Heterogeneous multi-core coexistence management system and method for cloud computing

Abstract

The invention provides a heterogeneous multi-core coexistence management system and method for cloud computing, and relates to the technical field of cloud platforms, and the system comprises a cluster information obtaining module which is used for obtaining equipment information and performance information of a plurality of server clusters with different cores in the system; the container establishing module is used for establishing a container by utilizing a preset container technology and establishing a mapping network between the container and the server cluster; the intercommunication establishment module is used for establishing a unified API interface and middleware and outputting cluster intercommunication information; and the multi-core resource scheduling module is used for acquiring a resource scheduling strategy and a resource scheduling method matched with the real-time demand-resource view, and distributing tasks in the application to nodes under the corresponding server cluster in combination with cluster intercommunication information. By utilizing the advantages of different cloud platforms, flexible and dynamic configuration of heterogeneous multi-core cloud platform resources is realized, the utilization rate of the resources is improved, and the risk caused by binding a single cloud platform is reduced.

Description

Technical Field

[0001] The present invention relates to the field of cloud platform technology, and in particular to a heterogeneous multi-core coexistence management system and method for cloud computing. Background Art

[0002] In cloud computing environments, tasks are usually calculated and stored on large-scale clusters to improve performance and efficiency. However, with the rapid growth of data volume, the demand for computing power from new businesses, and the steady development of domestically produced independent and controllable hardware, traditional single-type chip computing power can no longer meet the demand, and the development of diversified computing power has become a trend.

[0003] At present, due to the differences in the performance and characteristics of computing resources of different platform architectures, and the possible differences in data transmission speeds between heterogeneous computing nodes, there are problems such as low data transmission efficiency between platforms, large transmission delays, scattered resources on each platform, and low resource utilization.

[0004] Therefore, the present invention provides a heterogeneous multi-core coexistence management system and method for cloud computing. Summary of the invention

[0005] The present invention provides a heterogeneous multi-core coexistence management system and method for cloud computing, which is used to utilize the advantages of different cloud platforms, realize flexible and dynamic configuration of heterogeneous multi-core cloud platform resources, improve resource utilization, enhance the flexibility and scalability of heterogeneous multi-core cloud platforms, and reduce the risks caused by binding to a single cloud platform.

[0006] In one aspect, the present invention provides a heterogeneous multi-core coexistence management system for cloud computing, comprising: The cluster information acquisition module is used to obtain the device information and performance information of multiple server clusters with different cores in the system, and summarize and output the multi-core heterogeneous cluster information; A container establishment module, used to establish multiple containers based on the multi-core heterogeneous cluster information using a preset container technology, and to construct a mapping network between each of the containers and each server cluster; An intercommunication establishment module, used to establish a unified API interface and middleware based on the multi-core heterogeneous cluster information and the mapping network, and output cluster intercommunication information; The multi-core resource scheduling module is used to obtain the application demand information of the system and the resource information of each server cluster, output the real-time demand-resource view, and obtain the resource scheduling strategy and resource scheduling method that match the real-time demand-resource view in the strategy-method database. Based on the resource scheduling strategy and resource scheduling method, and in combination with the cluster intercommunication information, the tasks in the application are allocated to the nodes under the corresponding server cluster.

[0007] Preferably, the cluster information acquisition module includes: The cluster category information acquisition submodule is used to obtain the type information of each server cluster in the system and output the cluster category information; The chip-device-performance information acquisition submodule is used to obtain the chip environment information corresponding to each category of server clusters in the cluster category information and the device information and performance information under the corresponding chip environment, and summarize and output multi-core heterogeneous cluster information.

[0008] Preferably, the container establishment module includes: A feature extraction submodule, used to extract features from the multi-core heterogeneous cluster information and construct a first feature set based on the extracted features; A container construction submodule, used to select at least two preset container technologies matching the first feature set from a method database in combination with a preset feature-technology mapping table, and to establish multiple containers based on the preset container technologies; The network construction submodule is used to establish a mapping network between each container and each server cluster.

[0009] Preferably, the intercommunication establishment module includes: An interface information acquisition submodule, used to acquire the interface information of each server cluster in the multi-core heterogeneous cluster information and output the cluster interface information; The interface unification submodule is used to configure the API gateway of each server cluster based on the cluster interface information and establish a unified API interface; A middleware establishment submodule is used to select middleware matching the middleware function requirement information from a method database based on the obtained middleware function requirement information; The intercommunication information establishment submodule is used to output cluster intercommunication information based on the unified API interface and middleware.

[0010] Preferably, the middleware establishes a submodule, including: A demand acquisition unit, used to acquire functional demand information for mutual communication and data exchange between server clusters; A middleware matching unit, configured to select a matching middleware from a method database based on the functional requirement information; The middleware configuration and optimization unit is used to establish the architecture of the middleware, and develop and integrate the middleware based on the architecture. At the same time, the middleware is performance tested, and the performance optimization and function expansion of the middleware are performed in combination with the performance test results and the function requirement information.

