrates the first business to obtain an orchestration result. In this method, since the first guidance information is also input into the second model, the accuracy of the orchestration result output by the second model can be improved.

[0122] It should be understood that the arranged first guidance information, arrangement rules and first business information can be input into the second model. The arranged first guidance information, arrangement rules and first business information are shown in Figure 3, wherein the first guidance information and arrangement rules are generally unchanged, and the first business information is generally variable.

[0123] When the second model processes the first guidance information, it needs to calculate the feature vector of the first guidance information through each operator layer in the second model. Since the first guidance information is generally unchanging, the calculated feature vector of the first guidance information can be stored in the orchestration memory. Similarly, since the orchestration rules are generally unchanging, the calculated feature vector of the orchestration rules can also be stored in the orchestration memory.

[0124] In the case where the orchestration memory stores the feature vector of the first guidance information and the feature vector of the orchestration rule, if the OA inputs the first guidance information, the orchestration rule, and the first business information into the second model, each operator layer in the second model calculates the feature vector of the first business information, and no longer calculates the feature vector of the first guidance information and the feature vector of the orchestration rule. Instead, the operator layer directly obtains the feature vector of the first guidance information and the feature vector of the orchestration rule from the orchestration memory, and orchestrates the first business based on the feature vector of the first business information, the feature vector of the first guidance information, and the feature vector of the orchestration rule to obtain an orchestration result. The above method can avoid repeated calculation of the feature vector of the first guidance information and the feature vector of the orchestration rule, which is conducive to improving business orchestration efficiency.

[0125] Fourth, the second orchestration resource information is used to indicate the resources used to adjust the first orchestration result to obtain the second orchestration result. The second orchestration resource information includes the version information and location information of the first adaptation model. If the first orchestration resource information includes the version information of the second model, the second orchestration resource information also includes the dependency relationship between the first adaptation model and the second model; if the first orchestration resource information includes the version information of the orchestration tool, the second orchestration resource information also includes the dependency relationship between the first adaptation model and the orchestration tool. Different first adaptation models with different version information and location information will also result in different first adaptation models.

[0126] It should be understood that the first adaptation model can run on the second computing node, can also run on other computing nodes in the first network, and can also run on computing nodes in other networks, which is not limited here.

[0127] The dependency relationship between the first adaptation model and the second model includes any of the following:

[0128] The first dependency relationship is that the first adaptation model and the second model are in a series relationship. As shown in Figure 4(a), the first adaptation model and the second model are in a series relationship, and the first adaptation model is located before the second model. Alternatively, as shown in Figure 4(b), the first adaptation model and the second model are in a series relationship, and the first adaptation model is located after the second model. Alternatively, as shown in Figure 4(c), when there are two first adaptation models (i.e., the first adaptation model A and the first adaptation model B), the first adaptation model A, the first adaptation model B and the second model are in a series relationship, and the first adaptation model A is located before the second model, and the second model is located before the first adaptation model B.

[0129] The second dependency relationship is that the first adaptation model and the second model are in a parallel relationship. As shown in Figure 4(d), the first adaptation model and the second model are in a parallel relationship. The preprocessing module is connected in series with the first adaptation model and the second model, and the preprocessing module is located before the first adaptation model and the second model. The postprocessing module is connected in series with the first adaptation model and the second model, and the postprocessing module is located after the first adaptation model and the second model.

[0130] The dependency relationship between the first adaptation model and the orchestration tool can refer to the dependency relationship between the first adaptation model and the second model, which will not be described in detail here.

[0131] Fifth, orchestration memory information indicates the memory used for the obtained orchestration results. This information includes version and location information of the orchestration memory. The orchestration memory is used to store temporary information generated during the orchestration of services. Different orchestration memory versions and locations determine different orchestration memory locations.

[0132] It should be understood that the orchestration memory may be located in the second computing node, or in other computing nodes in the first network, or in computing nodes in other networks, which is not limited here.

[0133] The version information of the orchestration memory may correspond to the version information of the second model, and the temporary information generated by the second model may be stored in the orchestration memory corresponding to the version information.

[0134] Alternatively, the version information of the orchestration memory may correspond to the version information of the orchestration tool, and the temporary information generated by the orchestration tool may be stored in the orchestration memory corresponding to the version information.

[0135] Sixth, the orchestration output format information indicates the output format of the orchestration results. This information includes at least one of the following: task name, task type, dependencies between tasks, type of computing resources required for the task, amount of computing resources required for the task, memory capacity required for the task, storage capacity required for the task, the identifier of the compute node that schedules the task, and the identifier of the SA. The SA in the compute node that schedules the task can be determined based on the identifier of the compute node that schedules the task and the identifier of the SA.

[0136] In one possible implementation, after OA deployment is completed, OA obtains any business and orchestrates the business to obtain an orchestration result. The orchestration result includes S orchestration information. The S orchestration information corresponds one-to-one to the S tasks. Among the S orchestration information, the sth orchestration information includes the task name and task type of the sth task, the computing power resource type required for the sth task, the amount of computing power resources required for the sth task, the memory capacity required for the sth task, the storage capacity required for the sth task, the identifier of the computing node that schedules the sth task and the identifier of the SA, as well as the dependency relationship between the S tasks, where s is an integer that spans [1, S].

[0137] After determining the orchestration information corresponding to each of the S tasks, each task can be scheduled based on the orchestration information. For ease of understanding, the orchestration information can be referred to as information to be scheduled. How to schedule each task based on the orchestration information is described in detail later and is not detailed here.

[0138] It should be understood that for any task, there is a computing node and SA that schedules the task, and the computing node and SA that schedule the task are both one. The SA in a computing node can schedule one or more tasks.

[0139] For example, after the OA deployment is completed, if the OA receives the face recognition service, the OA orchestrates the face recognition service and obtains the orchestration result, as shown in Table 1. According to the orchestration result in Table 1, it can be determined that the OA orchestrates the face recognition service and obtains Task 1, Task 2, Task 3 and Task 4. The orchestration information corresponding to each of the above 4 tasks can be referred to Table 1.

[0140] Table 1

[0141] In addition, the above arrangement results also include the dependency relationship between the four tasks, as shown in Figure 5(a). Task 1 is connected in series with Task 2, Task 3, and Task 4, and Task 1 is located before Task 2, Task 3, and Task 4; Task 2 is connected in series with Task 1 and Task 4, and Task 2 is located after Task 1 and before Task 4, and Task 2 is connected in parallel with Task 3; Task 3 is connected in series with Task 1 and Task 4, and Task 3 is located after Task 1 and before Task 4, and Task 3 is connected in parallel with Task 2; Task 4 is connected in series with Task 1, Task 2, and Task 3, and Task 4 is located after Task 1, Task 2, and Task 3.

[0142] Regarding the orchestration information corresponding to Task 1 in Table 1, Task 1's task type is data acquisition. Scheduling Task 1 requires 0.5 CPU cores and 2 GB of memory. Task 1 can be scheduled by the SA identified as SA-2-1 on Compute Node 2. The description of Task 1 can be followed for the other tasks in Table 1 and is omitted here.

[0143] In one possible implementation, the SA deployment strategy may include the number of SAs to be deployed, the identifier of the computing node to be deployed for each SA to be deployed, the computing power resource information required to deploy each SA to be deployed, and the SA deployment resource information corresponding to each SA to be deployed, wherein the SA deployment resource information is used to deploy SA.

[0144] In one possible implementation, the SA deployment strategy may include the number of SAs to be deployed (N), the identifiers of N third computing nodes, the computing power resource information required to deploy the N SAs to be deployed, and N SA deployment resource information, wherein the third computing node is a computing node that can deploy SA, the identifiers of the N third computing nodes correspond one-to-one to the N SA deployment resource information, and N is a positive integer.

[0145] For ease of description, the steps in Figure 2 are discussed using the SA deployment strategy as an example, including the number of SAs to be deployed (1), the identifier of the third computing node (1), the computing power resource information required to deploy SA (1), and the SA deployment resource information (1).

[0146] In one possible implementation, the computing resource information required to deploy SA includes at least one of the following: the type of computing resources required to deploy SA, the amount of computing resources required to deploy SA, the memory capacity required to deploy SA, and the storage capacity required to deploy SA, etc.

[0147] In a possible implementation, the SA deployment resource information includes an SA deployment template or an SA executable code, and the SA deployment template or the SA executable code is used to deploy SA.

[0148] When the SA deployment resource information includes an SA executable code, the SA executable code includes the type of SA to be deployed. Different SA executable codes may result in different types of SA to be deployed.

[0149] When the SA deployment resource information includes an SA deployment template, the SA deployment template includes SA system information and SA user information. The SA system information includes the type of SA to be deployed. Different SA system information results in different SAs to be deployed. The SA user information includes the user's scheduling request description information.

[0150] In a possible implementation, the SA system information includes at least one of the following: role information, scheduling rules, first scheduling resource information, second scheduling resource information, scheduling memory information, and scheduling output format information.

[0151] The following is a detailed explanation of the contents of the SA system information:

[0152] First, the role information is used to indicate the type of the agent to be deployed. For example, if the role information is SA, it indicates that the agent to be deployed is a scheduling agent.

[0153] Second, scheduling rules are the criteria used to schedule received tasks. Scheduling rules include at least one of the following: scheduling time rules and resource usage rules. Scheduling time rules are used to indicate the time rules for scheduling received tasks, and resource usage rules are used to indicate resource usage rates or resource usage permissions.

[0154] In one possible implementation, the scheduling time rule may schedule tasks with the goal of minimizing task execution time, or the scheduling time rule may schedule tasks with the goal of ensuring that the task execution time does not exceed a first time upper limit, which is not limited here.

[0155] Resource utilization can be used to schedule tasks with the goal of minimizing resource utilization, without limitation. Resource usage permissions can be used to schedule tasks with the goal of preventing them from using a first resource, or with the goal of allowing them to use a first resource, without limitation. The first resource can be wireless transmission resources, computing power resources, or the like, without limitation.

[0156] For example, the scheduling rules are as follows:

[0157] <Scheduling time rule: Shortest task execution time>

[0158] <Resource Utilization: Lowest Resource Utilization>

[0159] <Resource usage rights: Wireless transmission resources cannot be used>

[0160] After OA and SA are deployed, OA orchestrates any business and obtains an orchestration result, which includes S orchestration information (information to be scheduled), and the S orchestration information corresponds to S tasks.

[0161] The above S tasks are all tasks to be scheduled. For the sth task among the S tasks, where s is an integer ranging from [1, S], OA performs the following steps:

[0162] The OA can send the sth information to be scheduled to the SA in the computing node that schedules the sth task based on the identifier of the computing node of the scheduled task in the sth information to be scheduled and the identifier of the SA. The SA in the computing node that schedules the sth task schedules the sth task according to the scheduling rules and the sth information to be scheduled, and determines the scheduling result of the sth task, where the sth task is the task to be scheduled.

[0163] Third, the first scheduling resource information is used to indicate the resources used to schedule the received task to obtain the first scheduling result. The first scheduling resource information includes at least one of the following: version information and location information of the third model, or version information and location information of the scheduling tool. The third model or scheduling tool is used to schedule the received task to obtain the first scheduling result. Different third model versions and location information result in different third models. Different scheduling tools result in different version information and location information.

[0164] It should be understood that the third model can run on the third computing node, other computing nodes in the first network, or computing nodes in other networks, without limitation herein. The scheduling tool can run on the third computing node, other computing nodes in the first network, or computing nodes in other networks, without limitation herein.

[0165] In the case where the first scheduling resource information includes the version information and location information of the scheduling tool, after the SA deployment is completed, SA receives the scheduling information of the first task sent by OA, wherein the first task is the task to be scheduled. SA can input the scheduling rules and the scheduling information of the first task into the scheduling tool, and the scheduling tool schedules the first task to obtain a scheduling result, wherein the scheduling result includes the identifier of the fourth computing node used to execute the first task.

[0166] In the case that the first scheduling resource information includes the version information and location information of the third model, after the SA deployment is completed, SA receives the scheduling information of the first task sent by OA. SA can input the scheduling rules and the scheduling information of the first task into the third model, and the third template schedules the first task to obtain the scheduling result, wherein the scheduling result includes the identifier of the fourth computing node used to execute the first task.

[0167] In one possible implementation, the SA can input the second guidance information, the scheduling rules, and the first task's to-be-scheduled information into a third model. The third template then schedules the first task to obtain a scheduling result. In this method, since the second guidance information is also input into the third model, the accuracy of the scheduling result output by the third model can be improved.

[0168] It should be understood that the arranged second guidance information, scheduling rules and the information to be scheduled of the first task can be input into the third model. The arranged second guidance information, scheduling rules and the information to be scheduled of the first task are shown in Figure 5(b), wherein the second guidance information and scheduling rules are generally unchanged, and the information to be scheduled of the first task is generally variable.

[0169] When the third model processes the second guidance information, it needs to calculate the feature vector of the second guidance information through each operator layer in the third model. Since the second guidance information is generally unchanged, the calculated feature vector of the second guidance information can be stored in the scheduling memory. Similarly, since the scheduling rules are generally unchanged, the calculated feature vector of the scheduling rules can also be stored in the scheduling memory.

[0170] In the case where the scheduling memory stores the feature vector of the second guidance information and the feature vector of the scheduling rule, if the SA inputs the second guidance information, the scheduling rule, and the information to be scheduled of the first task into the third model, each operator layer in the third model calculates the feature vector of the information to be scheduled of the first task, and no longer calculates the feature vector of the second guidance information and the feature vector of the scheduling rule. Instead, the feature vector of the second guidance information and the feature vector of the scheduling rule are directly obtained from the scheduling memory, and the first task is scheduled based on the feature vector of the information to be scheduled of the first task, the feature vector of the second guidance information, and the feature vector of the scheduling rule to obtain the scheduling result. The above method can avoid the repeated calculation of the feature vector of the second guidance information and the feature vector of the scheduling rule, which is conducive to improving the efficiency of task scheduling.

