Calculation power resource scheduling method, system and device and storage medium

Abstract

The invention belongs to the technical field of edge computing, and particularly relates to a computing power resource scheduling method, system and device and a storage medium. The objective of the invention is to solve the technical problem that a large number of cloud hosts compete for resources and the service influence range is large due to the fact that no automatic intervention mechanism exists when the node workload is too high in the prior art. The system comprises four core components: a business management scheduling component receives a user application instance and a strategy parameter, carries out force quantity calculation and generates business operation; the resource management scheduling component extracts the business job, splits the business job into a plurality of computing power sub-jobs matched with resource distribution, and deploys the computing power sub-jobs to corresponding edge nodes; the edge computing node component is responsible for executing sub-jobs and calling an interface to create a cloud host to complete resource deployment; after the job is created, the resource node scheduling component performs self-inspection according to a priority strategy, so that reasonable allocation of resources according to priorities is ensured, the stability of key tasks is ensured, and the overall scheduling efficiency is improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of edge computing, and particularly relates to a computing resource scheduling method and system, a device and a storage medium. BACKGROUND

[0002] With the development of 5G high-speed communication system, virtual reality, artificial intelligence and other technologies have gradually emerged and been applied in practice. Human production and life are gradually inseparable from Internet, Internet of Things and other technological products. The amount of data generated is increasing, and the demand for real-time data analysis is increasing. Edge computing services can be physically close to data sources to provide faster and more reliable services.

[0003] The resources of edge data centers are abundant, but they are far away from data sources and often cannot provide efficient and reliable computing services quickly. The resources of edge sites are limited, but they are close to data sources and can directly provide low-latency and reliable computing services. In addition, various computing power operation types are diverse, and task types are various.

[0004] In the prior art, computing resource scheduling often directly relies on cloud computing platforms, k8s container platforms and other data center dimension technologies. The computing power scheduling range is limited to a single cluster as a granularity, and the allocation of computing resources is uniformly executed by a scheduling program in the data center.

[0005] The prior art has the following technical problems: 1. When the node workload is too high, there is no automatic intervention mechanism, so that a large number of cloud hosts compete for resources, resulting in a large business impact range.

[0006] 2. When the overall cluster load is high, there is no automatic intervention mechanism to ensure the normal operation of core and key business resources.

[0007] 3. There is no multi-cluster, multi-data center and multi-site scheduling management method to realize the collaborative allocation of edge computing power. SUMMARY

[0008] The present application provides a computing resource scheduling method, system, device and storage medium, which aims to solve the technical problems in the prior art that when the node workload is too high, there is no automatic intervention mechanism, so that a large number of cloud hosts compete for resources, resulting in a large business impact range; when the overall cluster load is high, there is no automatic intervention mechanism to ensure the normal operation of core and key business resources; and there is no multi-cluster, multi-data center and multi-site scheduling management method to realize the collaborative allocation of edge computing power.

[0009] The technical solution of the present application to solve the above technical problems is as follows: a computing resource scheduling system, comprising: A business management scheduling component is configured to receive an application instance and a policy parameter uploaded by a user, perform a computing power measurement on the application instance based on the policy parameter, and generate a business job; A resource management scheduling component is configured to extract the business job and perform the same, generate a plurality of corresponding computing power resource sub-jobs according to the distribution positions of resources in the business job, and deploy the plurality of computing power resource sub-jobs to corresponding edge computing node components according to the distribution positions; the edge computing node components include edge data centers and edge terminal sites; An edge computing node component is configured to execute the computing power resource sub-jobs, call a resource creation interface, generate a cloud host on a corresponding computing node, and complete the deployment of resources; A resource node scheduling component is configured to perform a self-check on the business job through a priority policy after the creation of the business job, and perform a self-check on the computing power resource sub-job through the priority policy after the creation of the computing power resource sub-job.

[0010] Further, the job states of the above-mentioned jobs include an initial state, an executed state, an abandoned state, a completed state, a deletion-required state, and a destroyed state.