[0011] Preferably, the multi-core resource scheduling module includes: An application demand acquisition submodule, configured to output first demand information based on demand information of each application in the system acquired in real time; The management requirement acquisition submodule is used to acquire the unified management requirement information of the system and output the second requirement information; A demand information generating submodule, configured to aggregate the first demand information and the second demand information and output cluster demand information; The resource data acquisition submodule is used to obtain the resource data and corresponding resource status of each server cluster in the system in real time, and output cluster resource information; A view construction submodule, used to construct a real-time demand-resource view based on the cluster demand information and cluster resource information and using a preset view construction method; A demand-resource matching submodule, used to match demand information and resource information in the real-time demand-resource view, and output a demand-resource matching result; A factor acquisition submodule, used to acquire a first screening factor matching the demand-resource matching result in combination with a preset result-factor comparison table; A strategy-method matching submodule, used for selecting a resource scheduling strategy that meets a preset screening condition and a resource scheduling method corresponding to the resource scheduling strategy from a strategy-method database based on the first screening factor; A task allocation submodule is used to allocate various tasks in the application to the node resources under the corresponding server cluster based on the resource scheduling strategy and resource scheduling method and in combination with the cluster intercommunication information, and output resource scheduling process data; The indicator acquisition submodule is used to select and obtain corresponding performance evaluation indicators in the indicator database based on the obtained indicator selection instruction; An effect analysis submodule, used to perform effect analysis on the resource scheduling process data based on the performance evaluation index and output a first result; An optimization analysis submodule, used to perform optimization analysis on the first result in combination with the optimization index selected in the index database, and output the optimization analysis result; The strategy-method optimization submodule is used to select a preset optimization method that matches the optimization analysis result from the method database, and optimize the resource scheduling strategy and resource scheduling method based on the preset optimization method.

[0012] Preferably, the effect analysis submodule includes: An indicator division unit, used for dividing the performance evaluation indicators to obtain a main indicator set and a secondary indicator set; A weight acquisition unit, used to acquire a first weight of each performance evaluation indicator in the main indicator set and a second weight of each performance evaluation indicator in the secondary indicator set, and output an indicator-weight comparison table; An indicator relationship acquisition unit, used to acquire a first relationship between each performance evaluation indicator in the main indicator set, a second relationship between each performance evaluation indicator in the secondary indicator set, and a third relationship between each performance evaluation indicator in the main indicator set and the secondary indicator set, and construct an indicator relationship graph; A data parsing unit, used to parse the resource scheduling process data to obtain data to be evaluated; An effect analysis unit, used to perform effect analysis on the data to be evaluated based on the main indicator set, the sub-indicator set, the indicator-weight comparison table and the indicator relationship diagram, to obtain an effect analysis result; ; in, Indicates the comprehensive effect analysis value based on the main indicator set and the sub-indicator set; represents the exponential function; Indicates the mutual correlation coefficient between the main index set and the secondary index set; Indicates the calculation weight of the effect analysis results corresponding to the main indicator set; Represents the total number of performance evaluation indicators in the main indicator set; represents the i-th performance evaluation indicator in the main indicator set; represents the data value to be evaluated corresponding to the i-th performance evaluation indicator in the main indicator set; represents a first weight corresponding to the i-th performance evaluation indicator in the main indicator set; Indicates the influence factors of the remaining n-1 performance evaluation indicators in the main indicator set except the i-th performance evaluation indicator on the i-th performance evaluation indicator; Indicates the calculation weight of the effect analysis result corresponding to the sub-indicator set; Represents the total number of performance evaluation indicators in the sub-indicator set; represents the jth performance evaluation indicator in the sub-indicator set; represents the data value to be evaluated corresponding to the jth performance evaluation indicator in the secondary indicator set; represents a second weight corresponding to the jth performance evaluation indicator in the secondary indicator set; It represents the influence factors of the remaining m-1 performance evaluation indicators in the secondary indicator set except the j-th performance evaluation indicator on the j-th performance evaluation indicator.

[0013] On the other hand, the present invention also provides a heterogeneous multi-core coexistence management method for cloud computing, comprising: Step 1: Obtain the device information and performance information of multiple server clusters with different cores in the system, and summarize and output the multi-core heterogeneous cluster information; Step 2: Based on the multi-core heterogeneous cluster information, multiple containers are established using a preset container technology, and a mapping network between each container and each server cluster is constructed; Step 3: Based on the multi-core heterogeneous cluster information and the mapping network, a unified API interface and middleware are established to output cluster intercommunication information; Step 4: Obtain the application demand information of the system and the resource information of each server cluster, output the real-time demand-resource view, and obtain the resource scheduling strategy and resource scheduling method that match the real-time demand-resource view in the strategy-method database. Based on the resource scheduling strategy and resource scheduling method, and in combination with the cluster intercommunication information, assign the tasks in the application to the nodes under the corresponding server cluster.