[0171] Fourth, the second scheduling resource information is used to indicate the resources used to adjust the first scheduling result to obtain the second scheduling result. The second scheduling resource information includes the version information and location information of the second adaptation model. If the first scheduling resource information includes the version information of the third model, the second scheduling resource information also includes the dependency relationship between the second adaptation model and the third model. If the first scheduling resource information includes the version information of the scheduling tool, the second scheduling resource information also includes the dependency relationship between the second adaptation model and the scheduling tool. Different second adaptation models with different version information and location information will also result in different second adaptation models.

[0172] It should be understood that the second adaptation model can run on the third computing node, can also run on other computing nodes in the first network, and can also run on computing nodes in other networks, which is not limited here.

[0173] The dependency relationship between the second adaptation model and the third model includes any of the following:

[0174] In the first dependency relationship, the second adaptation model and the third model are in a series relationship. The second adaptation model and the third model are in a series relationship, and the second adaptation model is located before the third model. Alternatively, the second adaptation model and the third model are in a series relationship, and the second adaptation model is located after the third model. Alternatively, when there are two second adaptation models (i.e., the second adaptation model A and the second adaptation model B), the second adaptation model A, the second adaptation model B, and the third model are in a series relationship, and the second adaptation model A is located before the third model, and the third model is located before the second adaptation model B.

[0175] The second dependency relationship is that the second adaptation model and the third model are in parallel. The pre-processing module is connected in series with the second adaptation model and the third model respectively, and the pre-processing module is located before the second adaptation model and the third model. The post-processing module is connected in series with the second adaptation model and the third model respectively, and the post-processing module is located after the second adaptation model and the third model.

[0176] The dependency relationship between the second adaptation model and the scheduling tool can refer to the dependency relationship between the second adaptation model and the third model, which will not be elaborated here.

[0177] Fifth, scheduling memory information indicates the memory used for the obtained scheduling results. This information includes the scheduling memory's version information and location information. The scheduling memory is used to store temporary information generated when scheduling received tasks. Different scheduling memory versions and locations result in different scheduling memory usage.

[0178] It should be understood that the scheduling memory may be located in the third computing node, or in other computing nodes in the first network, or in computing nodes in other networks, which is not limited here.

[0179] The version information of the scheduling memory may correspond to the version information of the third model, and the temporary information generated by the third model may be stored in the scheduling memory corresponding to the version information.

[0180] Alternatively, the version information of the scheduling memory may correspond to the version information of the scheduling tool, and the temporary information generated by the scheduling tool may be stored in the scheduling memory corresponding to the version information.

[0181] Sixth, the scheduling output format information is used to indicate the output format of the scheduling result.

[0182] The scheduling output format information includes the following two possible forms:

[0183] The first scheduling relationship includes a correspondence between the tasks received by the SA and the identifier of the fourth computing node.

[0184] For example, after OA and SA are deployed, if OA receives a face recognition service, it orchestrates the face recognition service and obtains an orchestration result. The specific content of the orchestration result can be found in Table 1. Based on the orchestration result in Table 1, it can be determined that OA orchestrates the face recognition service and obtains Task 1, Task 2, Task 3, and Task 4. The above four tasks are all tasks to be scheduled.

[0185] OA sends the scheduling information of task 1 to the SA identified as SA-2-1 in computing node 2, sends the scheduling information of task 2 to the SA identified as SA-4-1 in computing node 4, sends the scheduling information of task 3 to the SA identified as SA-4-1 in computing node 4, and sends the scheduling information of task 4 to the SA identified as SA-4-2 in computing node 4.

[0186] After the SA identified as SA-2-1 in computing node 2 receives the information to be scheduled of task 1, it determines that the fourth computing node to execute task 1 is computing node 10 according to the scheduling rules and the information to be scheduled of task 1; after the SA identified as SA-4-1 in computing node 4 receives the information to be scheduled of task 2, it determines that the fourth computing node to execute task 2 is computing node 11 according to the scheduling rules and the information to be scheduled of task 2; after the SA identified as SA-4-1 in computing node 4 receives the information to be scheduled of task 3, it determines that the fourth computing node to execute task 3 is computing node 12 according to the scheduling rules and the information to be scheduled of task 3; after the SA identified as SA-4-2 in computing node 4 receives the information to be scheduled of task 4, it determines that the fourth computing node to execute task 4 is computing node 13 according to the scheduling rules and the information to be scheduled of task 4.

[0187] In summary, after SA schedules the four tasks, the first scheduling relationship can be determined as follows:

[0188] The correspondence between the task and the identifier of the fourth computing node

[0189] <Task 1: Compute Node 10>

[0190] <Task 2: Compute Node 11>

[0191] <Task 3: Compute Node 12>

[0192] <Task 4: Compute Node 13>

[0193] The second scheduling relationship includes the correspondence between the tasks received by the SA and the task types, the correspondence between the task types and the identifiers of the execution units, and the correspondence between the identifiers of the execution units and the identifiers of the fourth computing nodes.

[0194] For example, after OA and SA are deployed, if OA receives a face recognition service, OA orchestrates the face recognition service and obtains an orchestration result. The specific content of the orchestration result can be found in Table 1. Based on the orchestration result in Table 1, it can be determined that OA orchestrates the face recognition service to obtain Task 1, Task 2, Task 3, and Task 4. The above four tasks are all tasks to be scheduled. Among them, the task type of Task 1 is data acquisition type, the task type of Task 2 is reasoning type, the task type of Task 3 is reasoning type, and the task type of Task 4 is post-processing type.

[0195] OA sends the scheduling information of task 1 to the SA identified as SA-2-1 in computing node 2, sends the scheduling information of task 2 to the SA identified as SA-4-1 in computing node 4, sends the scheduling information of task 3 to the SA identified as SA-4-1 in computing node 4, and sends the scheduling information of task 4 to the SA identified as SA-4-2 in computing node 4.

[0196] After the SA identified as SA-2-1 in computing node 2 receives the scheduling information of task 1, it determines the task type of task 1 based on the scheduling information of task 1, and then determines the execution unit corresponding to the task type of task 1, namely, execution unit 1, based on the scheduling rules and the scheduling information of task 1, where the execution unit is located in computing node 10.

[0197] After the SA identified as SA-4-1 in computing node 4 receives the scheduling information of task 2, it determines the task type of task 2 based on the scheduling information of task 2, and then determines the execution unit corresponding to the task type of task 2, namely, execution unit 2, based on the scheduling rules and the scheduling information of task 2, where execution unit 2 is located in computing node 11.

[0198] After the SA identified as SA-4-1 in computing node 4 receives the scheduling information of task 3, it determines the task type of task 3 based on the scheduling information of task 3, and then determines the execution unit corresponding to the task type of task 3, namely, execution unit 4, based on the scheduling rules and the scheduling information of task 3, where execution unit 4 is located in computing node 13.

[0199] After the SA identified as SA-4-2 in computing node 4 receives the scheduling information of task 4, it determines the task type of task 4 based on the scheduling information of task 4, and then determines that the execution unit corresponding to the task type of task 4 is execution unit 3 based on the scheduling rules and the scheduling information of task 4, where execution unit 3 is located in computing node 12.

[0200] In summary, after SA schedules the four tasks, the second scheduling relationship can be determined as follows:

[0201] The correspondence between tasks and task types

[0202] <Task 1: Data Collection Type>

[0203] <Task 2: Reasoning Type>

[0204] <Task 3: Reasoning Type>

[0205] <Task 4: Post-processing Type>

[0206] The correspondence between task types and execution unit identifiers

[0207] <Data Collection Type: Execution Unit 1>

[0208] <Inference Type: Execution Unit 2, Execution Unit 4>

[0209] <Post-processing type: Execution unit 3>

[0210] Correspondence between the identifier of the execution unit and the identifier of the fourth computing node

[0211] <Execution Unit 1: Compute Node 10>

[0212] <Execution Unit 2: Compute Node 11>

[0213] <Execution Unit 4: Compute Node 13>

[0214] <Execution Unit 3: Compute Node 12>

[0215] S204: The MA sends an OA deployment request to the second computing node based on the identifier of the second computing node. Correspondingly, the second computing node receives the OA deployment request sent by the MA.

[0216] In an embodiment of the present application, the OA deployment request includes OA deployment resource information and computing power resource information required for deploying OA, wherein the OA deployment request is used to request deployment of OA in the second computing node based on the OA deployment resource information. The computing power resource information required for deploying OA includes at least one of the following: the type of computing power resources required for deploying OA, the number of computing power resources required for deploying OA, the memory capacity required for deploying OA, and the storage capacity required for deploying OA, etc. The OA deployed in the second computing node can not only be used to orchestrate the first business to obtain S tasks and determine the SA for scheduling S tasks, but can also be used to orchestrate any business to obtain S tasks and determine the SA for scheduling S tasks, where S is a positive integer.

[0217] S205: The second computing node deploys the OA on the second computing node based on the OA deployment resource information.

[0218] In one possible implementation, the node agent (host agent) in the second computing node can receive the OA deployment request sent by the MA, determine whether the second computing node can provide the type of computing resources required to deploy the OA, determine whether the second computing node can provide the amount of computing resources required to deploy the OA, determine whether the second computing node can provide the memory capacity required to deploy the OA, and determine whether the second computing node can provide the storage capacity required to deploy the OA. If the above conditions are met, the node agent in the second computing node can deploy the OA on the second computing node based on the OA deployment resource information.

[0219] In the embodiment of the present application, the second computing node deploys the OA in different ways according to different OA deployment resource information, including the following two possible implementations:

[0220] In a first possible implementation, when the OA deployment resource information includes the OA executable code, the second computing node uses the OA executable code to deploy the OA in the second computing node.

[0221] In one possible implementation, when the OA deployment resource information includes the OA executable code, the second computing node can first determine whether the OA executable code is complete. If the OA executable code is incomplete, the OA is not deployed in the second computing node; if the OA executable code is complete, the OA executable code is used to deploy the OA in the second computing node, and an OA identifier is generated for the deployed OA. The OA identifier is a unique identifier and can be used to distinguish the deployed OA.

[0222] In a second possible implementation, when the OA deployment resource information includes an OA deployment template, the second computing node first obtains an OA image resource based on the OA deployment template, where the OA image resource is used to deploy OA. The OA system information in the OA deployment template is then used to update the image parameters in the OA image resource to obtain an updated OA image resource. The updated OA image resource is then used to deploy OA on the second computing node. The second computing node may determine and obtain the OA image resource based on the role information in the OA deployment template.

[0223] In one possible implementation, when the OA deployment resource information includes an OA deployment template, the second computing node can first determine whether the OA system information in the OA deployment template is complete. If the OA system information is incomplete, the OA will not be deployed in the second computing node; if the OA system information is complete, the OA mirror resource is obtained based on the role information in the OA deployment template, and then the OA system information in the OA deployment template is used to update the mirror parameters in the OA mirror resource to obtain the updated OA mirror resource. Then, the updated OA mirror resource is used to deploy the OA in the second computing node, and an OA identifier is generated for the deployed OA.

[0224] It should be understood that determining whether the OA system information is complete may include at least one of the following: determining whether role information, orchestration rules, first orchestration resource information, second orchestration resource information, orchestration memory information, and orchestration output format information are contained in the OA system information; determining whether a usable first orchestration resource can be obtained based on the first orchestration resource information in the OA system information; determining whether a usable second orchestration resource can be obtained based on the second orchestration resource information in the OA system information; and determining whether a usable orchestration memory can be obtained based on the orchestration memory information in the OA system information.

[0225] S206: The second computing node sends an OA deployment response to the MA. Correspondingly, the MA receives the OA deployment response sent by the second computing node.

[0226] In the embodiment of the present application, if the second computing node fails to deploy the OA, an OA deployment failure message is sent to the MA. After receiving the OA deployment failure message, the MA may generate a new OA deployment strategy.

[0227] If the second computing node successfully deploys the OA, it sends an OA deployment success message to the MA, where the OA deployment success message includes the identifier of the second computing node and the OA identifier of the OA deployed in the second computing node. After receiving the OA deployment success message, the MA adds the identifier of the second computing node and the OA identifier of the OA deployed in the second computing node as a set of records to the first OA deployment relationship. The identifier of the second computing node and the OA identifier of the OA deployed in the second computing node can be stored in the first OA deployment relationship in the form of a key-value pair.

[0228] S207: The MA sends an SA deployment request to the third computing node based on the identifier of the third computing node. Correspondingly, the third computing node receives the SA deployment request sent by the MA.

[0229] In an embodiment of the present application, the SA deployment request includes SA deployment resource information and computing resource information required for deploying the SA, wherein the SA deployment request is used to request deployment of the SA in the third computing node based on the SA deployment resource information. The computing resource information required for deploying the SA includes at least one of the following: the type of computing resource required for deploying the SA, the amount of computing resource required for deploying the SA, the memory capacity required for deploying the SA, and the storage capacity required for deploying the SA. The SA deployed in the third computing node is used to determine a fourth computing node that executes the received task.

[0230] S208: The third computing node deploys SA on the third computing node based on the SA deployment resource information.

[0231] In one possible implementation, the node agent in the third computing node can receive the SA deployment request sent by the MA, determine whether the third computing node can provide the type of computing power resources required to deploy SA, determine whether the third computing node can provide the quantity of computing power resources required to deploy SA, determine whether the third computing node can provide the memory capacity required to deploy SA, and determine whether the third computing node can provide the storage capacity required to deploy SA. If the above conditions are met, the node agent in the third computing node can deploy SA on the third computing node based on the SA deployment resource information.

[0232] In the embodiment of the present application, the third computing node deploys SA in different ways according to different SA deployment resource information, including the following two possible implementations:

[0233] In a first possible implementation, when the SA deployment resource information includes an SA executable code, the third computing node uses the SA executable code to deploy the SA in the third computing node.