[0011] Further, the self-check on the business job specifically includes: Obtaining a plurality of business jobs in the executed state and the completed state, sorting the plurality of business jobs according to priorities to obtain a business job list; Iteratively checking the business job list according to the priorities from high to low to obtain a self-check job record; Based on the self-check job record, sequentially judging whether the business jobs in the business job list meet a preset business health threshold: If yes, the next business job is judged until all the business jobs in the business job list are judged; If no, monitoring data of the business job that does not meet the threshold is called by using a resource placement method, and computing power resources of the business job that does not meet the threshold are calculated by a measurement module according to the monitoring data, and a new edge data center or a new edge terminal site is obtained according to the computing power resources; The computing power resource sub-job corresponding to the business job that does not meet the threshold is deleted, a new computing power resource sub-job is generated, the corresponding computing power resources are deployed to the new edge data center or the new edge terminal site based on the new computing power resource sub-job, and the iterative checking is ended.

[0012] Further, the self-check on the computing power resource sub-job specifically includes: Obtaining a cloud host list corresponding to all the computing power resource sub-jobs; obtaining a node list where the cloud host list is located, and traversing the node list to obtain the workloads of each node; determining whether the workload of a node is higher than a load threshold value: if not, then the workload of the next node is determined in turn until the workloads of all nodes are determined; if higher, then the load monitoring information of all cloud hosts of the node is obtained, and based on the load monitoring information, a cloud host with insufficient resource computing power is obtained, and it is determined whether the cloud host with insufficient resource computing power belongs to a restricted cloud host: if not, then no processing is performed; if yes, then a resource placement method is called to obtain an idle node, and the cloud host is migrated to the idle node.

[0013] Further, the above further includes: periodically obtaining the workloads of resource nodes and virtual resources; based on the workloads of resource nodes and virtual resources, determining whether all computing power resources are sufficient: if not sufficient, then the priority order of all cloud hosts is predefined by a digital character, the cloud hosts are scheduled according to the priority order, and the cloud hosts are scheduled by a computing power competition scheduling method; if sufficient, then no processing is performed.

[0014] Further, the above, by which the priority order of all cloud hosts is predefined by a digital character, and the cloud hosts are scheduled according to the priority order, specifically includes: obtaining the node workload, and determining whether the node workload is higher than a first load threshold value: if higher than the first load threshold value, then the cloud hosts with interruptible computing power in the node are turned off one by one in descending order of priority, and after each cloud host is turned off, it is determined whether the node workload is higher than the first load threshold value; if higher, then the cloud hosts with interruptible computing power are turned off in descending order of priority until all cloud hosts with interruptible computing power are turned off or the node workload is lower than the first load threshold value; after all cloud hosts with interruptible computing power are turned off, it is determined whether the node workload is still higher than the first load threshold value, if higher than the first load threshold value, then the cloud hosts with limitable computing power in the node are limited one by one in descending order of priority, and after each cloud host is limited, it is determined whether the node workload is higher than the first load threshold value; if higher, then the cloud hosts with limitable computing power are limited in descending order of priority until all cloud hosts with limitable computing power are limited; If the node load is lower than the first load threshold, the cloud host with limited computing power in the node is released one by one according to the priority from high to low, and after the release of each cloud host, it is judged whether the node load is higher than the second load threshold; if not, the cloud host with limited computing power is continuously released according to the priority from high to low, until all the cloud hosts with limited computing power are released or the node load is higher than the second load threshold; after all the cloud hosts with limited computing power are released, it is judged whether the node load is still lower than the second load threshold, if lower than the second load threshold, the cloud host with interrupted computing power in the node is started one by one according to the priority from high to low, and after the start of each cloud host, it is judged whether the node load is lower than the second load threshold; if not, the cloud host with interrupted computing power is continuously started according to the priority from high to low, until all the cloud hosts with interrupted computing power are started.

[0015] Further, the above scheduling method for cloud host through computing power competition specifically includes: obtaining a list of cloud hosts to be limited, and judging whether there is a cloud host to be limited in the list of cloud hosts to be limited: if there is, performing computing power competition on all cloud hosts to be limited to obtain a winning cloud host; limiting the computing power of all cloud hosts to be limited except the winning cloud host, and not limiting the computing power of the winning cloud host; if not, ending.