[0014] The present invention provides a heterogeneous multi-core coexistence management system and method for cloud computing, which realizes effective management and resource scheduling of multi-core heterogeneous clusters to meet the needs of applications through the collaborative work of cluster information acquisition, container establishment, intercommunication establishment and multi-core resource scheduling modules. The present invention realizes efficient utilization of resources and intelligent scheduling of tasks, improves system performance and stability, simplifies the complexity of management and deployment, and improves overall operation and maintenance efficiency and user experience. Through the optimization of containerization technology and resource scheduling strategies, the heterogeneous multi-core cloud platform can better adapt to the needs of different application scenarios and improve the flexibility and scalability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 It is a schematic diagram of a framework of a heterogeneous multi-core coexistence management system for cloud computing provided by an embodiment of the present invention; Figure 2 It is a flow chart of a heterogeneous multi-core coexistence management method for cloud computing provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] like Figure 1 As shown, an embodiment of the present invention provides a heterogeneous multi-core coexistence management system for cloud computing, including: The cluster information acquisition module is used to obtain the device information and performance information of multiple server clusters with different cores in the system, and summarize and output the multi-core heterogeneous cluster information; The container establishment module is used to establish multiple containers based on the multi-core heterogeneous cluster information using the preset container technology, and to build a mapping network between each container and each server cluster; The intercommunication establishment module is used to establish a unified API interface and middleware based on multi-core heterogeneous cluster information and mapping network, and output cluster intercommunication information; The multi-core resource scheduling module is used to obtain the application demand information of the system and the resource information of each server cluster, output the real-time demand-resource view, and obtain the resource scheduling strategy and resource scheduling method that match the real-time demand-resource view in the strategy-method database. Based on the resource scheduling strategy and resource scheduling method, and combined with the cluster intercommunication information, the tasks in the application are allocated to the nodes under the corresponding server cluster.

[0019] In this embodiment, device information: hardware device related information of multiple server clusters with different cores, such as model, configuration, status, etc.; In this embodiment, performance information: performance parameters of multiple server clusters, such as CPU utilization, memory utilization, network bandwidth, etc.; In this embodiment, the multi-core heterogeneous cluster information: summarizes the device information and performance information of multiple server clusters with different cores in the system for overall management and scheduling; In this embodiment, the preset container technology refers to a pre-set containerization technology used to implement isolation and deployment of applications; In this embodiment, the mapping network: that is, the mapping relationship between the container and each server cluster, ensures that the container can correctly access the required resources; In this embodiment, the unified API interface: provides a unified interface so that different types of server clusters can communicate and operate with the system in the same way; In this embodiment, the middleware is used to connect different server clusters and containers as a communication bridge, providing data transmission and processing functions, including but not limited to message queues (such as RabbitMQ, Kafka), message middleware (such as ActiveMQ, RabbitMQ), data middleware (such as Redis), ESB (Enterprise Service Bus), etc. In this embodiment, cluster intercommunication information includes information about communication methods and rules between server clusters, such as a unified API interface and middleware to ensure that clusters can communicate and collaborate with each other; In this embodiment, application requirement information: describes the requirements of the application program in the system, including resource requirements, operating environment, etc.; In this embodiment, resource information: describes the resource status of each server cluster, including available CPU, memory, storage and other resources; In this embodiment, real-time demand-resource view: a real-time view generated based on application demand information and resource information, used for resource scheduling decisions; In this embodiment, the strategy-method database: stores various resource scheduling strategies and methods, and is used to select appropriate resource scheduling strategies and methods according to the real-time demand-resource view; In this embodiment, resource scheduling policy: describes the decision rules and algorithms for resource scheduling, which are used to allocate resources according to real-time demand; In this embodiment, resource scheduling method: a specific resource scheduling implementation method, including task allocation, load balancing and other methods.

[0020] The implementation principle and beneficial effects of this embodiment: The present invention realizes effective management and resource scheduling of multi-core heterogeneous clusters to meet the needs of applications through the collaborative work of cluster information acquisition, container establishment, intercommunication establishment and multi-core resource scheduling modules. The present invention realizes efficient utilization of resources and intelligent scheduling of tasks, improves system performance and stability, simplifies the complexity of management and deployment, and improves overall operation and maintenance efficiency and user experience. Through the optimization of containerization technology and resource scheduling strategies, the heterogeneous multi-core cloud platform can better adapt to the needs of different application scenarios and improve the flexibility and scalability of the system.

[0021] An embodiment of the present invention provides a heterogeneous multi-core coexistence management system for cloud computing, a cluster information acquisition module, including: The cluster category information acquisition submodule is used to obtain the type information of each server cluster in the system and output the cluster category information; The chip-device-performance information acquisition submodule is used to obtain the chip environment information corresponding to each category of server clusters in the cluster category information, as well as the device information and performance information under the corresponding chip environment, and summarize and output multi-core heterogeneous cluster information.