[0234] In one possible implementation, when the SA deployment resource information includes the SA executable code, the third computing node can first determine whether the SA executable code is complete. If the SA executable code is incomplete, the SA is not deployed in the third computing node; if the SA executable code is complete, the SA executable code is used to deploy the SA in the third computing node, and an SA identifier is generated for the deployed SA. The SA identifier is a unique identifier and can be used to distinguish the deployed SA.

[0235] In a second possible implementation, when the SA deployment resource information includes an SA deployment template, the third computing node first obtains an SA image resource based on the SA deployment template, where the SA image resource is used to deploy SA. The SA system information in the SA deployment template is then used to update the image parameters in the SA image resource to obtain an updated SA image resource. The updated SA image resource is then used to deploy SA on the third computing node. The second computing node may determine and obtain the SA image resource based on the role information in the SA deployment template.

[0236] In one possible implementation, when the SA deployment resource information includes an SA deployment template, the third computing node can first determine whether the SA system information in the SA deployment template is complete. If the SA system information is incomplete, the SA is not deployed in the third computing node; if the SA system information is complete, the SA mirror resource is obtained based on the role information in the SA deployment template, and then the SA system information in the SA deployment template is used to update the mirror parameters in the SA mirror resource to obtain the updated SA mirror resource. Then, the updated SA mirror resource is used to deploy the SA in the third computing node, and an SA identifier is generated for the deployed SA.

[0237] It should be understood that determining whether the SA system information is complete may include at least one of the following: determining whether role information, scheduling rules, first scheduling resource information, second scheduling resource information, scheduling memory information, and scheduling output format information are contained in the SA system information; determining whether a usable first scheduling resource can be obtained based on the first scheduling resource information in the SA system information; determining whether a usable second scheduling resource can be obtained based on the second scheduling resource information in the SA system information; and determining whether a usable scheduling memory can be obtained based on the scheduling memory information in the SA system information.

[0238] S209: The third computing node sends an SA deployment response to the MA. Correspondingly, the MA receives the SA deployment response sent by the third computing node.

[0239] In the embodiment of the present application, if the third computing node fails to deploy the SA, an SA deployment failure message is sent to the MA. After receiving the SA deployment failure message, the MA may generate a new SA deployment strategy.

[0240] If the third computing node successfully deploys the SA, an SA deployment success message is sent to the MA, wherein the SA deployment success message includes the identifier of the third computing node and the SA identifier of the SA deployed in the third computing node. After receiving the SA deployment success message, the MA adds the identifier of the third computing node and the SA identifier of the SA deployed in the third computing node as a set of records to the first SA deployment relationship. The identifier of the third computing node and the SA identifier of the SA deployed in the third computing node can be stored in the first SA deployment relationship in the form of a key-value pair.

[0241] In the above method, MA can directly determine the OA deployment strategy and the SA deployment strategy based on the first network information and / or the first business information, and deploy OA in the second computing node based on the OA deployment strategy, and deploy SA in the third computing node based on the SA deployment strategy. The above OA and SA deployment process is completed independently by MA without the participation of developers, which can improve the deployment efficiency of OA and SA.

[0242] The OA deployed in the second computing node can orchestrate any business to obtain S tasks. The above OA orchestration process is completed independently by OA and does not rely on the experience of developers, which is conducive to improving business orchestration efficiency. At the same time, it can also reduce costs and shorten the business orchestration cycle.

[0243] The SA deployed in the third computing node can schedule the received tasks and determine the computing node to execute the received tasks. The process of SA scheduling tasks is completed independently by SA and does not rely on the experience of developers, which is conducive to improving task scheduling efficiency.

[0244] The following describes in detail how to deploy OA and SA in conjunction with specific embodiments. The deployment is interactively performed by MA, computing node A, computing node B, and computing node C in the first computing node, and may include the following steps as shown in FIG6 :

[0245] S601: The MA receives a first request, wherein the first request is used to request the generation of an OA and a SA.

[0246] S602: The MA obtains first network information and / or first service information of the first network based on the first request.

[0247] S603: The MA determines an OA deployment strategy and an SA deployment strategy based on the first network information and / or the first service information.

[0248] In the embodiment of the present application, the OA deployment strategy is as follows:

[0249] <Number of OAs to be deployed: 2>

[0250] <Identifier of the computing node where OA1 is to be deployed: Computing Node A>

[0251] Computing power resource information required for deploying OA1

[0252] OA1 deployment resource information (where OA1 deployment resource information includes the OA1 deployment template)

[0253] <Identifier of the computing node where OA2 is to be deployed: Computing Node C>

[0254] Computing power resource information required for deploying OA2

[0255] OA2 deployment resource information (where OA2 deployment resource information includes the OA2 executable code)

[0256] In the embodiment of the present application, the SA deployment strategy is as follows:

[0257] <Number of SAs to be deployed: 2>

[0258] <Identifier of the computing node where SA1 is to be deployed: Computing Node B>

[0259] Computing power resource information required for deploying SA1

[0260] SA1 deployment resource information (where SA1 deployment resource information includes the SA1 executable code)

[0261] <Identifier of the computing node where SA2 is to be deployed: Computing Node C>

[0262] Computing power resource information required for deploying SA2

[0263] SA2 deployment resource information (where SA2 deployment resource information includes the SA2 deployment template)

[0264] S604. MA sends an OA1 deployment request to computing node A. Correspondingly, computing node A receives the OA1 deployment request sent by MA.

[0265] In the embodiment of the present application, the OA1 deployment request includes the computing power resource information required for deploying OA1 and the OA1 deployment resource information. Among them, the computing power resource information required for deploying OA1 includes the type of computing power resources required for deploying OA1, the quantity of computing power resources required for deploying OA1, the memory capacity required for deploying OA1, and the storage capacity required for deploying OA1. The OA1 deployment resource information includes the OA1 deployment template.

[0266] S605. Computing node A deploys OA1 on computing node A based on the OA1 deployment resource information.

[0267] In an embodiment of the present application, the node agent in computing node A can receive an OA1 deployment request sent by MA. After determining that computing node A can provide the type of computing resources required to deploy OA1, the amount of computing resources required to deploy OA1, the memory capacity required to deploy OA1, and the storage capacity required to deploy OA1, the node agent in computing node A determines that the OA1 system information in the OA1 deployment template is complete, and based on the OA1 deployment template, obtains the OA1 mirror resource, and then uses the OA1 system information in the OA1 deployment template to update the mirror parameters in the OA1 mirror resource to obtain the updated OA1 mirror resource. Then, the updated OA1 mirror resource is used to deploy OA1 in computing node A.

[0268] After OA1 is successfully deployed, the node agent in computing node A can generate an OA identifier for the deployed OA1, namely OA-A-1, and add the identifier of computing node A (computing node A) and the OA identifier of the deployed OA1 (OA-A-1) to the OA1 deployment success message.

[0269] S606: Computing node A sends an OA1 deployment success message to the MA. Correspondingly, the MA receives the OA1 deployment success message sent by computing node A.

[0270] In one possible implementation, the MA receives the OA1 deployment success message and adds the identifier of the computing node A and the identifier of the deployed OA1 as a set of records to the first OA deployment relationship. The first OA deployment relationship is as follows:

[0271] <Compute Node A: OA-A-1>

[0272] S607: The MA sends an OA2 deployment request to the computing node C. Correspondingly, the computing node C receives the OA2 deployment request sent by the MA.

[0273] In an embodiment of the present application, the OA2 deployment request includes computing resource information required for deploying OA2 and OA2 deployment resource information. The computing resource information required for deploying OA2 includes the type of computing resources required for deploying OA2, the amount of computing resources required for deploying OA2, the memory capacity required for deploying OA2, and the storage capacity required for deploying OA2. The OA2 deployment resource information includes OA2 executable code.

[0274] S608 : The computing node C deploys OA2 on the computing node C based on the OA2 deployment resource information.

[0275] In an embodiment of the present application, the node agent in the computing node C can receive the OA2 deployment request sent by the MA. After determining that the computing node C can provide the computing power resource type required to deploy OA2, the computing power resource quantity required to deploy OA2, the memory capacity required to deploy OA2, and the storage capacity required to deploy OA2, the node agent in the computing node C determines that the OA2 executable code is incomplete, and then does not deploy OA2 in the computing node C. It also determines that the reason for the OA2 deployment failure is "OA2 executable code is incomplete", and adds "OA2 executable code is incomplete" to the OA2 deployment failure message.

[0276] S609: Computing node C sends an OA2 deployment failure message to the MA. Correspondingly, the MA receives the OA2 deployment failure message sent by computing node C.

[0277] S610 , the MA sends an SA1 deployment request to the computing node B. Correspondingly, the computing node B receives the SA1 deployment request sent by the MA.

[0278] In this embodiment of the present application, the SA1 deployment request includes computing resource information required for deploying SA1 and SA1 deployment resource information. The computing resource information required for deploying SA1 includes the type of computing resource required for deploying SA1, the amount of computing resource required for deploying SA1, the memory capacity required for deploying SA1, and the storage capacity required for deploying SA1. The SA1 deployment resource information includes the SA1 executable code.

[0279] S611 , computing node B deploys SA1 on computing node B based on SA1 deployment resource information.

[0280] In an embodiment of the present application, the node agent in the computing node B can receive the SA1 deployment request sent by the MA. After determining that the computing node B can provide the type of computing resources required to deploy SA1, the amount of computing resources required to deploy SA1, the memory capacity required to deploy SA1, and the storage capacity required to deploy SA1, the node agent in the computing node B determines that the SA1 executable code is complete and, based on the SA1 executable code, deploys SA1 in the computing node B.

[0281] After SA1 is successfully deployed, the node agent in computing node B can generate an SA identifier for the deployed SA1, namely SA-B-1, and add the identifier of computing node B (computing node B) and the SA identifier of the deployed SA1 (SA-B-1) to the SA1 deployment success message.

[0282] S612: Computing node B sends an SA1 deployment success message to MA. Correspondingly, MA receives the SA1 deployment success message sent by computing node B.

[0283] In one possible implementation, the MA receives the SA1 deployment success message and adds the identifier of the computing node B and the identifier of the deployed SA1 as a set of records to the first SA deployment relationship. The first SA deployment relationship is as follows:

[0284] <Compute Node B: SA-B-1>

[0285] S613, the MA sends an SA2 deployment request to the computing node C. Correspondingly, the computing node C receives the SA2 deployment request sent by the MA.

[0286] In an embodiment of the present application, the SA2 deployment request includes computing resource information required for deploying SA2 and SA2 deployment resource information. The computing resource information required for deploying SA2 includes the type of computing resource required for deploying SA2, the amount of computing resource required for deploying SA2, the memory capacity required for deploying SA2, and the storage capacity required for deploying SA2. The SA2 deployment resource information includes an SA2 deployment template.

[0287] S614 , computing node C deploys SA2 on computing node C based on the SA2 deployment resource information.

[0288] In an embodiment of the present application, the node agent in the computing node C can receive the SA2 deployment request sent by the MA. After determining that the computing node C can provide the computing power resource type required to deploy SA2, the number of computing power resources required to deploy SA2, the memory capacity required to deploy SA2, and the storage capacity required to deploy SA2, the node agent in the computing node C determines that the SA2 system information in the SA2 deployment template is incomplete, and then does not deploy SA2 in the computing node C, and determines that the reason for the SA2 deployment failure is "SA2 system information is incomplete", and adds "SA2 system information is incomplete" to the SA2 deployment failure message.

[0289] S615: Computing node C sends an SA2 deployment failure message to the MA. Correspondingly, the MA receives the SA2 deployment failure message sent by computing node C.

[0290] After executing steps S601 to S615, the first OA deployment relationship stored in the first computing node is as follows:

[0291] <Compute Node A: OA-A-1>

[0292] The first SA deployment relationship stored in the first computing node is as follows:

[0293] <Compute Node B: SA-B-1>

[0294] It should be understood that, in the above steps, S604-S606, S607-S609, S610-S612, and S613-S615, the execution order of these four steps is not particular.

[0295] Figure 7 shows a flow chart of the OA and SA deployment method provided in an embodiment of the present application, which is interactively executed by the MA in the first computing node, the second computing node, and the third computing node, wherein the first computing node can be any node in the first network, the second computing node can be any node in the first network, and the third computing node can be any node in the first network. The first computing node, the second computing node, and the third computing node can be the same computing node or different computing nodes. The OA and SA deployment method provided in an embodiment of the present application may include the following steps.

[0296] S701: The MA receives a first request, wherein the first request is used to request the generation of an OA.

[0297] S702: The MA obtains first network information and / or first service information of the first network based on the first request.

[0298] S703: The MA determines an OA deployment strategy based on the first network information and / or the first service information.

[0299] In a possible implementation, the MA inputs the first network information and / or the first business information into the first model to obtain the OA deployment strategy, wherein the first model has the ability to generate the OA deployment strategy.

[0300] S704: The MA sends an OA deployment request to the second computing node based on the identifier of the second computing node. Correspondingly, the second computing node receives the OA deployment request sent by the MA.

[0301] S705: The second computing node deploys the OA on the second computing node based on the OA deployment resource information.

[0302] S706: The second computing node sends an OA deployment response to the MA. Correspondingly, the MA receives the OA deployment response sent by the second computing node.

[0303] In the embodiment of the present application, if the second computing node fails to deploy the OA, an OA deployment failure message is sent to the MA. After receiving the OA deployment failure message, the MA may generate a new OA deployment strategy.

[0304] If the second computing node successfully deploys the OA, it sends an OA deployment success message to the MA, including the identifier of the second computing node and the OA identifier of the OA deployed on the second computing node. After receiving the OA deployment success message, the MA adds the identifier of the second computing node and the OA identifier of the OA deployed on the second computing node as a set of records to the first OA deployment relationship.

[0305] S707: The OA receives a third request, where the third request is used to request the generation of an SA.

[0306] In the embodiment of the present application, the third request can be sent from the developer to the OA via any computing node in the first network, or from any computing node in another network. The third request can be sent from a network device in the first network to the OA, or from a network device in another network, without limitation.