[0016] In a second aspect, the present application also provides a scheduling method for computing power resources, comprising: receiving an application instance and a policy parameter uploaded by a user through a business management scheduling component, performing computing power measurement on the application instance based on the policy parameter, and generating a business job; extracting the business job through a resource management scheduling component and executing, generating a plurality of corresponding computing power resource sub-jobs according to the distribution position of resources in the business job, and deploying the plurality of computing power resource sub-jobs to corresponding edge computing node components according to the distribution position; wherein the edge computing node component includes an edge data center and an edge terminal site; executing the computing power resource sub-job through the edge computing node component, calling a resource creation interface, generating a cloud host on the corresponding computing node, and completing the deployment of resources; further through the resource node scheduling component, self-checking the business job through a priority strategy after the creation of the business job is completed; and self-checking the computing power resource sub-job through a priority strategy after the creation of the computing power resource sub-job is completed.

[0017] In a third aspect, the present application provides an electronic device to solve the above technical problems, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the computing resource scheduling method of the present application when executing the computer program.

[0018] In a fourth aspect, the present application provides a computer readable storage medium to solve the above technical problems, which stores a computer program, and the computer program is executable on the processor to implement the computing resource scheduling method of the present application.

[0019] Compared with the prior art, the present application has the following advantages: 1. The present application pursues full allocation of computing power for critical services, limited allocation of computing power for important services, and no allocation of computing power for ordinary services through the priority mechanism. According to the computing power competition mechanism, services with high priority can compete for more computing power time, which can ensure the continuity of high-priority services in the case of insufficient resources, and ultimately improve resource utilization.

[0020] 2. When the node where the critical service resource is located is continuously at a peak load, the present application can trigger the resource migration mechanism to select an idle machine to run the resource, i.e., allocate computing power to the idle.

[0021] 3. The present application simplifies the human configuration through the standardized policy module.

[0022] 4. The present application recycles and redeploys resources at the data center level, and allocates virtual resources required by services to more stable or idle data centers or terminal sites. Other features and advantages of the present application will be described in the following description, and some will become apparent from the description, or will be understood by those skilled in the art. The purpose and other advantages of the present application can be achieved and obtained by the structure indicated in the specification, claims and drawings. BRIEF DESCRIPTION OF DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0024] Figure 1 An execution schematic diagram of a computing resource scheduling system according to an embodiment of the present application is shown; Figure 2 A job master state and sub-state schematic diagram according to an embodiment of the present application is shown; Figure 3 Fig. 1 shows a schematic diagram of a flow process of automatic allocation of computing power resources by a service management scheduling component according to an embodiment of the present application; Figure 4 Fig. 2 shows a schematic diagram of a flow process of automatic allocation of computing power resources by a resource management scheduling component according to an embodiment of the present application; Figure 5 Fig. 3 shows a schematic diagram of a flow process of automatic allocation of computing power resources by a computing node scheduling component according to an embodiment of the present application; Figure 6 Fig. 4 shows a schematic diagram of a flow process of a priority scheduling method according to an embodiment of the present application; Figure 7 Fig. 5 shows a schematic diagram of a flow process of a computing power competition scheduling method according to an embodiment of the present application; Figure 8 Fig. 6 shows a schematic diagram of an electronic device structure according to an embodiment of the present application. DETAILED DESCRIPTION

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the following will clearly and completely explain the technical solutions in the embodiments of the present application with reference to the accompanying drawings. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of the present application.

[0026] Figure 1 Fig. 1 shows a schematic diagram of a flow process of automatic allocation of computing power resources by a service management scheduling component according to an embodiment of the present application; Figure 1 Fig. 1 shows a schematic diagram of a flow process of automatic allocation of computing power resources by a service management scheduling component according to an embodiment of the present application; The service management scheduling component is configured to receive an application instance and policy parameters uploaded by a user, perform computing power measurement on the application instance based on the policy parameters, and generate a service job. In this embodiment, the service job is put into a job pool, used for asynchronous consumption of the job, and returns job information to the submitter for tracking and management.