[0022] In this embodiment, the category information: describes the category or type of each server cluster in the system, for example, a high-performance computing cluster, a distributed storage cluster, a virtualization cluster, etc.; In this embodiment, cluster category information refers to information for classifying and identifying different types of server clusters, and is used to distinguish clusters of various categories; In this embodiment, chip environment information: environmental information of a specific chip or processor, including supported instruction sets, architectural features, performance parameters, etc. Chip environment information is very important for system resource management and scheduling, because different chip environments may require different optimization strategies and resource allocation methods.

[0023] The implementation principle and beneficial effects of this embodiment: Through the collaborative work of the cluster category information acquisition submodule and the chip-device-performance information acquisition submodule, the system can accurately identify various types of server clusters and obtain their corresponding chip environment information and performance data, thereby constructing multi-core heterogeneous cluster information. The present invention can achieve more refined resource management and scheduling based on different types of server clusters and chip environment information, improve the performance and efficiency of the system, and at the same time, through accurate grasp of cluster categories and chip environments, better optimize resource allocation, improve the operating efficiency of applications, and reduce energy consumption and costs. This refined management and scheduling capability can bring better performance and user experience to cloud computing environments.

[0024] An embodiment of the present invention provides a heterogeneous multi-core coexistence management system for cloud computing, a container establishment module, including: A feature extraction submodule, used to extract features from multi-core heterogeneous cluster information and construct a first feature set based on the extracted features; A container construction submodule, for selecting at least two preset container technologies matching the first feature set from a method database in combination with a preset feature-technology mapping table, and establishing a plurality of containers based on the preset container technologies; The network construction submodule is used to establish a mapping network between each container and each server cluster.

[0025] In this embodiment, the first feature set is a set of key features extracted from the multi-core heterogeneous cluster information, which is used as a reference for subsequent container construction; In this embodiment, the preset feature-technology mapping table: describes the mapping relationship between features and container technologies, is pre-set, and can help the system select an appropriate container technology to meet the requirements of specific features; In this embodiment, the method database: stores various methods and strategies used in the system, including container construction methods, network construction methods, etc.; In this embodiment, container: a virtualization technology used to encapsulate an application and all its runtime environments and dependencies to achieve isolation and deployment of the application, for example, container technologies such as Docker and Kubernetes.

[0026] The implementation principle and beneficial effects of this embodiment: The present invention extracts key features through the feature extraction submodule, the container construction submodule selects the appropriate container technology, and the network construction submodule establishes the connection between the container and the server cluster, and the system can realize containerized deployment and management. The present invention can improve the flexibility and portability of the system, simplify the deployment and maintenance of applications, and select the appropriate container technology by extracting the data features. The system can better adapt to different application scenarios and requirements, improve the efficiency and performance of the system, and at the same time, the network construction submodule ensures smooth communication between the container and the server cluster, further improving the stability and reliability of the system.

[0027] An embodiment of the present invention provides a heterogeneous multi-core coexistence management system for cloud computing, and an intercommunication establishment module, including: The interface information acquisition submodule is used to acquire the interface information of each server cluster in the multi-core heterogeneous cluster information and output the cluster interface information; The interface unification submodule is used to configure the API gateway of each server cluster based on the cluster interface information and establish a unified API interface; The middleware establishment submodule is used to select the middleware matching the middleware function requirement information from the method database based on the obtained middleware function requirement information; The intercommunication information establishment submodule is used to output cluster intercommunication information based on a unified API interface and middleware.

[0028] In this embodiment, cluster interface information refers to the interface information of each server cluster, including communication protocol, data format, access permission, etc., which is used to describe the communication method and rules between servers; In this embodiment, API Gateway: a server for managing, monitoring and protecting APIs, and for uniformly managing interface access between different platforms or server clusters; In this embodiment, the middleware functional requirement information refers to the description of the system's functional requirements for the middleware, including requirements in data processing, message transmission, security, etc. Selecting appropriate middleware according to these requirements can better support the system's functions and performance.

[0029] The implementation principle and beneficial effects of this embodiment: Through the collaborative work of the interface information acquisition submodule, the interface unification submodule, the middleware establishment submodule and the intercommunication information establishment submodule, the system can uniformly manage the interfaces of each server cluster, select appropriate middleware, establish intercommunication channels, and realize the mutual communication and collaboration of various modules within the system. The present invention can realize the standardization and unified management of interfaces between server clusters, simplify the communication and integration process of the system, improve the maintainability and scalability of the system, and can also support the system function requirements by selecting appropriate middleware, improve the performance and efficiency of the system, and reduce the complexity of system integration and development. The design of this intercommunication establishment module can strengthen the collaboration between various modules within the system and improve the overall operation efficiency and stability of the system.

[0030] An embodiment of the present invention provides a heterogeneous multi-core coexistence management system for cloud computing, wherein a middleware establishes a submodule, including: A demand acquisition unit, used to acquire functional demand information for mutual communication and data exchange between server clusters; A middleware matching unit, used for selecting a matching middleware in a method database based on the functional requirement information; The middleware configuration and optimization unit is used to establish the middleware architecture and develop and integrate the middleware based on the architecture. At the same time, it performs performance testing on the middleware and optimizes the performance and expands the functions of the middleware based on the performance test results and functional requirement information.