[0307] S708: The OA obtains first network information and / or first service information of the first network based on the third request.

[0308] S709: The OA determines an SA deployment strategy based on the first network information and / or the first service information.

[0309] In a possible implementation, the OA inputs the first network information and / or the first business information into the first model to obtain the SA deployment strategy, wherein the first model has the ability to generate the SA deployment strategy.

[0310] S710: The OA sends an SA deployment request to the third computing node based on the identifier of the third computing node. Correspondingly, the third computing node receives the SA deployment request sent by the OA.

[0311] S711: The third computing node deploys SA on the third computing node based on the SA deployment resource information.

[0312] S712: The third computing node sends an SA deployment response to the OA. Correspondingly, the OA receives the SA deployment response sent by the third computing node.

[0313] In the embodiment of the present application, if the third computing node fails to deploy the SA, an SA deployment failure message is sent to the OA. After receiving the SA deployment failure message, the OA may generate a new SA deployment strategy.

[0314] If the SA is successfully deployed on the third computing node, an SA deployment success message is sent to the OA, including the identifier of the third computing node and the SA identifier of the SA deployed on the third computing node. After receiving the SA deployment success message, the OA adds the identifier of the third computing node and the SA identifier of the SA deployed on the third computing node as a set of records to the first SA deployment relationship.

[0315] After the OA deploys the SA, the first SA deployment relationship may be synchronized to the MA in the first computing node and stored in the first computing node.

[0316] In the above method, the MA can directly determine the OA deployment strategy based on the first network information and / or the first business information, and deploy the OA in the second computing node based on the OA deployment strategy. The above OA deployment process is completed autonomously by the MA, without the involvement of developers, which can improve the efficiency of OA deployment. In addition, the OA deployed in the second computing node can orchestrate any business to obtain S tasks. The above OA business orchestration process is completed autonomously by the OA and does not rely on the experience of developers, which is conducive to improving business orchestration efficiency, while also reducing costs and shortening the business orchestration cycle.

[0317] After the OA is deployed, the OA can directly determine the SA deployment strategy based on the first network information and / or the first business information, and deploy the SA in the third computing node based on the SA deployment strategy. The above SA deployment process is completed autonomously by the OA without the involvement of developers, which can improve the deployment efficiency of SA. In addition, the SA deployed in the third computing node can schedule the received tasks and determine the computing node to execute the received tasks. The process of SA scheduling tasks is completed autonomously by the SA and does not rely on the experience of developers, which is conducive to improving task scheduling efficiency.

[0318] In the above method, since the MA is responsible for deploying the OA, and the deployed OA is responsible for deploying the SA, the above deployment scheme can reduce the load on the MA. In addition, if the MA fails, the SA can still be deployed in the first network through the OA, which can effectively ensure that the deployment of the SA is not affected.

[0319] The following describes in detail how to deploy OA and SA in conjunction with specific embodiments. The deployment is interactively performed by MA, computing node A, computing node B, and computing node C in the first computing node, and may include the following steps as shown in FIG8 :

[0320] S801: The MA receives a first request, wherein the first request is used to request the generation of an OA.

[0321] S802: The MA obtains first network information and / or first service information of the first network based on the first request.

[0322] S803. MA determines the OA deployment strategy based on the first network information and / or the first service information.

[0323] In the embodiment of this application, the OA deployment strategy is as follows:

[0324] <Number of OAs to be deployed: 1>

[0325] <Identifier of the computing node where OA1 is to be deployed: Computing Node A>

[0326] Computing power resource information required for deploying OA1

[0327] OA1 deployment resource information (where the OA1 deployment resource information includes the OA1 deployment template)

[0328] S804. MA sends an OA1 deployment request to Computing Node A. Correspondingly, Computing Node A receives the OA1 deployment request sent by MA.

[0329] In the embodiment of this application, the OA1 deployment request includes the computing power resource information required for deploying OA1 and the OA1 deployment resource information. Among them, the computing power resource information required for deploying OA1 includes the type of computing power resource required for deploying OA1, the quantity of computing power resource required for deploying OA1, the memory capacity required for deploying OA1, and the storage capacity required for deploying OA1. The OA1 deployment resource information includes the OA1 deployment template.

[0330] S805. Computing Node A deploys OA1 based on the OA1 deployment resource information in Computing Node A.

[0331] In the embodiment of this application, the node agent in Computing Node A can receive the OA1 deployment request sent by MA. After determining that Computing Node A can provide the type of computing power resource required for deploying OA1, the quantity of computing power resource required for deploying OA1, the memory capacity required for deploying OA1, and the storage capacity required for deploying OA1, the node agent in Computing Node A determines that the OA1 system information in the OA1 deployment template is complete, obtains the OA1 image resource based on the OA1 deployment template, then updates the image parameters in the OA1 image resource with the OA1 system information in the OA1 deployment template to obtain the updated OA1 image resource, and then deploys OA1 in Computing Node A using the updated OA1 image resource.

[0332] After the successful deployment of OA1, the node agent in Computing Node A can generate an OA identifier for the deployed OA1, that is, OA-A-1, and add the identifier of Computing Node A (Computing Node A) and the OA identifier of the deployed OA1 (OA-A-1) to the OA1 deployment success message.

[0333] S806, The computing node A sends an OA1 deployment success message to the MA. Correspondingly, the MA receives the OA1 deployment success message sent by the computing node A.

[0334] In a possible implementation, the MA receives the OA1 deployment success message, and takes the identifier of the computing node A and the identifier of the deployed OA1 as a set of records, and adds them to the first OA deployment relationship. The first OA deployment relationship is as follows:

[0335] <Computing node A: OA - A - 1>

[0336] S807, OA1 receives the third request. Among them, the first request is used to request the generation of SA.

[0337] S808, Based on the third request, OA1 obtains the first network information and / or the first service information of the first network.

[0338] S809, Based on the first network information and / or the first service information, OA1 determines the SA deployment strategy.

[0339] In the embodiment of the present application, the SA deployment strategy is as follows:

[0340] <Number of SAs to be deployed: 2>

[0341] <Identifier of the computing node where SA1 is to be deployed: Computing node B>

[0342] Computing power resource information required for deploying SA1

[0343] SA1 deployment resource information (where the SA1 deployment resource information includes the SA1 executable code)

[0344] <Identifier of the computing node where SA2 is to be deployed: Computing node C>

[0345] Computing power resource information required for deploying SA2

[0346] SA2 deployment resource information (where the SA2 deployment resource information includes the SA2 deployment template)

[0347] S810, OA1 sends a SA1 deployment request to the computing node B. Correspondingly, the computing node B receives the SA1 deployment request sent by OA1.

[0348] In the embodiment of the present application, the SA1 deployment request includes the computing power resource information required for deploying SA1 and the SA1 deployment resource information. Among them, the computing power resource information required for deploying SA1 includes the type of computing power resources required for deploying SA1, the quantity of computing power resources required for deploying SA1, the memory capacity required for deploying SA1, and the storage capacity required for deploying SA1. The SA1 deployment resource information includes the SA1 executable code.

[0349] S811 , computing node B deploys SA1 on computing node B based on SA1 deployment resource information.

[0350] In an embodiment of the present application, the node agent in the computing node B can receive the SA1 deployment request sent by the MA. After determining that the computing node B can provide the type of computing resources required to deploy SA1, the amount of computing resources required to deploy SA1, the memory capacity required to deploy SA1, and the storage capacity required to deploy SA1, the node agent in the computing node B determines that the SA1 executable code is complete and, based on the SA1 executable code, deploys SA1 in the computing node B.

[0351] After SA1 is successfully deployed, the node agent in computing node B can generate an SA identifier for the deployed SA1, namely SA-B-1, and add the identifier of computing node B (computing node B) and the SA identifier of the deployed SA1 (SA-B-1) to the SA1 deployment success message.

[0352] S812: Computing node B sends an SA1 deployment success message to OA1. Correspondingly, OA1 receives the SA1 deployment success message sent by computing node B.

[0353] In one possible implementation, OA1 receives the SA1 deployment success message and adds the identifier of the computing node B and the identifier of the deployed SA1 as a set of records to the first SA deployment relationship. The first SA deployment relationship is as follows:

[0354] <Compute Node B: SA-B-1>

[0355] S813, OA1 sends an SA2 deployment request to computing node C. Correspondingly, computing node C receives the SA2 deployment request sent by OA1.

[0356] In an embodiment of the present application, the SA2 deployment request includes computing resource information required for deploying SA2 and SA2 deployment resource information. The computing resource information required for deploying SA2 includes the type of computing resource required for deploying SA2, the amount of computing resource required for deploying SA2, the memory capacity required for deploying SA2, and the storage capacity required for deploying SA2. The SA2 deployment resource information includes an SA2 deployment template.

[0357] S814 , computing node C deploys SA2 on computing node C based on the SA2 deployment resource information.

[0358] In an embodiment of the present application, the node agent in the computing node C can receive the SA2 deployment request sent by the MA. After determining that the computing node C can provide the computing power resource type required to deploy SA2, the number of computing power resources required to deploy SA2, the memory capacity required to deploy SA2, and the storage capacity required to deploy SA2, the node agent in the computing node C determines that the SA2 system information in the SA2 deployment template is incomplete, and then does not deploy SA2 in the computing node C, and determines that the reason for the SA2 deployment failure is "SA2 system information is incomplete", and adds "SA2 system information is incomplete" to the SA2 deployment failure message.

[0359] S815: Computing node C sends an SA2 deployment failure message to OA1. Correspondingly, OA1 receives the SA2 deployment failure message sent by computing node C.

[0360] After OA1 completes deploying the SA, the first SA deployment relationship may be synchronized to the MA in the first computing node and stored in the first computing node.

[0361] After executing steps S801 to S815, the first OA deployment relationship stored in the first computing node is as follows:

[0362] <Compute Node A: OA-A-1>

[0363] The first SA deployment relationship stored in the first computing node is as follows:

[0364] <Compute Node B: SA-B-1>

[0365] It should be understood that, in the above steps, S810 - S812 and S813 - S815 are executed in no particular order.

[0366] Figure 9 shows a flow chart of the OA and SA adjustment method provided in an embodiment of the present application, which is interactively executed by the MA in the first computing node, the second computing node, and the third computing node, wherein the first computing node can be any node in the first network, the second computing node can be any node in the first network, and the third computing node can be any node in the first network. The first computing node, the second computing node, and the third computing node can be the same computing node or different computing nodes. The OA and SA adjustment method provided in an embodiment of the present application may include the following steps.

[0367] S901: MA receives a second request, wherein the second request is used to request adjustment of OA and SA.

[0368] In the embodiment of the present application, the second request can be sent by the developer to the MA via any computing node in the first network or any computing node in another network. The second request can be sent by a network device in the first network or a network device in another network, without limitation.

[0369] S902: The MA obtains second network information and / or second service information of the first network based on the second request.

[0370] In the embodiment of the present application, the second network information is the adjusted first network information, and the second service information is the adjusted first service information.

[0371] S903 , the MA determines an OA adjustment policy and an SA adjustment policy based on the second network information and / or the second service information, the first OA deployment relationship, and the first SA deployment relationship.

[0372] In the embodiment of the present application, the first model has the ability to generate an OA adjustment strategy and an SA adjustment strategy. The OA adjustment strategy and the SA adjustment strategy can be determined by the following two possible implementation methods:

[0373] In a first possible implementation, the MA inputs the second network information and / or the second service information, and the first OA deployment relationship into the first model to obtain an OA adjustment policy. The MA inputs the second network information and / or the second service information, and the first SA deployment relationship into the first model to obtain an SA adjustment policy.

[0374] In a second possible implementation, the MA inputs the second network information and / or the second service information, as well as the first OA deployment relationship and the first SA deployment relationship into the first model to obtain the OA adjustment policy and the SA adjustment policy.

[0375] In one possible implementation, the OA adjustment strategy may include the number of OAs to be adjusted, the identifier of the computing node corresponding to each OA to be adjusted, and OA adjustment information corresponding to each OA to be adjusted, wherein the OA adjustment information is used to indicate the adjustment of the OA.

[0376] In one possible implementation, the OA adjustment strategy may include the number of OAs to be adjusted (m), the identifiers of m second computing nodes, and m OA adjustment information, wherein the identifiers of the m second computing nodes correspond one-to-one to the m OA adjustment information, m is a positive integer, and m is less than or equal to M.

[0377] For ease of description, the steps in FIG9 are discussed using the OA adjustment strategy including the number of OAs to be adjusted (1), the identifier of the second computing node (1), and the OA adjustment information (1) as an example.

[0378] In a possible implementation, the OA adjustment policy includes an OA deletion policy or an OA addition policy.

[0379] In the case where the OA adjustment policy is the OA deletion policy, the OA deletion policy may include the number of OAs to be deleted, the identifier of the computing node corresponding to each OA to be deleted, and the OA deletion information corresponding to each OA to be deleted, wherein the OA deletion information is used to indicate the deletion of the OA, and the OA deletion information includes the identifier of the OA to be deleted, wherein the identifier of the OA to be deleted may be obtained from the first OA deployment relationship.

[0380] In the case where the OA adjustment strategy is an OA addition strategy, the OA addition strategy may include the number of OAs to be added, the identifiers of the computing nodes to be deployed for each OA to be added, and the OA addition information corresponding to each OA to be added, wherein the OA addition information is used to indicate that the OA is to be added, and the OA addition information includes the computing power resource information required to deploy the OA to be added and the OA deployment resource information corresponding to the OA to be added. The computing power resource information required to deploy the OA to be added includes at least one of the following: the type of computing power resources required to deploy the OA to be added, the number of computing power resources required to deploy the OA to be added, the memory capacity required to deploy the OA to be added, and the storage capacity required to deploy the OA to be added. The OA deployment resource information includes the OA deployment template or the OA executable code, which has been described in detail above and will not be repeated here.