[0027] In this embodiment, the policy parameters include but are not limited to priority, computing power competition victory duration, resource topology relationship (expressing the distribution of cloud instances in edge data centers and edge terminal sites), label (using the label to classify assets such as edge data centers, edge terminal sites, and computing nodes, and according to the label, the deployment target of computing power resources can be determined), specification information, life cycle time, redistribution threshold, and other configuration parameters. The business management component will perform computing power measurement according to these configuration parameters, convert product-related configuration information into computing power resource-related configuration information, and determine the specific distribution position of each cloud instance at this moment.

[0028] In this embodiment, the business management scheduling component tracks the completion of the sub-job to update the job state.

[0029] The resource management scheduling component is configured to extract the business job and perform, generate a plurality of corresponding computing power resource sub-jobs according to the distribution position of the resources in the business job, and deploy the plurality of computing power resource sub-jobs to corresponding edge computing node components according to the distribution position; wherein the edge computing node component includes an edge data center and an edge terminal site. In this embodiment, 5 cloud hosts need to be created in edge data center z1, 10 cloud hosts need to be created in edge data center z2, and 1 cloud host needs to be created in edge terminal site ez1. Then 3 sub-jobs will be generated.

[0030] The edge computing node component is configured to execute the computing power resource sub-job, call a resource creation interface, generate a cloud host on a corresponding computing node, and complete the deployment of the resource. The resource node scheduling component is configured to perform self-checking on the business job through a priority policy after the business job is created, and perform self-checking on the computing power resource sub-job through the priority policy after the computing power resource sub-job is created.

[0031] Optionally, the job state of the job includes an initial state, an executed state, an abandoned state, an execution completed state, a need to delete state, and a destroyed state.

[0032] In this embodiment, as shown in Figure 2 The job attributes include: Business application driven job (app): this type of job follows the job flow shown in Figure 1 from submission to final destruction.

[0033] Self-checking job (app-check): the scheduling component periodically performs a historical business job self-checking task, and there is no deletion process and state.

[0034] The job state includes a main state and a sub-state. The main state includes: pending, Scheduling, Aborted, Finished, Deleted, Ceased. The sub-state includes: self-checking, ceasing.

[0035] Job state conversion flow pending→scheduling: When a job is submitted, its initial state is pending. Then, the job enters the scheduling state and starts to perform tasks in the expected manner.

[0036] scheduling→finished or aborted: In the scheduling state, if the job is successfully completed, the state becomes finished; if it is abandoned by external intervention, the state becomes aborted.

[0037] finished and aborted→deleted: For app-type jobs, in the finished or aborted state, when the job needs to be deleted, the state will first be set to deleted.

[0038] deleted→ceased: In the deleted state, after a certain period of time (to preserve data security for a period of time), the job enters the destruction process, and the state becomes ceased, indicating that the job and its related computing power resources have been confirmed to be destroyed.

[0039] Self-checking job (app-check): The self-checking job does not have a deleted and ceased state, so it will not go through the deletion and destruction process.

[0040] Optionally, the self-checking of the business job specifically includes: Obtaining a plurality of business jobs in an executed state and a completed execution state, sorting the plurality of business jobs according to priority to obtain a business job list; Iteratively checking the business job list according to the priority from high to low to obtain a self-checking job record; Based on the self-checking job record, sequentially judging whether the business job in the business job list meets a preset business health threshold: If it meets, the next business job is judged until all business jobs in the business job list are judged; If not, the monitoring data of the non-compliant business operation is called by using the resource placement method, and the computing resource of the non-compliant business operation is calculated by the metric module according to the monitoring data, and a new edge data center or a new edge terminal site is obtained according to the computing resource; The corresponding computing resource sub-operation of the non-compliant business operation is deleted, a new computing resource sub-operation is generated, the corresponding computing resource is deployed in the new edge data center or the new edge terminal site based on the new computing resource sub-operation, and the iteration check is ended.

[0041] As shown in Figure 3 In the embodiment, the business operation is iterated during the checking process of the business operation, the business health degree health information is obtained, and it is judged whether the business operation belongs to a long-time business unavailability according to the business health degree health information. If it belongs, other edge computing resources are obtained, then the unusable resource is recycled, and a new resource is deployed for the business operation. If it does not belong, the business operation is continuously iterated until the iteration is ended.

[0042] In the embodiment, the business management and scheduling component is a cross-data center scheduling from the perspective of a business application. The scheduling mechanism used by the business management and scheduling component for self-checking of the app-type business main operation includes a life cycle time and a resource placement mechanism.