[0031] In this embodiment, functional requirement information: the functional requirement description of the system for mutual communication and data exchange between different server clusters, including but not limited to data exchange, communication protocol, security requirements, etc. Such requirement information can guide the selection and development of middleware; In this embodiment, architecture refers to the overall design structure of the middleware, including components, interfaces, interaction methods, etc. A good architecture can support the functional requirements of the system and provide good performance; In this embodiment, performance test results: results obtained after the performance test of the middleware, including but not limited to test results under indicators such as response time, throughput, and concurrency performance; In this embodiment, performance optimization: optimization work based on performance test results is aimed at improving the performance of the middleware, reducing resource usage, and improving the response speed and stability of the system; In this embodiment, function expansion refers to expanding the middleware based on the original functions to add new functions or support more application scenarios.

[0032] The implementation principle and beneficial effects of this embodiment: Through the collaborative work of the demand acquisition unit, the middleware matching unit and the middleware configuration and optimization unit, the system can select appropriate middleware according to functional requirements, design and optimize the middleware architecture to meet the communication and data exchange requirements of the system. The present invention can select the most suitable middleware according to system requirements, improve the performance and stability of the middleware through performance testing and optimization, and at the same time, expand the functions of the middleware through functional expansion, so that the heterogeneous multi-core cloud platform can cope with more application scenarios and requirements.

[0033] An embodiment of the present invention provides a heterogeneous multi-core coexistence management system for cloud computing, a multi-core resource scheduling module, including: An application demand acquisition submodule, configured to output first demand information based on demand information of each application in the system acquired in real time; The management requirement acquisition submodule is used to acquire the unified management requirement information of the system and output the second requirement information; A demand information generating submodule, used to aggregate the first demand information and the second demand information and output cluster demand information; The resource data acquisition submodule is used to obtain the resource data and corresponding resource status of each server cluster in the system in real time, and output cluster resource information; A view building submodule is used to build a real-time demand-resource view based on cluster demand information and cluster resource information and using a preset view building method; The demand-resource matching submodule is used to match the demand information and resource information in the real-time demand-resource view and output the demand-resource matching result; A factor acquisition submodule, used to obtain a first screening factor matching the demand-resource matching result in combination with a preset result-factor comparison table; A strategy-method matching submodule, used to select a resource scheduling strategy that meets a preset screening condition and a resource scheduling method corresponding to the resource scheduling strategy from a strategy-method database based on a first screening factor; The task allocation submodule is used to allocate various tasks in the application to the node resources under the corresponding server cluster based on the resource scheduling strategy and resource scheduling method, and in combination with the cluster intercommunication information, and output the resource scheduling process data; The indicator acquisition submodule is used to select and obtain corresponding performance evaluation indicators in the indicator database based on the obtained indicator selection instruction; The effect analysis submodule is used to perform effect analysis on the resource scheduling process data based on the performance evaluation index and output a first result; An optimization analysis submodule, used to perform optimization analysis on the first result in combination with the optimization index selected in the index database, and output the optimization analysis result; The strategy-method optimization submodule is used to select a preset optimization method that matches the optimization analysis result from the method database, and optimize the resource scheduling strategy and resource scheduling method based on the preset optimization method.

[0034] In this embodiment, the first demand information refers to the real-time demand information of each application in the system, such as the demand for computing resources, storage resources, etc.; In this embodiment, the second requirement information: unified management requirement information of the system, such as requirements on security, reliability, etc.; In this embodiment, cluster demand information: cluster-level demand information integrating application demand and management demand; In this embodiment, cluster resource information: includes resource data and status information of each server cluster in the system, including but not limited to CPU utilization, memory usage, and load status of each server cluster; In this embodiment, the preset view construction method: the method for constructing the real-time demand-resource view is preset; In this embodiment, demand-resource matching result: the result of matching real-time demand and resource information, matching the application demand with available resources to determine the best resource allocation solution; In this embodiment, the preset result-factor comparison table is a predefined table containing a mapping relationship between the demand-resource matching result and the first screening factor; In this embodiment, the first screening factor: a screening factor of a resource scheduling strategy and a resource scheduling method selected according to a demand-resource matching result; In this embodiment, the preset screening conditions: preset conditions for screening resource scheduling strategies and methods; In this embodiment, resource scheduling process data refers to data generated during the resource scheduling process, including task allocation, resource utilization, and other information; In this embodiment, the indicator selection instruction: an instruction for selecting a performance evaluation indicator; In this embodiment, the indicator database: a database storing various performance evaluation indicators and optimization indicators; In this embodiment, the performance evaluation index: an index used to evaluate the resource scheduling effect, such as response time, throughput, etc.; In this embodiment, the first result: that is, the result of evaluating the resource scheduling effect based on the performance evaluation index; In this embodiment, optimization index: an index used to guide resource scheduling optimization, such as cost reduction, performance improvement, etc.; In this embodiment, the optimization analysis result is the result obtained after optimizing the resource scheduling effect; In this embodiment, the preset optimization method is a predefined method or strategy for optimizing resource scheduling.