[0381] In a possible implementation, the SA adjustment strategy may include the number of SAs to be adjusted, the identifier of the computing node corresponding to each SA to be adjusted, and SA adjustment information corresponding to each SA to be adjusted, wherein the SA adjustment information is used to indicate adjustment of the SA.

[0382] In one possible implementation, the SA adjustment strategy may include the number of SAs to be adjusted (n), the identifiers of n third computing nodes, and n SA deployment resource information, wherein the identifiers of the n third computing nodes correspond one-to-one to the n SA adjustment information, n is a positive integer, and n is less than or equal to N.

[0383] For ease of description, the steps in FIG9 are discussed by taking the SA adjustment strategy including the number of SAs to be adjusted (1), the identifier of the third computing node (1), and the SA adjustment information (1) as an example.

[0384] In a possible implementation, the SA adjustment policy includes an SA deletion policy or an SA addition policy.

[0385] In the case where the SA adjustment policy is the SA deletion policy, the SA deletion policy may include the number of SAs to be deleted, the identifier of the computing node corresponding to each SA to be deleted, and the SA deletion information corresponding to each SA to be deleted, wherein the SA deletion information is used to indicate the deletion of the SA, and the SA deletion information includes the identifier of the SA to be deleted, wherein the identifier of the SA to be deleted may be obtained from the first SA deployment relationship.

[0386] In the case where the SA adjustment policy is an SA addition policy, the SA addition policy may include the number of SAs to be added, the identifiers of the computing nodes to be deployed for each SA to be added, and the SA addition information corresponding to each SA to be added, wherein the SA addition information is used to indicate that the SA is to be added, and the SA addition information includes the computing power resource information required to deploy the SA to be added and the SA deployment resource information corresponding to the SA to be added. The computing power resource information required to deploy the SA to be added includes at least one of the following: the type of computing power resources required to deploy the SA to be added, the number of computing power resources required to deploy the SA to be added, the memory capacity required to deploy the SA to be added, and the storage capacity required to deploy the SA to be added. The SA deployment resource information includes an SA deployment template or SA executable code, which has been described in detail above and will not be repeated here.

[0387] S904: The MA sends an OA adjustment request to the second computing node based on the identifier of the second computing node. Correspondingly, the second computing node receives the OA adjustment request sent by the MA.

[0388] In an embodiment of the present application, the OA adjustment request includes OA adjustment information, wherein the OA adjustment request is used to request adjustment of the deployed OA in the second computing node based on the OA adjustment information.

[0389] In one possible implementation, the OA adjustment request includes two types, namely, an OA deletion request and an OA addition request.

[0390] When the OA adjustment request is an OA deletion request, the OA adjustment information is OA deletion information, wherein the OA deletion information includes an identifier of the OA to be deleted. The OA deletion request is used to request the second computing node to delete the corresponding OA based on the OA deletion information.

[0391] If the OA adjustment request is an OA addition request, the OA adjustment information is OA addition information, where the OA addition information includes computing resource information required to deploy the OA to be added and OA deployment resource information corresponding to the OA to be added. The OA addition request is used to request that an OA be added to the second computing node based on the OA deployment resource information corresponding to the OA to be added.

[0392] S905: The second computing node adjusts the OA deployed in the second computing node based on the OA adjustment information.

[0393] In a possible implementation, the node agent in the second computing node may receive the OA adjustment request and adjust the OA deployed in the second computing node based on the OA adjustment information.

[0394] In the case that the OA adjustment information is OA deletion information, the node agent in the second computing node may delete the corresponding OA according to the identifier of the OA to be deleted in the OA deletion information.

[0395] In the case where the OA adjustment information is new OA information, the node agent in the second computing node can determine whether the second computing node can provide the type of computing resources required for deploying the OA to be added, determine whether the second computing node can provide the amount of computing resources required for deploying the OA to be added, determine whether the second computing node can provide the memory capacity required for deploying the OA to be added, and determine whether the second computing node can provide the storage capacity required for deploying the OA to be added. If the above conditions are met, the node agent in the second computing node can add the OA to the second computing node based on the OA deployment resource information corresponding to the OA to be added.

[0396] S906: The second computing node sends an OA adjustment response to the MA. Correspondingly, the MA receives the OA adjustment response sent by the second computing node.

[0397] If the second computing node fails to adjust the deployed OA, it sends an OA adjustment failure message to the MA.

[0398] If the second computing node successfully adjusts the deployed OA, it sends an OA adjustment success message to the MA, where the OA adjustment success message includes the identifier of the second computing node and the OA identifier of the adjusted OA in the second computing node. After receiving the OA adjustment success message, the MA modifies the first OA deployment relationship using the identifier of the second computing node and the OA identifier of the adjusted OA in the second computing node.

[0399] When the second computing node deletes the deployed OA, if the second computing node fails to delete the deployed OA, the second computing node sends an OA deletion failure message to the MA. If the second computing node successfully deletes the deployed OA, the second computing node sends an OA deletion success message to the MA, where the OA deletion success message includes the identifier of the second computing node and the OA identifier of the deleted OA. Based on the OA deletion success message, the MA deletes the set of records including the identifier of the second computing node and the OA identifier of the deleted OA from the first OA deployment relationship.

[0400] When adding an OA to the second computing node, if the addition of the OA fails, the MA sends an OA addition failure message. If the addition of the OA succeeds, the MA sends an OA addition success message, including the identifier of the second computing node and the OA identifier of the newly added OA. Based on the OA addition success message, the MA adds the identifier of the second computing node and the OA identifier of the newly added OA as a set of records to the first OA deployment relationship.

[0401] S907: The MA sends an SA adjustment request to the third computing node based on the identifier of the third computing node. Correspondingly, the third computing node receives the SA adjustment request sent by the MA.

[0402] In an embodiment of the present application, the SA adjustment request includes SA adjustment information, wherein the SA adjustment request is used to request adjustment of the deployed SA in the third computing node based on the SA adjustment information.

[0403] In one possible implementation, SA adjustment requests include two types: SA deletion requests and SA addition requests. When the SA adjustment request is an SA deletion request, the SA adjustment information is SA deletion information, where the SA deletion information includes an identifier of the SA to be deleted. The SA deletion request is used to request that the third computing node delete the corresponding SA based on the SA deletion information.

[0404] If the SA adjustment request is an SA addition request, the SA adjustment information is SA addition information, where the SA addition information includes computing resource information required to deploy the SA to be added and SA deployment resource information corresponding to the SA to be added. The SA addition request is used to request the addition of a new SA in the third computing node based on the SA deployment resource information corresponding to the SA to be added.

[0405] S908: The third computing node adjusts the SA deployed in the third computing node based on the SA adjustment information.

[0406] In a possible implementation, the node agent in the third computing node may receive the SA adjustment request and adjust the deployed SA in the third computing node based on the SA adjustment information.

[0407] In the case that the SA adjustment information is SA deletion information, the node agent in the third computing node may delete the corresponding SA according to the identifier of the SA to be deleted in the SA deletion information.

[0408] When the SA adjustment information is SA new addition information, the node agent in the third computing node can determine whether the third computing node can provide the type of computing power resources required for deploying the SA to be added, determine whether the third computing node can provide the quantity of computing power resources required for deploying the SA to be added, determine whether the third computing node can provide the memory capacity required for deploying the SA to be added, and determine whether the third computing node can provide the storage capacity required for deploying the SA to be added. If the above conditions are met, the node agent in the third computing node can add the SA to the third computing node based on the SA deployment resource information corresponding to the SA to be added.

[0409] S909: The third computing node sends an SA adjustment response to the MA. Correspondingly, the MA receives the SA adjustment response sent by the third computing node.

[0410] If the third computing node fails to adjust the deployed SA, it sends an SA adjustment failure message to the MA.

[0411] If the third computing node successfully adjusts the deployed SA, it sends an SA adjustment success message to the MA, where the SA adjustment success message includes the identifier of the third computing node and the SA identifier of the adjusted SA in the third computing node. After receiving the SA adjustment success message, the MA modifies the first SA deployment relationship using the identifier of the third computing node and the SA identifier of the adjusted SA in the third computing node.

[0412] When the third computing node deletes a deployed SA, if the third computing node fails to delete the deployed SA, the third computing node sends an SA deletion failure message to the MA. If the third computing node successfully deletes the deployed SA, the third computing node sends an SA deletion success message to the MA, where the SA deletion success message includes the identifier of the third computing node and the SA identifier of the deleted SA. Based on the SA deletion success message, the MA deletes the set of records including the identifier of the third computing node and the SA identifier of the deleted SA from the first SA deployment relationship.

[0413] When the third computing node adds a new SA, if the third computing node fails to add the SA, the MA sends an SA addition failure message. If the third computing node successfully adds the SA, the MA sends an SA addition success message, which includes the identifier of the third computing node and the SA identifier of the newly added SA. Based on the SA addition success message, the MA adds the identifier of the third computing node and the SA identifier of the newly added SA as a set of records to the first SA deployment relationship.

[0414] In the above method, when the first network information and / or the first service information changes, the MA can directly determine the OA adjustment policy and the SA adjustment policy based on the changed first network information and / or the first service information, as well as the first OA deployment relationship and the first SA deployment relationship. The MA then adjusts the OA already deployed in the second computing node according to the OA adjustment policy and adjusts the SA already deployed in the third computing node according to the SA deployment policy. The above adjustment process of the OA and the SA is completed independently by the MA without the participation of developers, which can improve the adjustment efficiency of the OA and the SA.

[0415] In addition, the OA adjustment policy and the SA adjustment policy can be adjusted as the first network information and / or the first service information changes, which can effectively avoid the problem of unreasonable numbers of deployed OAs and SAs.

[0416] The following describes in detail how to adjust the OA and the SA in combination with specific embodiments. The interaction and execution by the MA, computing node A, computing node B, and computing node C in the first computing node may include the following steps shown in FIG. 10:

[0417] The first OA deployment relationship stored in the first computing node is as follows:

[0418] <Computing node A: OA-A-1>

[0419] The first SA deployment relationship stored in the first computing node is as follows:

[0420] <Computing node B: SA-B-1>

[0421] S1001. The MA receives a second request. The second request is used to request the adjustment of the OA and the SA.

[0422] S1002. Based on the second request, the MA obtains the second network information and / or the second service information of the first network.

[0423] S1003. Based on the second network information and / or the second service information, as well as the first OA deployment relationship and the first SA deployment relationship, the MA determines the OA adjustment policy and the SA adjustment policy.

[0424] In the embodiment of the present application, the OA adjustment policy is as follows:

[0425] <Number of OAs to be deleted: 1>

[0426] <Identifier of the computing node corresponding to OA1: Computing node A>

[0427] OA1 deletion information (where the OA1 deletion information includes the identifier OA-A-1 of the OA1 to be deleted)

[0428] <Number of OAs to be added: 1>

[0429] <Identity of the computing node where OA2 is to be deployed: Computing node C>

[0430] OA2 new information (where OA2 new information includes computing power resource information required for deploying OA2 and OA2 deployment template)

[0431] In the embodiment of the present application, the SA adjustment strategy is as follows:

[0432] <Number of SAs to be deleted: 1>

[0433] <Identity of the computing node corresponding to SA1: Computing node B>

[0434] SA1 deletion information (where SA1 deletion information includes the identity of the SA to be deleted, SA-B-1)

[0435] <Number of SAs to be newly added: 1>

[0436] <Identity of the computing node where SA2 is to be deployed: Computing node C>

[0437] SA2 new information (where SA2 new information includes computing power resource information required for deploying SA2 and SA2 deployment template)

[0438] S1004, MA sends an OA1 deletion request to computing node A. Correspondingly, computing node A receives the OA1 deletion request sent by MA.

[0439] In the embodiment of the present application, the OA1 deletion request includes OA1 deletion information. Among them, OA1 deletion information includes the identity of the OA1 to be deleted (OA-A-1).

[0440] S1005, computing node A deletes OA1 in computing node A based on the identity of the OA1 to be deleted in the OA1 deletion information.

[0441] In the embodiment of the present application, computing node A fails to delete OA1 and generates an OA1 deletion failure message.

[0442] S1006, computing node A sends an OA1 deletion failure message to MA. Correspondingly, MA receives the OA1 deletion failure message sent by computing node A.

[0443] S1007, MA sends an OA2 addition request to computing node C. Correspondingly, computing node C receives the OA2 addition request sent by MA.

[0444] In an embodiment of the present application, the OA2 new information in the OA2 new addition request includes computing resource information required for deploying OA2. The computing resource information required for deploying OA2 includes the type of computing resources required for deploying OA2, the amount of computing resources required for deploying OA2, the memory capacity required for deploying OA2, and the storage capacity required for deploying OA2.

[0445] S1008 , computing node C adds OA2 in computing node C based on the new information of OA2.

[0446] In an embodiment of the present application, the node agent in the computing node C can receive a new OA2 request. After determining that the computing node C can provide the computing power resource type required for deploying OA2, the number of computing power resources required for deploying OA2, the memory capacity required for deploying OA2, and the storage capacity required for deploying OA2, the node agent in the computing node C determines that the OA2 system information in the OA2 deployment template is complete, and based on the OA2 deployment template, obtains the OA2 mirror resource, and then uses the OA2 system information in the OA2 deployment template to update the mirror parameters in the OA2 mirror resource to obtain the updated OA2 mirror resource. Then, the updated OA2 mirror resource is used to deploy OA2 in the computing node C.

[0447] After OA2 is successfully deployed, the node agent in computing node C can generate an OA identifier for the newly added OA2, namely OA-C-1, and add the identifier of computing node C (computing node C) and the OA identifier of the newly added OA2 (OA-C-1) to the OA2 addition success message.

[0448] S1009: Computing node C sends an OA2 addition success message to the MA. Correspondingly, the MA receives the OA2 addition success message sent by computing node C.