[0043] Optionally, the self-checking of the computing resource sub-operation specifically includes: obtaining a cloud host list corresponding to all computing resource sub-operations; obtaining a node list where the cloud host list is located, and iterating the node list to obtain the workloads of each node; judging whether the workload of the node is higher than a load threshold value: If not, the workloads of the next nodes are sequentially judged until the workloads of all nodes are judged; If yes, the load monitoring information of all cloud hosts of the node is obtained, and a cloud host with insufficient resource computing power is obtained based on the load monitoring information. It is judged whether the cloud host with insufficient resource computing power belongs to a restricted cloud host: If not, no processing is performed; If yes, a resource placement method is called to obtain an idle node, and the cloud host is migrated to the idle node.

[0044] As shown in Figure 4The diagram illustrates an automatic resource allocation process for computing power by the resource management and scheduling component. This process includes: traversing the node list; if a node's workload exceeds a load threshold, obtaining the load monitoring information of the corresponding cloud host to identify those hosts with insufficient computing power for an extended period; then, calling the resource placement method to obtain an idle node and migrating the cloud host with insufficient computing power to that idle node; and finally, allocating computing power resources. The process then continues traversing the remaining nodes, checking if their workload exceeds a load threshold, until all nodes have been checked.

[0045] Optional, also includes: Periodically obtain the workload of resource nodes and the workload of virtual resources; Based on the workload of resource nodes and virtual resources, determine whether all computing resources are sufficient: If insufficient, the priority order of all cloud hosts is predefined by numerical characters, and the cloud hosts are scheduled according to the priority order; and the cloud hosts are scheduled by a computing power competition scheduling method. If sufficient, no action is taken.

[0046] like Figure 5 As shown, in this embodiment, the resource node scheduling component automatically allocates computing resources, obtains the node workload, and determines whether the node's computing resources are sufficient. If sufficient, the process ends; if insufficient, a priority mechanism is used to restrict all cloud hosts with loads exceeding the load threshold, and a computing power contention scheduling method is used to remove the restrictions on each restricted cloud host.

[0047] Optionally, the priority order of all cloud hosts is predefined using numeric characters, and scheduling the cloud hosts according to the priority order specifically includes: Obtain the node workload and determine if the node workload exceeds the first load threshold: If the workload exceeds the first load threshold, the cloud hosts with interruptible computing power in the node are shut down one by one according to priority from high to low. After each cloud host is shut down, it is determined whether the node workload exceeds the first load threshold. If it does, the shutdown of cloud hosts with interruptible computing power continues according to priority from high to low until all cloud hosts with interruptible computing power are shut down or the node workload is lower than the first load threshold. After all cloud hosts with interruptible computing power are shut down, it is determined whether the node workload still exceeds the first load threshold. If it does, the restriction of cloud hosts with finite computing power in the node is performed one by one according to priority from high to low. After each cloud host is restricted, it is determined whether the node workload exceeds the first load threshold. If it does, the restriction of cloud hosts with finite computing power continues according to priority from high to low until all cloud hosts with finite computing power are restricted. If the node load is lower than the first load threshold, the cloud host with limited computing power in the node is released one by one according to the priority from high to low, and after each cloud host is released, it is judged whether the node load is higher than the second load threshold; if not, the cloud host with limited computing power is continued to be released according to the priority from high to low, until all the cloud hosts with limited computing power are released or the node load is higher than the second load threshold; after all the cloud hosts with limited computing power are released, it is judged whether the node load is still lower than the second load threshold, if lower than the second load threshold, the cloud host with interrupted computing power in the node is started one by one according to the priority from high to low, and after each cloud host is started, it is judged whether the node load is lower than the second load threshold; if not, the cloud host with interrupted computing power is continued to be started according to the priority from high to low, until all the cloud hosts with interrupted computing power are started.

[0048] As shown in Figure 6 , for the node with high load, first judge whether the priority of the cloud host in the node belongs to interruptable cloud host or limitable cloud host, when the street load is high, first shut down the interruptable cloud host, when all the interruptable cloud hosts are shut down, judge whether the load is high, if high, limit the limitable cloud host, if not high, end. When the load is not high for the first time, limit the limitable cloud host.