[0035] The implementation principle and beneficial effects of this embodiment: The present invention can make intelligent resource scheduling decisions based on application requirements and management requirements in combination with resource conditions through the collaborative work of each submodule under the multi-core resource scheduling module, and continuously optimize the scheduling strategy and method through performance evaluation and optimization analysis. The present invention can perform intelligent scheduling according to real-time requirements and resource conditions, improve resource utilization and system performance, and optimize according to the evaluation results to make resource scheduling more efficient and flexible, thereby improving the overall resource utilization, operating efficiency and user experience of the system.

[0036] An embodiment of the present invention provides a heterogeneous multi-core coexistence management system for cloud computing, and an effect analysis submodule includes: An indicator division unit, used for dividing the performance evaluation indicators into a main indicator set and a sub-indicator set; A weight acquisition unit, used to acquire a first weight of each performance evaluation indicator in the main indicator set and a second weight of each performance evaluation indicator in the secondary indicator set, and output an indicator-weight comparison table; An indicator relationship acquisition unit, used to acquire a first relationship between each performance evaluation indicator in the main indicator set, a second relationship between each performance evaluation indicator in the secondary indicator set, and a third relationship between each performance evaluation indicator in the main indicator set and the secondary indicator set, and construct an indicator relationship diagram; A data analysis unit, used to analyze the resource scheduling process data to obtain the data to be evaluated; The effect analysis unit is used to perform effect analysis on the evaluation data based on the main indicator set, the sub-indicator set, the indicator-weight comparison table and the indicator relationship diagram to obtain the effect analysis result; ; in, It indicates the comprehensive effect analysis value based on the main indicator set and the sub-indicator set; represents the exponential function; Indicates the mutual correlation coefficient between the main index set and the secondary index set; Indicates the calculation weight of the effect analysis results corresponding to the main indicator set; Represents the total number of performance evaluation indicators in the main indicator set; represents the i-th performance evaluation indicator in the main indicator set; represents the data value to be evaluated corresponding to the i-th performance evaluation indicator in the main indicator set; represents the first weight corresponding to the i-th performance evaluation indicator in the main indicator set; It represents the influence factor of the remaining n-1 performance evaluation indicators in the main indicator set except the i-th performance evaluation indicator on the i-th performance evaluation indicator; Indicates the calculation weight of the effect analysis result corresponding to the sub-indicator set; Indicates the total number of performance evaluation indicators in the sub-indicator set; represents the jth performance evaluation indicator in the sub-indicator set; represents the data value to be evaluated corresponding to the jth performance evaluation indicator in the sub-indicator set; represents the second weight corresponding to the jth performance evaluation indicator in the sub-indicator set; It represents the influence factor of the remaining m-1 performance evaluation indicators in the sub-indicator set except the j-th performance evaluation indicator on the j-th performance evaluation indicator.

[0037] In this embodiment, the main indicator set is a key indicator set obtained by dividing the performance evaluation indicators, which is usually used to measure important indicators of system performance, such as response time, resource utilization, throughput, etc. In this embodiment, the secondary indicator set: corresponding to the primary indicator set, the secondary indicator set is a set of indicators in the performance evaluation indicators that are secondary but still have a certain influence, such as stability, reliability, cost, etc. It should be noted that the division rules of the primary indicator set and the secondary indicator set can be divided according to the specific performance requirements in actual situations; In this embodiment, the first weight is a weight coefficient for each performance evaluation indicator in the main indicator set. For example, the response time may be assigned a higher weight because it has a more direct impact on the user experience. In this embodiment, the second weight is: the weight coefficient for each performance evaluation indicator in the secondary indicator set. These weights are usually relatively low, but still have a certain impact on the overall effect of the system; In this embodiment, the indicator-weight comparison table is used to compare each performance evaluation indicator with its corresponding weight for use in the subsequent effect analysis process; In this embodiment, the first relationship: that is, the correlation relationship between the performance evaluation indicators within the main indicator set; In this embodiment, the second relationship: that is, the correlation relationship between the performance evaluation indicators within the secondary indicator set; In this embodiment, the third relationship: that is, the correlation relationship between each performance evaluation indicator between the primary indicator set and the secondary indicator set; In this embodiment, the indicator relationship diagram: a graphical representation constructed based on the main indicator set, the secondary indicator set and the relationship between them, used to show the correlation and influence degree between the various performance evaluation indicators; In this embodiment, the data to be evaluated refers to the data that needs to be analyzed for effect after being processed by the data analysis unit, including the data generated in the resource scheduling process that corresponds to various performance evaluation indicators and needs to be evaluated and analyzed.