[0449] In the embodiment of the present application, the MA receives the OA2 addition success message and adds the identifier of the computing node C and the newly added OA identifier of the OA2 to the first OA deployment relationship. The first OA deployment relationship is as follows:

[0450] <Compute Node A: OA-A-1>

[0451] <Compute Node C: OA-C-1>

[0452] S1010, the MA sends an SA1 deletion request to the computing node B. Correspondingly, the computing node B receives the SA1 deletion request sent by the MA.

[0453] In the embodiment of the present application, the SA1 deletion request includes SA1 deletion information, wherein the SA1 deletion information includes the identifier (SA-B-1) of the SA1 to be deleted.

[0454] S1011 , computing node B deletes SA1 in computing node B based on the identifier of the SA1 to be deleted in the SA1 deletion information.

[0455] In the embodiment of the present application, the computing node B fails to delete SA1 and generates an SA1 deletion failure message.

[0456] S1012: Computing node A sends an SA1 deletion failure message to MA. Correspondingly, MA receives the SA1 deletion failure message sent by computing node A.

[0457] S1013 , the MA sends an SA2 adding request to the computing node C. Correspondingly, the computing node C receives the SA2 adding request sent by the MA.

[0458] In the embodiment of the present application, the SA2 new information in the SA2 new request includes the computing resource information required for deploying SA2. The computing resource information required for deploying SA2 includes the type of computing resource required for deploying SA2, the amount of computing resource required for deploying SA2, the memory capacity required for deploying SA2, and the storage capacity required for deploying SA2.

[0459] S1014 , computing node C adds SA2 in computing node C based on the new information of SA2.

[0460] In an embodiment of the present application, the node agent in the computing node C can receive a new SA2 request. After determining that the computing node C can provide the type of computing power resources required to deploy SA2, the amount of computing power resources required to deploy SA2, the memory capacity required to deploy SA2, and the storage capacity required to deploy SA2, the node agent in the computing node C determines that the SA2 system information in the SA2 deployment template is complete, and based on the SA2 deployment template, obtains the SA2 image resource, and then uses the SA2 system information in the SA2 deployment template to update the image parameters in the SA2 image resource to obtain the updated SA2 image resource. Then, the updated SA2 image resource is used to deploy SA2 in the computing node C.

[0461] After SA2 is successfully deployed, the node agent in computing node C can generate an SA identifier for the newly added SA2, namely SA-C-1, and add the identifier of computing node C (computing node C) and the SA identifier of the newly added SA2 (SA-C-1) to the SA2 addition success message.

[0462] S1015: Computing node C sends an SA2 addition success message to the MA. Correspondingly, the MA receives the SA2 addition success message sent by computing node C.

[0463] In the embodiment of the present application, the MA receives the SA2 addition success message and adds the identifier of the computing node C and the newly added SA identifier of SA2 to the first SA deployment relationship. The first SA deployment relationship is as follows:

[0464] <Compute Node B: SA-B-1>

[0465] <Compute Node C: SA-C-1>

[0466] After executing steps S1001 to S1015, the first OA deployment relationship stored in the first computing node is as follows:

[0467] <Compute Node A: OA-A-1>

[0468] <Compute Node C: OA-C-1>

[0469] The first SA deployment relationship stored in the first computing node is as follows:

[0470] <Compute Node B: SA-B-1>

[0471] <Compute Node C: SA-C-1>

[0472] It should be understood that, in the above steps, S1004-S1006, S1007-S1009, S1010-S1012, and S1013-S1015, the execution order of these four steps is not particular.

[0473] Figure 11 shows a flow chart of the OA and SA adjustment method provided in an embodiment of the present application, which is interactively executed by the MA in the first computing node, the second computing node, and the third computing node, wherein the first computing node can be any node in the first network, the second computing node can be any node in the first network, and the third computing node can be any node in the first network. The first computing node, the second computing node, and the third computing node can be the same computing node or different computing nodes. The OA and SA adjustment method provided in an embodiment of the present application may include the following steps.

[0474] S1101: MA receives a second request, wherein the second request is for requesting adjustment of OA.

[0475] S1102: The MA obtains second network information and / or second service information of the first network based on the second request.

[0476] S1103: The MA determines an OA adjustment strategy based on the second network information and / or the second service information and the first OA deployment relationship.

[0477] In an embodiment of the present application, the MA inputs the second network information and / or the second service information and the first OA deployment relationship into the first model to obtain the OA adjustment policy, wherein the first model has the ability to generate the OA adjustment policy.

[0478] S1104: The MA sends an OA adjustment request to the second computing node based on the identifier of the second computing node. Correspondingly, the second computing node receives the OA adjustment request sent by the MA.

[0479] S1105: The second computing node adjusts the OA deployed in the second computing node based on the OA adjustment information.

[0480] S1106: The second computing node sends an OA adjustment response to the MA. Correspondingly, the MA receives the OA adjustment response sent by the second computing node.

[0481] S1107: The OA receives a fourth request, wherein the fourth request is used to request adjustment of the SA.

[0482] In the embodiment of the present application, the fourth request can be sent by the developer to the OA via any computing node in the first network, or by the developer to the OA via any computing node in another network. The fourth request can be sent by a network device in the first network to the OA, or by a network device in another network to the OA, without limitation.

[0483] S1108: The OA obtains second network information and / or second service information of the first network based on the fourth request.

[0484] S1109: The OA determines an SA adjustment strategy based on the second network information and / or the second service information and the first SA deployment relationship.

[0485] In an embodiment of the present application, the OA may obtain the first SA deployment relationship from the first computing node. The OA inputs the second network information and / or the second service information, as well as the first SA deployment relationship, into the first model to obtain the SA adjustment policy. The first model is capable of generating the SA adjustment policy.

[0486] S1110: The OA sends an SA adjustment request to the third computing node based on the identifier of the third computing node. Correspondingly, the third computing node receives the SA adjustment request sent by the OA.

[0487] S1111: The third computing node adjusts the SA deployed in the third computing node based on the SA adjustment information.

[0488] S1112: The third computing node sends an SA adjustment response to the OA. Correspondingly, the OA receives the SA adjustment response sent by the third computing node.

[0489] If the third computing node fails to adjust the deployed SA, it sends an SA adjustment failure message to the OA.

[0490] If the third computing node successfully adjusts the deployed SA, it sends an SA adjustment success message to the OA, where the SA adjustment success message includes the identifier of the third computing node and the SA identifier of the adjusted SA in the third computing node. After receiving the SA adjustment success message, the OA modifies the first SA deployment relationship using the identifier of the third computing node and the SA identifier of the adjusted SA in the third computing node.

[0491] After the OA completes adjusting the SA, the modified first SA deployment relationship may be synchronized to the MA in the first computing node to replace the first SA deployment relationship stored in the first computing node.

[0492] In the case where the third computing node deletes the deployed SA, if the third computing node fails to delete the deployed SA, an SA deletion failure message is sent to the OA; if the third computing node successfully deletes the deployed SA, an SA deletion success message is sent to the OA, wherein the SA deletion success message includes the identifier of the third computing node and the SA identifier of the deleted SA. Based on the SA deletion success message, the OA deletes the set of records including the identifier of the third computing node and the SA identifier of the deleted SA from the first SA deployment relationship, and synchronizes the first SA deployment relationship after deleting this set of records to the MA in the first computing node, replacing the first SA deployment relationship stored in the first computing node.

[0493] In the case of adding a new SA to the third computing node, if the third computing node fails to add the SA, an SA addition failure message is sent to the OA; if the third computing node successfully adds the SA, an SA addition success message is sent to the OA, wherein the SA addition success message includes the identifier of the third computing node and the SA identifier of the newly added SA. Based on the SA addition success message, the OA adds the identifier of the third computing node and the SA identifier of the newly added SA as a set of records to the first SA deployment relationship, and synchronizes the first SA deployment relationship with this set of records to the MA in the first computing node, replacing the first SA deployment relationship stored in the first computing node.

[0494] In the above method, when the first network information and / or the first business information changes, the MA can directly determine the OA adjustment policy based on the changed first network information and / or the first business information, as well as the first OA deployment relationship. The MA then adjusts the deployed OA in the second computing node based on the OA adjustment policy. The above OA adjustment process is completed autonomously by the MA without the involvement of developers, which can improve the efficiency of OA adjustment. In addition, the OA adjustment policy can be adjusted as the first network information and / or the first business information changes, which can effectively avoid the problem of an unreasonable number of deployed OAs.

[0495] When the first network information and / or the first business information changes, the OA can directly determine the SA adjustment strategy based on the changed first network information and / or the first business information, as well as the first SA deployment relationship. The OA then adjusts the deployed SA in the second computing node based on the SA adjustment strategy. The above SA adjustment process is completed autonomously by the OA without the involvement of developers, which can improve the efficiency of SA adjustment. In addition, the SA adjustment strategy can be adjusted as the first network information and / or the first business information changes, which can effectively avoid the problem of an unreasonable number of deployed SAs.

[0496] In this approach, since the MA is responsible for adjusting the deployed OA, and the OA is responsible for adjusting the deployed SA, this adjustment scheme can reduce the MA's load. Furthermore, if the MA fails, the OA can still adjust the deployed SA, effectively ensuring that the SA's adjustment is not affected.

[0497] The following describes in detail how to adjust OA and SA in conjunction with a specific embodiment. The adjustment is interactively performed by the MA, computing node A, computing node B, and computing node C in the first computing node, and may include the following steps as shown in FIG12 :

[0498] The first OA deployment relationship stored in the first computing node is as follows:

[0499] <Compute Node A: OA-A-1>

[0500] The first SA deployment relationship stored in the first computing node is as follows:

[0501] <Compute Node B: SA-B-1>

[0502] S1201: MA receives a second request, wherein the second request is for requesting adjustment of OA.

[0503] S1202: The MA obtains second network information and / or second service information of the first network based on the second request.

[0504] S1203: The MA determines an OA adjustment strategy based on the second network information and / or the second service information and the first OA deployment relationship.

[0505] In the embodiment of the present application, the OA adjustment strategy is as follows:

[0506] <Number of OAs to be deleted: 1>

[0507] <Identifier of the computing node corresponding to OA1: Computing Node A>

[0508] OA1 deletion information (where the OA1 deletion information includes the identifier OA-A-1 of the OA1 to be deleted)

[0509] <Number of OAs to be added: 1>

[0510] <Identifier of the computing node where OA2 is to be deployed: Computing Node C>

[0511] OA2 addition information (where the OA2 addition information includes the computing power resource information required for deploying OA2 and the OA2 deployment template)

[0512] S1204, MA sends an OA1 deletion request to Computing Node A. Correspondingly, Computing Node A receives the OA1 deletion request sent by MA.

[0513] In the embodiment of the present application, the OA1 deletion request includes OA1 deletion information. Among them, the OA1 deletion information includes the identifier (OA-A-1) of the OA1 to be deleted.

[0514] · S1205, Computing Node A deletes OA1 in Computing Node A based on the identifier of the OA1 to be deleted in the OA1 deletion information.<q

[0515] In the embodiment of the present application, Computing Node A fails to delete OA1 and generates an OA1 deletion failure message.

[0516] S1206, Computing Node A sends an OA1 deletion failure message to MA. Correspondingly, MA receives the OA1 deletion failure message sent by Computing Node A.

[0517] S1207, MA sends an OA2 addition request to Computing Node C. Correspondingly, Computing Node C receives the OA2 addition request sent by MA.

[0518] In the embodiment of the present application, the OA2 addition request includes OA2 addition information. Among them, the OA2 addition information includes the computing power resource information required for deploying OA2. Among them, the computing power resource information required for deploying OA2 includes the type of computing power resource required for deploying OA2, the quantity of computing power resource required for deploying OA2, the memory capacity required for deploying OA2, and the storage capacity required for deploying OA2.

[0519] S1208, Computing Node C adds OA2 in Computing Node C based on the OA2 addition information.

[0520] In an embodiment of the present application, the node agent in the computing node C can receive a new OA2 request. After determining that the computing node C can provide the computing power resource type required for deploying OA2, the number of computing power resources required for deploying OA2, the memory capacity required for deploying OA2, and the storage capacity required for deploying OA2, the node agent in the computing node C determines that the OA2 system information in the OA2 deployment template is complete, and based on the OA2 deployment template, obtains the OA2 mirror resource, and then uses the OA2 system information in the OA2 deployment template to update the mirror parameters in the OA2 mirror resource to obtain the updated OA2 mirror resource. Then, the updated OA2 mirror resource is used to deploy OA2 in the computing node C.

[0521] After OA2 is successfully deployed, the node agent in computing node C can generate an OA identifier for the newly added OA2, namely OA-C-1, and add the identifier of computing node C (computing node C) and the OA identifier of the newly added OA2 (OA-C-1) to the OA2 addition success message.

[0522] S1209: Computing node C sends an OA2 addition success message to the MA. Correspondingly, the MA receives the OA2 addition success message sent by computing node C.

[0523] In the embodiment of the present application, the MA receives the OA2 addition success message and adds the identifier of the computing node C and the newly added OA identifier of the OA2 to the first OA deployment relationship. The first OA deployment relationship is as follows:

[0524] <Compute Node A: OA-A-1>

[0525] <Compute Node C: OA-C-1>

[0526] S1210: OA1 receives a fourth request, where the fourth request is used to request adjustment of the SA.

[0527] S1211. OA1 obtains second network information and / or second service information of the first network based on the fourth request.

[0528] S1212: OA1 determines an SA adjustment strategy based on the second network information and / or the second service information and the first SA deployment relationship.

[0529] In this embodiment of the present application, OA1 may obtain the first SA deployment relationship from the first computing node.