[0049] Optionally, the scheduling of the cloud host by the computing power competition scheduling method specifically includes: obtaining a list of cloud hosts to be limited, and judging whether there is a cloud host to be limited in the list of cloud hosts to be limited: if there is, performing computing power competition on all cloud hosts to be limited to obtain a winning cloud host; limiting the computing power of all cloud hosts to be limited except the winning cloud host, and the winning cloud host is not limited in computing power; if not, end.

[0050] As shown in Figure 7 , the flow of the computing power competition scheduling method is used to judge whether there is a limited cloud host, when there is, the winning cloud host is calculated by the computing power competition method, and the limitation is released.

[0051] In this embodiment, the priority mechanism is to set the priority, which is a natural number, starting from 0, and the maximum value is 1024, and the larger the value is, the lower the priority is. Among them, 0 represents that full computing power must be provided, 1-999 represents that limited computing power is allowed to be provided, and 1000-1024 represents that computing power can be interrupted. The most basic design goal is that the resources with a priority of 0 are not intervened by the application; when the computing power is insufficient, the computing power is interrupted or limited according to the priority from low to high; when the computing power is sufficient, limited or full computing power is provided according to the priority from high to low.

[0052] The computing power competition mechanism: this computing power competition mechanism only runs between tasks with a priority of 1-999 (i.e. "allowing to provide limited computing power"). It does not involve tasks with a priority of 0 (which must be fully allocated), nor does it schedule interruptable tasks with a priority of 1000-1024. When multiple limited computing power tasks compete for resources, the system decides resource allocation through two rounds of calculation, sets a "computing power competition winning duration", the unit is second; within this time period, the winning task will continuously obtain the computing power resources it expects; after the time is up, it will participate in the next round of competition.

[0053] Based on the same principle as the method shown in Figure 1 Based on the same principle as the method shown in The business management scheduling component receives the application instance and the policy parameter uploaded by the user, performs computing power measurement on the application instance based on the policy parameter, and generates a business job; The resource management scheduling component extracts and executes the business job, generates a plurality of corresponding computing power resource sub-jobs according to the distribution position of the resources in the business job, and deploys the plurality of computing power resource sub-jobs to corresponding edge computing node components according to the distribution position; wherein the edge computing node component includes an edge data center and an edge terminal site; The edge computing node component executes the computing power resource sub-job, calls a resource creation interface, generates a cloud host on the corresponding computing node, and completes the deployment of the resource; The business management scheduling component also performs self-checking on the business job using the priority policy, and the resource management scheduling component also performs self-checking on the computing power resource sub-job using the priority policy.

[0054] The computing power resource scheduling system of the embodiment of the application can execute the computing power resource scheduling method provided by the embodiment of the application, and the implementation principle is similar. The actions performed by each module and unit in the computing power resource scheduling system in each embodiment of the application correspond to the steps in the computing power resource scheduling method in each embodiment of the application. For detailed descriptions of the functions of each module of the computing power resource scheduling system, please refer to the descriptions of the corresponding computing power resource scheduling method shown in the foregoing.

[0055] The scheduling system of the computing resource can be a computer program (including program code) running in a computer device, for example, the scheduling system of the computing resource is an application software; the application software can be used to execute the corresponding steps in the method provided by the embodiments of the present application.

[0056] The modules described in the embodiments of the present application can be implemented in the form of software or in the form of hardware. In some cases, the name of the module does not constitute a limitation on the module itself.

[0057] Based on the same principle as the method shown in the embodiments of the present application, the embodiments of the present application also provide an electronic device, which can include but is not limited to: a processor and a memory; the memory is used to store a computer program; the processor is used to execute the method shown in any of the embodiments of the present application by calling the computer program.

[0058] In an optional embodiment, an electronic device is provided, as shown in Figure 8 As shown in the electronic device shown in Figure 8 The electronic device includes a processor and a memory. The processor and the memory are connected, such as through a bus. Optionally, the electronic device can also include a transceiver, which can be used for data interaction between the electronic device and other electronic devices, such as data transmission and / or data reception, etc. It should be noted that in actual application, the transceiver is not limited to one, and the structure of the electronic device does not constitute a limitation on the embodiments of the present application.