[0038] Implementation principle and beneficial effects of this embodiment: This method evaluates the effect of resource scheduling by weight allocation of the main indicator set and the secondary indicator set, modeling of the indicator relationship and data analysis, and combining the formula to calculate the comprehensive effect analysis value E. The present invention can help system administrators or decision makers to understand the effect of resource scheduling more comprehensively. By quantitatively analyzing the relationship and weight between indicators, it can comprehensively consider the importance and mutual relationship of each indicator, improve the accuracy of the resource scheduling effect analysis results, and provide data support for the optimization of subsequent scheduling strategies and scheduling methods, thereby improving the resource utilization efficiency of the system.

[0039] like Figure 2 As shown, an embodiment of the present invention provides a heterogeneous multi-core coexistence management method for cloud computing, including: Step 1: Obtain the device information and performance information of multiple server clusters with different cores in the system, and summarize and output the multi-core heterogeneous cluster information; Step 2: Based on the multi-core heterogeneous cluster information, use the preset container technology to establish multiple containers, and build a mapping network between each container and each server cluster; Step 3: Based on the multi-core heterogeneous cluster information and mapping network, a unified API interface and middleware are established to output cluster intercommunication information; Step 4: Obtain the application demand information of the system and the resource information of each server cluster, output the real-time demand-resource view, and obtain the resource scheduling strategy and resource scheduling method that match the real-time demand-resource view in the strategy-method database. Based on the resource scheduling strategy and resource scheduling method, and combined with the cluster intercommunication information, the tasks in the application are allocated to the nodes under the corresponding server cluster.

[0040] The implementation principle and beneficial effects of this embodiment: The present invention realizes effective management and resource scheduling of multi-core heterogeneous clusters to meet the needs of applications through the collaborative work of cluster information acquisition, container establishment, intercommunication establishment and multi-core resource scheduling modules. The present invention realizes efficient utilization of resources and intelligent scheduling of tasks, improves system performance and stability, simplifies the complexity of management and deployment, and improves overall operation and maintenance efficiency and user experience. Through the optimization of containerization technology and resource scheduling strategies, the heterogeneous multi-core cloud platform can better adapt to the needs of different application scenarios and improve the flexibility and scalability of the system.

[0041] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A heterogeneous multi-core coexistence management system for cloud computing, characterized in that: include: The cluster information acquisition module is used to obtain the device information and performance information of multiple server clusters with different cores in the system, and summarize and output the multi-core heterogeneous cluster information; A container establishment module, used to establish multiple containers based on the multi-core heterogeneous cluster information using a preset container technology, and to construct a mapping network between each of the containers and each server cluster; An intercommunication establishment module, used to establish a unified API interface and middleware based on the multi-core heterogeneous cluster information and the mapping network, and output cluster intercommunication information; The multi-core resource scheduling module is used to obtain the application demand information of the system and the resource information of each server cluster, output the real-time demand-resource view, and obtain the resource scheduling strategy and resource scheduling method that match the real-time demand-resource view in the strategy-method database. Based on the resource scheduling strategy and resource scheduling method, and in combination with the cluster intercommunication information, the tasks in the application are allocated to the nodes under the corresponding server cluster.

2. A heterogeneous multi-core coexistence management system for cloud computing according to claim 1, characterized in that: The cluster information acquisition module includes: The cluster category information acquisition submodule is used to obtain the type information of each server cluster in the system and output the cluster category information; The chip-device-performance information acquisition submodule is used to obtain the chip environment information corresponding to each category of server clusters in the cluster category information and the device information and performance information under the corresponding chip environment, and summarize and output multi-core heterogeneous cluster information.

3. The heterogeneous multi-core coexistence management system for cloud computing according to claim 1, characterized in that: The container establishment module includes: A feature extraction submodule, used to extract features from the multi-core heterogeneous cluster information and construct a first feature set based on the extracted features; A container construction submodule, used to select at least two preset container technologies matching the first feature set from a method database in combination with a preset feature-technology mapping table, and to establish multiple containers based on the preset container technologies; The network construction submodule is used to establish a mapping network between each container and each server cluster.

4. The heterogeneous multi-core coexistence management system for cloud computing according to claim 1, characterized in that: The intercommunication establishment module includes: An interface information acquisition submodule, used to acquire the interface information of each server cluster in the multi-core heterogeneous cluster information and output the cluster interface information; The interface unification submodule is used to configure the API gateway of each server cluster based on the cluster interface information and establish a unified API interface; A middleware establishment submodule is used to select middleware matching the middleware function requirement information from a method database based on the obtained middleware function requirement information; The intercommunication information establishment submodule is used to output cluster intercommunication information based on the unified API interface and middleware.

5. A heterogeneous multi-core coexistence management system for cloud computing according to claim 4, characterized in that: The middleware establishes a submodule, including: A demand acquisition unit, used to acquire functional demand information for mutual communication and data exchange between server clusters; A middleware matching unit, configured to select a matching middleware from a method database based on the functional requirement information; The middleware configuration and optimization unit is used to establish the architecture of the middleware, and develop and integrate the middleware based on the architecture. At the same time, the middleware is performance tested, and the performance optimization and function expansion of the middleware are performed in combination with the performance test results and the function requirement information.