[0530] In the embodiment of the present application, the SA adjustment strategy is as follows:

[0531] <Number of SAs to be deleted: 1>

[0532] <Identifier of the computing node corresponding to SA1: Computing node B>

[0533] SA1 deletion information (where the SA1 deletion information includes the identifier SA-B-1 of the SA to be deleted)

[0534] <Number of SAs to be newly added: 1>

[0535] <Identifier of the computing node where SA2 is to be deployed: Computing node C><P

[0536] SA2 addition information (where the SA2 addition information includes the computing power resource information required for deploying SA2 and the SA2 deployment template)

[0537] S1213. OA1 sends a SA1 deletion request to computing node B. Correspondingly, computing node B receives the SA1 deletion request sent by OA1.

[0538] In the embodiment of the present application, the SA1 deletion request includes SA1 deletion information. Among them, the SA1 deletion information includes the identifier (SA-B-1) of the SA1 to be deleted.

[0539] S1214. Computing node B deletes SA1 in computing node B based on the identifier of the SA1 to be deleted in the SA1 deletion information.

[0540] In the embodiment of the present application, computing node B fails to delete SA1 and generates a SA1 deletion failure message.

[0541] S1215. Computing node B sends a SA1 deletion failure message to OA1. Correspondingly, OA1 receives the SA1 deletion failure message sent by computing node B.

[0542] S1216. OA1 sends a SA2 addition request to computing node C. Correspondingly, computing node C receives the SA2 addition request sent by OA1.

[0543] In the embodiment of the present application, the SA2 addition request includes SA2 addition information. Among them, the SA2 addition information includes the computing power resource information required for deploying SA2. Among them, the computing power resource information required for deploying SA2 includes the type of computing power resource required for deploying SA2, the quantity of computing power resource required for deploying SA2, the memory capacity required for deploying SA2, and the storage capacity required for deploying SA2.

[0544] S1217. Computing node C adds SA2 in computing node C based on the SA2 addition information.

[0545] In an embodiment of the present application, the node agent in the computing node C can receive a new SA2 request. After determining that the computing node C can provide the type of computing power resources required to deploy SA2, the amount of computing power resources required to deploy SA2, the memory capacity required to deploy SA2, and the storage capacity required to deploy SA2, the node agent in the computing node C determines that the SA2 system information in the SA2 deployment template is complete, and based on the SA2 deployment template, obtains the SA2 image resource, and then uses the SA2 system information in the SA2 deployment template to update the image parameters in the SA2 image resource to obtain the updated SA2 image resource. Then, the updated SA2 image resource is used to deploy SA2 in the computing node C.

[0546] After SA2 is successfully deployed, the node agent in computing node C can generate an SA identifier for the newly added SA2, namely SA-C-1, and add the identifier of computing node C (computing node C) and the SA identifier of the newly added SA2 (SA-C-1) to the SA2 addition success message.

[0547] S1218: Computing node C sends a SA2 addition success message to OA1. Correspondingly, OA1 receives the SA2 addition success message sent by computing node C.

[0548] In the embodiment of the present application, OA1 receives the SA2 addition success message and adds the identifier of the computing node C and the newly added SA identifier of SA2 to the first SA deployment relationship. The first SA deployment relationship is as follows:

[0549] <Compute Node B: SA-B-1>

[0550] <Compute Node C: SA-C-1>

[0551] After OA1 completes adjusting the SA, the first SA deployment relationship may be synchronized to the MA in the first computing node and stored in the first computing node.

[0552] After executing steps S1201 to S1218, the first OA deployment relationship stored in the first computing node is as follows:

[0553] <Compute Node A: OA-A-1>

[0554] <Compute Node C: OA-C-1>

[0555] The first SA deployment relationship stored in the first computing node is as follows:

[0556] <Compute Node B: SA-B-1>

[0557] <Compute Node C: SA-C-1>

[0558] It should be understood that, in the above steps, S1204-S1206 and S1207-S1209 are executed in a certain order, while S1213-S1215 and S1216-S1218 are executed in any certain order.

[0559] In the embodiments provided by the present application above, the methods provided by the embodiments of the present application are introduced from the perspective of interaction between each device. In order to implement the various functions in the methods provided by the embodiments of the present application above, the MA in the first computing node or the second computing node or the third computing node may include a hardware structure and / or a software module to implement the above functions in the form of a hardware structure, a software module, or a hardware structure plus a software module. Whether a function in the above functions is executed in the form of a hardware structure, a software module, or a hardware structure plus a software module depends on the specific application and design constraints of the technical solution.

[0560] The division of modules in the embodiments of the present application is illustrative and is merely a logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional modules in the various embodiments of the present application may be integrated into a single processor, or may exist physically separately, or two or more modules may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or software functional modules.

[0561] Similar to the above concept, as shown in FIG13 , an embodiment of the present application further provides a service processing device 1300 for implementing the functions of the MA in the first computing node, the second computing node, or the third computing node in the above method. For example, the service processing device may be a software module or a chip system. In the embodiment of the present application, the chip system may be composed of a chip or may include a chip and other discrete components. The service processing device 1300 may include: a communication unit 1301 and a processing unit 1302.

[0562] In the embodiment of the present application, the communication unit and processing unit included in the service processing apparatus 1300 are respectively configured to execute the sending and receiving steps of the MA in the first computing node, the second computing node, or the third computing node in the method embodiment described above. The communication unit and processing unit may be integrated into a single unit or two independent units.

[0563] The following describes in detail the service processing device provided in the embodiment of the present application in conjunction with Figures 13 and 14. 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, they will not be repeated here.

[0564] In one implementation, the service processing device 1300 may perform the following functions:

[0565] The communication unit 1301 is configured to receive a first request, where the first request is used to request the generation of an orchestration agent OA.

[0566] Processing unit 1302 is used to obtain first network information and / or first business information of the first network based on the first request, and determine the OA deployment strategy based on the first network information and / or the first business information, wherein the OA deployment strategy includes the identifier of the second computing node and the OA deployment resource information.

[0567] The communication unit 1301 is also used to send an OA deployment request to the second computing node based on the identifier of the second computing node, wherein the OA deployment request includes OA deployment resource information, and the OA deployment request is used to request the deployment of OA in the second computing node based on the OA deployment resource information; OA is used to orchestrate any business to obtain S tasks, and to determine the scheduling agent SA for scheduling the S tasks, where S is a positive integer.

[0568] In one possible design, the OA deployment resource information includes an OA deployment template or an OA executable code, which is used to deploy the OA; the OA system information in the OA deployment template includes the type of OA to be deployed; and the OA executable code includes the type of OA to be deployed.

[0569] In one possible design, the processing unit 1302 is specifically used to input the first network information and / or the first business information into the first model to obtain the OA deployment strategy, and the first model has the ability to generate the OA deployment strategy.

[0570] In one possible design, the communication unit 1301 is also used to receive an OA deployment failure message sent by the second computing node when the OA deployment on the second computing node fails; and to receive an OA deployment success message sent by the second computing node when the OA deployment on the second computing node succeeds, the OA deployment success message including the identifier of the second computing node and the OA identifier of the OA deployed in the second computing node; and to add the identifier of the second computing node and the OA identifier of the OA deployed in the second computing node to the first OA deployment relationship.

[0571] In one possible design, the communication unit 1301 is further configured to receive a second request, where the second request is used to request adjustment of the OA.

[0572] The processing unit 1302 is further configured to obtain, based on the second request, second network information and / or second service information of the first network, where the second network information is the adjusted first network information, and the second service information is the adjusted first service information;

[0573] The processing unit 1302 is further configured to determine an OA adjustment policy based on the second network information and / or the second service information and the first OA deployment relationship, where the OA adjustment policy includes an identifier of the second computing node and the OA adjustment information;

[0574] The communication unit 1301 is further configured to send an OA adjustment request to the second computing node based on the identifier of the second computing node. The OA adjustment request includes OA adjustment information. The OA adjustment request is configured to request the second computing node to adjust the deployed OA based on the OA adjustment information.

[0575] In one possible design, the communication unit 1301 is also used to receive an OA adjustment failure message sent by the second computing node when the second computing node fails to adjust the deployed OA; and to receive an OA adjustment success message sent by the second computing node when the second computing node successfully adjusts the deployed OA, the OA adjustment success message including the identifier of the second computing node and the OA identifier of the adjusted OA in the second computing node; and to modify the first OA deployment relationship based on the identifier of the second computing node and the OA identifier of the adjusted OA.

[0576] In one possible design, the OA system information includes at least one of the following: role information, orchestration rules, first orchestration resource information, second orchestration resource information, orchestration memory information, and orchestration output format information; the orchestration role information is used to indicate the type to be deployed, the orchestration rules are the criteria used to orchestrate the business, the first orchestration resource information is used to indicate the resources used to orchestrate the business to obtain a first orchestration result, the second orchestration resource information is used to indicate the resources used to adjust the first orchestration result to obtain a second orchestration result, the orchestration memory information is used to indicate the memory used for the obtained orchestration result, and the orchestration output format information is used to indicate the output format of the orchestration result.

[0577] In one possible design, the first request is also used to request the generation of an SA;

[0578] The processing unit 1302 is further configured to determine an SA deployment strategy based on the first network information and / or the first service information, where the SA deployment strategy includes an identifier of the third computing node and SA deployment resource information.

[0579] The communication unit 1301 is also used to send an SA deployment request to the third computing node based on the identifier of the third computing node. The SA deployment request includes SA deployment resource information. The SA deployment request is used to request the deployment of SA in the third computing node based on the SA deployment resource information; SA is used to determine the fourth computing node that executes the received task.

[0580] In one possible design, the SA deployment resource information includes an SA deployment template or SA executable code, which is used to deploy SA; the SA system information in the SA deployment template includes the type of SA to be deployed; and the SA executable code includes the type of SA to be deployed.

[0581] In one possible design, the processing unit 1302 is specifically used to input the first network information and / or the first business information into the first model to obtain the SA deployment strategy. The first model also has the ability to generate the SA deployment strategy.

[0582] In one possible design, the communication unit 1301 is also used to receive an SA deployment failure message sent by the third computing node when the deployment of SA on the third computing node fails; and to receive an SA deployment success message sent by the third computing node when the deployment of SA on the third computing node succeeds; the SA deployment success message includes the identifier of the third computing node and the SA identifier of the SA deployed by the third computing node; and the identifier of the third computing node and the SA identifier of the SA deployed in the third computing node are added to the first SA deployment relationship.

[0583] In one possible design, the communication unit 1301 is further used to receive a second request, where the second request is used to request adjustment of the SA.

[0584] The processing unit 1302 is further configured to obtain second network information and / or second service information of the first network based on the second request, where the second network information is the adjusted first network information and the second service information is the adjusted first service information.

[0585] The processing unit 1302 is further configured to determine an SA adjustment policy based on the second network information and / or the second service information and the first SA deployment relationship, where the SA adjustment policy includes an identifier of the third computing node and SA adjustment information.

[0586] The communication unit 1301 is further configured to send an SA adjustment request to the third computing node based on the identifier of the third computing node. The SA adjustment request includes SA adjustment information. The SA adjustment request is used to request the third computing node to adjust the deployed SA based on the SA adjustment information.

[0587] In one possible design, the communication unit 1301 is also used to receive an SA adjustment failure message sent by the third computing node when the third computing node fails to adjust the deployed SA; and to receive an SA adjustment success message sent by the third computing node when the third computing node successfully adjusts the deployed SA, the SA adjustment success message including the identifier of the third computing node and the SA identifier of the adjusted SA in the third computing node; and to modify the first SA deployment relationship based on the identifier of the third computing node and the SA identifier of the adjusted SA.

[0588] In one possible design, the SA system information includes at least one of the following: role information, scheduling rules, first scheduling resource information, second scheduling resource information, scheduling memory information, and scheduling output format information; the scheduling role information is used to indicate the type to be deployed, the scheduling rules are the criteria used for scheduling the received tasks, the first scheduling resource information is used to indicate the resources used to schedule the received tasks to obtain the first scheduling result, the second scheduling resource information is used to indicate the resources used to adjust the first scheduling result to obtain the second scheduling result, the scheduling memory information is used to indicate the memory used for the obtained scheduling result, and the scheduling output format information is used to indicate the output format of the scheduling result.

[0589] In one implementation, the service processing device 1300 may perform the following functions:

[0590] The communication unit 1301 is configured to receive an OA deployment request sent by the MA in the first computing node, where the OA deployment request includes OA deployment resource information.

[0591] The processing unit 1302 is configured to deploy the OA on the second computing node based on the OA deployment resource information; the OA is configured to orchestrate any business to obtain S tasks, and to determine the SA for scheduling the S tasks, where S is a positive integer.

[0592] In one possible design, the OA deployment resource information includes an OA deployment template or an OA executable code, which is used to deploy the OA; the OA system information in the OA deployment template includes the type of OA to be deployed; and the OA executable code includes the type of OA to be deployed.

[0593] In one possible design, the OA deployment resource information includes an OA deployment template;

[0594] The processing unit 1302 is specifically used to obtain OA mirror resources based on the OA deployment template; the OA mirror resources are used to deploy OA; the OA system information in the OA deployment template is used to update the mirror parameters in the OA mirror resources to obtain the updated OA mirror resources; the updated OA mirror resources are used to deploy OA in the second computing node.

[0595] In one possible design, the OA deployment resource information includes OA executable code;

[0596] The processing unit 1302 is specifically configured to deploy the OA in the second computing node using the OA executable code.

[0597] In one possible design, the communication unit 1301 is further used to receive an OA adjustment request sent by the MA, where the OA adjustment request includes OA adjustment information.

[0598] The processing unit 1302 is further configured to adjust the deployed OA in the second computing node based on the OA adjustment information.

[0599] In one implementation, the service processing device 1300 may perform the following functions:

[0600] The communication unit 1301 is configured to receive an SA deployment request, where the SA deployment request includes SA deployment resource information.

[0601] The processing unit 1302 is configured to deploy the SA on the third computing node based on the SA deployment resource information, wherein the SA is used to determine a fourth computing node that executes the received task.

[0602] In one possible design, the SA deployment request comes from the MA in the first computing node, or the OA in the second computing node.