[0059] The memory is used to store the application program code (computer program) for executing the scheme of the present application, and the processor is used to control the execution. The processor is used to execute the application program code stored in the memory to realize the content shown in the foregoing method embodiments.

[0060] The electronic device can also be a terminal device, Figure 8 The electronic device shown is only an example and should not limit the function and use range of the embodiments of the present application.

[0061] The embodiments of the present application provide a computer readable storage medium, which stores a computer program, when the computer program runs on a computer, the computer can execute the corresponding content in the foregoing method embodiments.

[0062] According to another aspect of the present application, there is also provided a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. A processor of a computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to cause the computer device to perform the methods provided in the various embodiments.

[0063] It should be understood that the flowchart and the block diagram in the drawings illustrate the possible implementation of the methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart and the block diagram can represent a module, a segment, or a portion of code, which comprises one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions noted in the flowchart or the block diagram can occur out of the order noted in the drawings. For example, two blocks shown in succession may, in fact, be executed substantially concurrently or in the reverse order, depending on the functionality involved. It will also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can be implemented by dedicated hardware-based systems that perform the specified functions or operations, or combinations of hardware and software.

[0064] The above computer readable storage medium stores one or more programs, which, when executed by the electronic device, cause the electronic device to perform the methods described in the above embodiments.

[0065] The above description merely illustrates the preferred embodiments of the present application and the principles of the technology applied. It is understood that the disclosed scope of the present application is not limited to the specific combinations of technical features described above, but also covers other technical solutions formed by any combinations of the above technical features or equivalent features thereof without departing from the above disclosed concept. For example, the above technical features can be replaced with other technical features disclosed in the present application (but not limited to) having similar functions to form technical solutions.

Claims

1. A computing resource scheduling system, characterized in that, The system includes: The business management and scheduling component is used to receive application instances and policy parameters uploaded by users, calculate the computational power of the application instances based on the policy parameters, and generate business jobs. A resource management and scheduling component is used to extract and execute the business jobs, generate multiple corresponding computing resource sub-jobs based on the distribution location of resources in the business jobs, and deploy the multiple computing resource sub-jobs to the corresponding edge computing node components according to their distribution location; wherein, the edge computing node components include edge data centers and edge terminal sites; Edge computing node components are used to execute the computing resource sub-jobs, call the resource creation interface, generate cloud hosts on the corresponding computing nodes, and complete the deployment of resources; The resource node scheduling component is used to perform self-checks on business jobs after they are created, based on priority strategies; it is also used to perform self-checks on computing resource sub-jobs after they are created, based on priority strategies.

2. The computing resource scheduling system according to claim 1, characterized in that, The job status includes: initial status, executed status, abandoned status, completed status, need to be deleted status, and destroyed status.

3. The computing resource scheduling system according to claim 2, characterized in that, Self-inspection of business operations specifically includes: Retrieve multiple business jobs with statuses of "being executed" and "execution completed", sort the multiple business jobs according to priority, and obtain a list of business jobs; The business task list is traversed and checked in descending order of priority to obtain the self-check task record; Based on the self-inspection work records, sequentially determine whether the business jobs in the business job list meet the preset business health threshold: If the condition is met, proceed to the next business task, and continue until all business tasks in the business task list have been evaluated. If it does not meet the requirements, the monitoring data of the non-compliant business operation is called using the resource placement method, and the computing power resources of the non-compliant business operation are calculated by the measurement module based on the monitoring data. A new edge data center or a new edge terminal site is obtained based on the computing power resources. Delete the computing resource sub-jobs corresponding to the non-compliant business jobs, generate new computing resource sub-jobs, deploy the corresponding computing resources to the new edge data center or new edge terminal site based on the new computing resource sub-jobs, and end the traversal check.

4. The computing resource scheduling system according to claim 1, characterized in that, The self-check of computing resource sub-jobs specifically includes: Get the list of cloud hosts corresponding to all computing power resource sub-jobs; Obtain the list of nodes containing the cloud host list, and traverse the list of nodes to obtain the workload of each node; Determine if the node's workload exceeds the load threshold: If it is not higher, then the workload of the next node is judged in turn, until the workload of all nodes is judged. If the load is higher, then obtain the load monitoring information of all cloud hosts on that node. Based on the load monitoring information, identify the cloud hosts with insufficient computing power and determine whether they are restricted cloud hosts. If not, no action will be taken; If so, the resource placement method is called to obtain an idle node, and the cloud host is migrated to the idle node.