6. The heterogeneous multi-core coexistence management system for cloud computing according to claim 1, characterized in that: The multi-core resource scheduling module includes: An application demand acquisition submodule, configured to output first demand information based on demand information of each application in the system acquired in real time; The management requirement acquisition submodule is used to acquire the unified management requirement information of the system and output the second requirement information; A demand information generating submodule, configured to aggregate the first demand information and the second demand information and output cluster demand information; The resource data acquisition submodule is used to obtain the resource data and corresponding resource status of each server cluster in the system in real time, and output cluster resource information; A view construction submodule, used to construct a real-time demand-resource view based on the cluster demand information and cluster resource information and using a preset view construction method; A demand-resource matching submodule, used to match demand information and resource information in the real-time demand-resource view, and output a demand-resource matching result; A factor acquisition submodule, used to acquire a first screening factor matching the demand-resource matching result in combination with a preset result-factor comparison table; A strategy-method matching submodule, used for selecting a resource scheduling strategy that meets a preset screening condition and a resource scheduling method corresponding to the resource scheduling strategy from a strategy-method database based on the first screening factor; A task allocation submodule is used to allocate various tasks in the application to the node resources under the corresponding server cluster based on the resource scheduling strategy and resource scheduling method and in combination with the cluster intercommunication information, and output resource scheduling process data; The indicator acquisition submodule is used to select and obtain corresponding performance evaluation indicators in the indicator database based on the obtained indicator selection instruction; An effect analysis submodule, used to perform effect analysis on the resource scheduling process data based on the performance evaluation index and output a first result; An optimization analysis submodule, used to perform optimization analysis on the first result in combination with the optimization index selected in the index database, and output the optimization analysis result; The strategy-method optimization submodule is used to select a preset optimization method that matches the optimization analysis result from the method database, and optimize the resource scheduling strategy and resource scheduling method based on the preset optimization method.

7. The heterogeneous multi-core coexistence management system for cloud computing according to claim 6, characterized in that: The effect analysis submodule includes: An indicator division unit, used for dividing the performance evaluation indicators to obtain a main indicator set and a secondary indicator set; A weight acquisition unit, used to acquire a first weight of each performance evaluation indicator in the main indicator set and a second weight of each performance evaluation indicator in the secondary indicator set, and output an indicator-weight comparison table; An indicator relationship acquisition unit, used to acquire a first relationship between each performance evaluation indicator in the main indicator set, a second relationship between each performance evaluation indicator in the secondary indicator set, and a third relationship between each performance evaluation indicator in the main indicator set and the secondary indicator set, and construct an indicator relationship graph; A data parsing unit, used to parse the resource scheduling process data to obtain data to be evaluated; An effect analysis unit, used to perform effect analysis on the data to be evaluated based on the main indicator set, the sub-indicator set, the indicator-weight comparison table and the indicator relationship diagram, to obtain an effect analysis result; ； in, Indicates the comprehensive effect analysis value based on the main indicator set and the sub-indicator set; represents the exponential function; Indicates the mutual correlation coefficient between the main index set and the secondary index set; Indicates the calculation weight of the effect analysis results corresponding to the main indicator set; Represents the total number of performance evaluation indicators in the main indicator set; represents the i-th performance evaluation indicator in the main indicator set; represents the data value to be evaluated corresponding to the i-th performance evaluation indicator in the main indicator set; represents a first weight corresponding to the i-th performance evaluation indicator in the main indicator set; Indicates the influence factors of the remaining n-1 performance evaluation indicators in the main indicator set except the i-th performance evaluation indicator on the i-th performance evaluation indicator; Indicates the calculation weight of the effect analysis result corresponding to the sub-indicator set; Represents the total number of performance evaluation indicators in the sub-indicator set; represents the jth performance evaluation indicator in the sub-indicator set; represents the data value to be evaluated corresponding to the jth performance evaluation indicator in the secondary indicator set; represents a second weight corresponding to the jth performance evaluation indicator in the secondary indicator set; It represents the influence factors of the remaining m-1 performance evaluation indicators in the secondary indicator set except the j-th performance evaluation indicator on the j-th performance evaluation indicator.

8. A heterogeneous multi-core coexistence management method for cloud computing, characterized in that: include: Step 1: Obtain the device information and performance information of multiple server clusters with different cores in the system, and summarize and output the multi-core heterogeneous cluster information; Step 2: Based on the multi-core heterogeneous cluster information, multiple containers are established using a preset container technology, and a mapping network between each container and each server cluster is constructed; Step 3: Based on the multi-core heterogeneous cluster information and the mapping network, a unified API interface and middleware are established to output cluster intercommunication information; Step 4: Obtain the application demand information of the system and the resource information of each server cluster, output the real-time demand-resource view, and obtain the resource scheduling strategy and resource scheduling method that match the real-time demand-resource view in the strategy-method database. Based on the resource scheduling strategy and resource scheduling method, and in combination with the cluster intercommunication information, assign the tasks in the application to the nodes under the corresponding server cluster.

Images