[0603] In one possible design, the SA deployment resource information includes an SA deployment template or SA executable code, which is used to deploy SA; the SA system information in the SA deployment template includes the type of SA to be deployed; and the SA executable code includes the type of SA to be deployed.

[0604] In one possible design, the SA deployment resource information includes an SA deployment template;

[0605] The processing unit 1302 is specifically used to obtain SA mirror resources based on the SA deployment template in the SA deployment template; the SA mirror resources are used to deploy SA; the SA system information in the SA deployment template is used to update the mirror parameters in the SA mirror resources to obtain the updated SA mirror resources; the updated SA mirror resources are used to deploy SA in the third computing node.

[0606] In one possible design, the SA deployment resource information includes SA executable code;

[0607] The processing unit 1302 is specifically configured to deploy the SA in the third computing node using the SA executable code.

[0608] In one possible design, the communication unit 1301 is also used to receive an SA adjustment request, which includes SA adjustment information.

[0609] The processing unit 1302 is further configured to adjust the deployed SA in the third computing node based on the SA adjustment information.

[0610] The above are just examples. The communication unit and processing unit in the business processing device 1300 can also perform other functions. For more detailed descriptions, please refer to the relevant descriptions in the method embodiment shown above, which will not be repeated here.

[0611] In the embodiments provided above, the methods provided in the embodiments of the present application are described from the perspective of a business processing device as the execution subject. To implement the various functions of the methods provided in the embodiments of the present application, the business processing device may include a hardware structure and / or a software module, and implement the aforementioned functions in the form of a hardware structure, a software module, or a hardware structure plus a software module. Whether a particular one of the aforementioned functions is implemented in the form of a hardware structure, a software module, or a hardware structure plus a software module depends on the specific application and design constraints of the technical solution.

[0612] For example, when hardware implementation is adopted, the hardware implementation of the service processing device may refer to FIG. 14 and its related description.

[0613] Referring to Figure 14 , the service processing apparatus may include: one or more processors 1402; a memory 1403; one or more application programs (not shown); and one or more computer programs 1404. The aforementioned components may be connected via one or more communication buses 1401. The one or more computer programs 1404 are stored in the aforementioned memory 1403 and configured to be executed by the one or more processors 1402. The one or more computer programs 1404 include instructions that can be used to execute the method of any of the aforementioned embodiments. The one or more processors 1402 may perform the functions of the communication unit 1301 and the processing unit 1302.

[0614] An embodiment of the present application further provides a computer-readable storage medium, which stores computer instructions. When the computer instructions are executed on a service processing device, the service processing device implements the service processing method in the above embodiment.

[0615] An embodiment of the present application further provides a computer program product, which, when executed on a computer, enables the computer to execute the business processing method in the above embodiment.

[0616] Among them, the business processing device, computer-readable storage medium, computer program product or chip provided in the embodiments of the present application are all used to execute the corresponding methods provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0617] Through the description of the above implementation methods, technical personnel in the relevant field can understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0618] 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 merely schematic. For example, the division of modules or units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that 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.

[0619] Units described as separate components may or may not be physically separate, and components shown as units may be one physical unit or multiple physical units, that is, they may be located in one place or distributed in multiple places. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0620] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0621] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling a device (which can be a single-chip microcomputer, chip, etc.) or a processor (processor) to execute all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0622] The above content 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 the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A service processing method, characterized in that The method is applied to a management agent MA deployed on a first computing node in a first network. The first network includes at least one computing node, and the first computing node is any one of the at least one computing node. The method includes: Receiving a first request for requesting to generate an orchestration agent OA; Based on the first request, obtaining first network information and / or first service information of the first network; Based on the first network information and / or the first service information, determining an OA deployment policy, where the OA deployment policy includes an identifier of a second computing node and OA deployment resource information; Based on the identifier of the second computing node, sending an OA deployment request to the second computing node. The OA deployment request includes the OA deployment resource information, and is used to request to deploy the OA in the second computing node based on the OA deployment resource information. The OA is used to orchestrate any service to obtain S tasks, and determine a scheduling agent SA for scheduling the S tasks, where S is a positive integer.

2. The method according to claim 1, characterized in that, The OA deployment resource information includes an OA deployment template or OA executable code, which is used to deploy the OA. The OA system information in the OA deployment template includes the type of the OA to be deployed. The OA executable code includes the type of the OA to be deployed.

3. The method according to claim 1, characterized in that The determining the OA deployment policy based on the first network information and / or the first service information includes: Inputting the first network information and / or the first service information into a first model to obtain the OA deployment policy, where the first model has the ability to generate the OA deployment policy.

4. The method according to any one of claims 1-3, characterized in that The method further includes: In the case that the deployment of the OA on the second computing node fails, receiving an OA deployment failure message sent by the second computing node; In the case that the deployment of the OA on the second computing node is successful, receiving an OA deployment success message sent by the second computing node, where the identifier of the second computing node and the OA identifier of the OA deployed in the second computing node are included in the OA deployment success message; Adding the identifier of the second computing node and the OA identifier of the OA deployed in the second computing node to a first OA deployment relationship.

5. The method according to any one of claims 1-4, characterized in that The method further includes: Receiving a second request for requesting to adjust the OA; Based on the second request, obtaining second network information and / or second service information of the first network. The second network information is the adjusted first network information, and the second service information is the adjusted first service information; Based on the second network information and / or the second service information, and the first OA deployment relationship, determining an OA adjustment policy, where the OA adjustment policy includes the identifier of the second computing node and OA adjustment information; Based on the identifier of the second computing node, send the OA adjustment request to the second computing node, where the OA adjustment request includes the OA adjustment information, and the OA adjustment request is used to request to adjust the deployed OA in the second computing node based on the OA adjustment information.

6. The method according to claim 5, wherein The method further includes: In the case where the adjustment of the deployed OA in the second computing node fails, receive the OA adjustment failure message sent by the second computing node; In the case where the adjustment of the deployed OA in the second computing node is successful, receive the OA adjustment success message sent by the second computing node, where the OA adjustment success message includes the identifier of the second computing node and the OA identifier of the adjusted OA in the second computing node; based on the identifier of the second computing node and the OA identifier of the adjusted OA, modify the first OA deployment relationship.

7. The method according to claim 2, wherein The OA system information includes at least one of the following: Role information, orchestration rules, first orchestration resource information, second orchestration resource information, orchestration memory information, orchestration output format information; The orchestration role information is used to indicate the type to be deployed, the orchestration rules are the criteria used to orchestrate the service, the first orchestration resource information is used to indicate the resources used to obtain the first orchestration result by orchestrating the service, the second orchestration resource information is used to indicate the resources used to adjust the first orchestration result to obtain the second orchestration result, the orchestration memory information is used to indicate the memory used for the obtained orchestration result, and the orchestration output format information is used to indicate the output format of the orchestration result.

8. The method according to claim 1, wherein The first request is further used to request the generation of the SA; the method further includes: Based on the first network information and / or the first service information, determine the SA deployment policy, where the SA deployment policy includes the identifier of the third computing node and the SA deployment resource information; Based on the identifier of the third computing node, send the SA deployment request to the third computing node, where the SA deployment request includes the SA deployment resource information, and the SA deployment request is used to request to deploy the SA in the third computing node based on the SA deployment resource information; the SA is used to determine the fourth computing node that executes the received task.

9. The method according to claim 8, wherein The SA deployment resource information includes an SA deployment template or SA executable code, and the SA deployment template or the SA executable code is used to deploy the SA; the SA system information in the SA deployment template includes the type of the SA to be deployed; the SA executable code includes the type of the SA to be deployed.

10. The method according to claim 8, characterized in that, The determining the SA deployment policy based on the first network information and / or the first service information includes: Input the first network information and / or the first service information into the first model to obtain the SA deployment policy, and the first model also has the ability to generate the SA deployment policy.

11. The method according to any one of claims 8-10, characterized in that, The method further includes: In the case where the deployment of the SA in the third computing node fails, receive the SA deployment failure message sent by the third computing node; Upon successful deployment of the SA at the third computing node, receive the SA deployment success message sent by the third computing node; the SA deployment success message includes the identifier of the third computing node and the SA identifier of the SA deployed at the third computing node; add the identifier of the third computing node and the SA identifier of the SA deployed in the third computing node to the first SA deployment relationship.

12. The method according to any one of claims 8-11, characterized in that, The method further includes: Receive the second request for requesting adjustment of the SA. Based on the second request, obtain the second network information and / or second service information of the first network, where the second network information is the adjusted first network information and the second service information is the adjusted first service information. Based on the second network information and / or the second service information, and the first SA deployment relationship, determine an SA adjustment policy, where the SA adjustment policy includes the identifier of the third computing node and SA adjustment information. Based on the identifier of the third computing node, send the SA adjustment request to the third computing node, where the SA adjustment request includes the SA adjustment information, and the SA adjustment request is used to request adjustment of the deployed SA in the third computing node based on the SA adjustment information.

13. The method according to claim 12, characterized in that, The method further includes: Upon failure of the third computing node to adjust the deployed SA, receive the SA adjustment failure message sent by the third computing node. Upon successful adjustment of the deployed SA by the third computing node, receive the SA adjustment success message sent by the third computing node, where the SA adjustment success message includes the identifier of the third computing node and the SA identifier of the SA that has been adjusted in the third computing node; modify the first SA deployment relationship based on the identifier of the third computing node and the SA identifier of the adjusted SA.

14. The method according to claim 9, characterized in that, The SA system information includes at least one of the following: Role information, scheduling rules, first scheduling resource information, second scheduling resource information, scheduling memory information, scheduling output format information. The scheduling role information is used to indicate the type to be deployed, the scheduling rules are the criteria used to schedule the received tasks, the first scheduling resource information is used to indicate the resources used to schedule the received tasks to obtain the first scheduling result, the second scheduling resource information is used to indicate the resources used to adjust the first scheduling result to obtain the second scheduling result, the scheduling memory information is used to indicate the memory used for the obtained scheduling result, and the scheduling output format information is used to indicate the output format of the scheduling result.

15. A service processing method, characterized in that, The method is applied to a second computing node in a first network, where the first network includes at least one computing node, and the second computing node is any one of the at least one computing node. The method includes: Receive the OA deployment request sent by the MA in the first computing node, where the OA deployment request includes OA deployment resource information. Deploy OA on the second computing node based on the OA deployment resource information; the OA is used to orchestrate any service to obtain S tasks, and determine an SA for scheduling the S tasks, where S is a positive integer.

16. The method according to claim 15, wherein The OA deployment resource information includes an OA deployment template or OA executable code, and the OA deployment template or the OA executable code is used to deploy the OA; the OA system information in the OA deployment template includes the type of the OA to be deployed; the OA executable code includes the type of the OA to be deployed.

17. The method according to claim 16, wherein The OA deployment resource information includes the OA deployment template. The deploying OA on the second computing node based on the OA deployment resource information includes: Obtain OA image resources based on the OA deployment template; the OA image resources are used to deploy the OA. Update the image parameters in the OA image resources by using the OA system information in the OA deployment template to obtain updated OA image resources. Deploy the OA in the second computing node by using the updated OA image resources.

18. The method according to claim 16, wherein The OA deployment resource information includes the OA executable code. The deploying the OA on the second computing node based on the OA deployment resource information includes: Deploy the OA in the second computing node by using the OA executable code.

19. The method according to any one of claims 15-18, characterized in that, The method further includes: Receive an OA adjustment request sent by the MA, where the OA adjustment request includes OA adjustment information. Adjust the OA deployed in the second computing node based on the OA adjustment information.

20. A service processing method, characterized in that, The method is applied to a third computing node in a first network, the first network includes at least one computing node, the third computing node is any one of the at least one computing node, and the method includes: Receive an SA deployment request, where the SA deployment request includes SA deployment resource information. Deploy an SA on the third computing node based on the SA deployment resource information; the SA is used to determine a fourth computing node for executing the received task.

21. The method according to claim 20, wherein The SA deployment request comes from the MA in the first computing node or the OA in the second computing node.

22. The method according to any one of claims 20-21, characterized in that, The SA deployment resource information includes an SA deployment template or SA executable code, and the SA deployment template or the SA executable code is used to deploy the SA; the SA system information in the SA deployment template includes the type of the SA to be deployed; the SA executable code includes the type of the SA to be deployed.

23. The method according to claim 22, wherein The SA deployment resource information includes the SA deployment template. The deploying the SA on the third computing node based on the SA deployment resource information includes: Obtain SA image resources based on the SA deployment template in the SA deployment template; the SA image resources are used to deploy the SA. Update the image parameters in the SA image resources by using the SA system information in the SA deployment template to obtain updated SA image resources. Deploy the SA in the third computing node by using the updated SA image resources.

24. The method according to claim 22, wherein The SA deployment resource information includes the SA executable code; Deploying the SA on the third computing node based on the SA deployment resource information includes: Using the SA executable code to deploy the SA in the third computing node.

25. The method according to any one of claims 20-24, characterized in that, The method further includes: Receiving an SA adjustment request, where the SA adjustment request includes SA adjustment information; Adjusting the SA deployed in the third computing node based on the SA adjustment information.

26. A service processing device, characterized in that, Including a processor and a memory; The processor is configured to execute the computer program or instructions stored in the memory, so that the communication device implements the method according to any one of claims 1-25.

27. A computer-readable storage medium, characterized in that, Storing a computer program or instructions, when the computer program or instructions run on a computer, enabling the computer-readable storage medium to implement the method according to any one of claims 1-25.

28. A chip, characterized in that, Including a processor, the processor is coupled to the memory and is configured to execute the computer program or instructions stored in the memory, so that the chip implements the method according to any one of claims 1-25.

29. A service processing system, characterized in that, Including: The MA of the first computing node, the second computing node, and the third computing node, where the MA of the first computing node is configured to implement the method according to any one of claims 1-14; The second computing node, which is configured to implement the method according to any one of claims 15-19; The third computing node, which is configured to implement the method according to any one of claims 20-25.

Images