5. A computing resource scheduling system according to claim 1, characterized in that, Also includes: Periodically obtain the workload of resource nodes and the workload of virtual resources; Based on the workload of resource nodes and virtual resources, determine whether all computing resources are sufficient: If insufficient, the priority order of all cloud hosts is predefined by numerical characters, and the cloud hosts are scheduled according to the priority order; and the cloud hosts are scheduled by a computing power competition scheduling method. If sufficient, no action is taken.

6. A computing resource scheduling system according to claim 5, characterized in that, The priority order of all cloud hosts is predefined using numeric characters, and the scheduling of cloud hosts according to the priority order specifically includes: Obtain the node workload and determine if the node workload exceeds the first load threshold: If the workload exceeds the first load threshold, the cloud hosts with interruptible computing power in the node are shut down one by one according to priority from high to low. After each cloud host is shut down, it is determined whether the node workload exceeds the first load threshold. If it does, the shutdown of cloud hosts with interruptible computing power continues according to priority from high to low until all cloud hosts with interruptible computing power are shut down or the node workload is lower than the first load threshold. After all cloud hosts with interruptible computing power are shut down, it is determined whether the node workload still exceeds the first load threshold. If it does, the restriction of cloud hosts with finite computing power in the node is performed one by one according to priority from high to low. After each cloud host is restricted, it is determined whether the node workload exceeds the first load threshold. If it does, the restriction of cloud hosts with finite computing power continues according to priority from high to low until all cloud hosts with finite computing power are restricted. If the load is below the first load threshold, the cloud hosts with restricted computing power in the node are unrestricted one by one according to priority from high to low. After unrestricting each cloud host, it is determined whether the node workload is higher than the second load threshold. If it is lower, the cloud hosts with restricted computing power are unrestricted one by one according to priority from high to low until all cloud hosts with restricted computing power are unrestricted or the node workload is higher than the second load threshold. After all cloud hosts with restricted computing power are unrestricted, it is determined whether the node workload is still lower than the second load threshold. If it is lower than the second load threshold, the cloud hosts with interrupted computing power in the node are powered on one by one according to priority from high to low. After each cloud host is powered on, it is determined whether the node workload is lower than the second load threshold. If it is lower, the cloud hosts with interrupted computing power are powered on one by one according to priority from high to low until all cloud hosts with interrupted computing power are powered on.

7. A computing resource scheduling system according to claim 5, characterized in that, The scheduling of cloud servers through computing power contention scheduling methods specifically includes: Retrieve the list of cloud servers awaiting restriction, and determine if any cloud servers in the list are subject to restriction: If they exist, compete for computing power among all cloud hosts to be restricted, and obtain a winning cloud host; Computing power limits are imposed on all cloud servers to be limited except for the winning cloud server; the winning cloud server is not subject to computing power limits. If it does not exist, then the process ends.

8. A method for scheduling computing resources, characterized in that, include: The business management and scheduling component receives application instances and policy parameters uploaded by users, calculates the application instance based on the policy parameters, and generates a business job. The resource management and scheduling component extracts and executes the business jobs, generates multiple corresponding computing resource sub-jobs based on the distribution location of resources in the business jobs, and deploys the multiple computing resource sub-jobs to the corresponding edge computing node components according to their distribution location; wherein, the edge computing node components include edge data centers and edge terminal sites; The computing resource sub-jobs are executed by the edge computing node components, the resource creation interface is called, and cloud hosts are generated on the corresponding computing nodes to complete the deployment of resources; It also uses a resource node scheduling component to perform self-checks on business jobs after they are created, based on a priority strategy; and to perform self-checks on computing resource sub-jobs after they are created, based on a priority strategy.

9. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method of claim 8.

10. A computer storage medium, characterized in that, The computer storage medium stores a computer program, which, when executed by a processor, implements the method of claim 8.

Images