Resource adaptive management system and method based on multi-cloud environment

Abstract

The invention relates to the technical field of cloud service management, in particular to a resource self-adaptive management system and method based on a multi-cloud environment. The system comprises a resource acquisition module used for acquiring dynamic data of a plurality of cloud platforms; the statistical analysis module receives the dynamic data and is used for carrying out statistical analysis on the current use situation and trend of the cloud resources according to the service requirements and the load changes, outputting a resource analysis result and predicting the capacity expansion requirements of the cloud resources in the future; the judgment module is used for judging whether the current resource configuration is appropriate; and the scheduling optimization module is used for dynamically configuring, optimizing and adjusting the resources according to the judgment result of the judgment module. According to a resource monitoring and analysis result, resource distribution is automatically adjusted, dynamic allocation and optimization management are performed on multi-cloud resources in an automatic and intelligent manner, the granularity size can be adjusted according to needs, and the resource utilization efficiency and management flexibility of the system are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of cloud business management, and in particular to a system and method for adaptive resource management based on a multi-cloud environment. Background Art

[0002] In cloud business management, the size of granularity depends on multiple factors, including business needs, application architecture, technology stack, and operation and maintenance capabilities. Generally speaking, a coarser granularity means a larger management unit, which simplifies the management process and reduces the complexity of operation and maintenance while sacrificing certain flexibility and scalability. On the contrary, a finer granularity means a smaller management unit, which is more flexible and scalable, but increases the complexity and cost of operation and maintenance. In the prior art, the granularity of cloud business management cannot be weighed and adjusted according to specific circumstances. In different application scenarios, each service or application is limited by a management unit with a set granularity. The granularity cannot be adjusted, and adaptive management cannot be performed independently and efficiently. Summary of the invention

[0003] The purpose of the present invention is to provide a system based on adaptive resource management in a multi-cloud environment to solve the above technical problems;

[0004] The present invention also aims to provide a method for adaptive resource management in a multi-cloud environment to solve the above technical problems.

[0005] The technical problem solved by the present invention can be achieved by adopting the following technical solutions:

[0006] A system for adaptive resource management based on a multi-cloud environment, comprising:

[0007] Resource collection module, used to collect dynamic data from multiple cloud platforms;

[0008] A statistical analysis module, connected to the resource acquisition module to receive the dynamic data, is used to statistically analyze the current status and trend of cloud resource usage according to business needs and load changes, output resource analysis results, and predict future demand for cloud resource expansion;

[0009] A judgment module, connected to the statistical analysis module, for judging whether the current resource configuration is appropriate;

[0010] A scheduling optimization module is connected to the judgment module and is used to dynamically configure, optimize and adjust resources according to the judgment result of the judgment module. The scheduling optimization module includes a cost optimization unit for performing cost optimization, a performance optimization unit for performing performance optimization, a host scheduling unit for performing designated host scheduling, a protocol scheduling unit for performing task scheduling according to a service level agreement, a processor resource scheduling unit for scheduling processor resources, and a processor load scheduling unit for scheduling processor load, which are respectively connected to the statistical analysis module.

[0011] Preferably, the dynamic data collected by the resource collection module is performance data, resource utilization and operation status information collected from the cloud platform, and the information sources of the dynamic data include performance data, alarm data, logs, network data and environmental information.

[0012] Preferably, the statistical analysis module acquires the dynamic data, and a resource analysis unit for analyzing resource usage is provided in the statistical analysis module. The resource analysis unit uses a machine learning model or a rule engine to statistically analyze the current status and trend of cloud resource usage and outputs the resource analysis result. The business demand and load changes based on which the statistical analysis module predicts future cloud resource expansion needs include at least application response time, business traffic changes, storage space changes, link quality changes, host load changes, system log alarms, and user access details.

[0013] The judgment module is used to judge whether the current resource configuration meets the business needs and load changes. If the current resource configuration meets the business needs and load changes, the current resource configuration is maintained. If the current resource configuration does not meet the business needs and load changes, the scheduling optimization module dynamically configures, optimizes and adjusts the resources.

[0014] It also includes an interaction module connected to the scheduling optimization module, and the interaction module has a graphical interface and an interaction interface for interaction.

[0015] Preferably, the cost optimization unit evaluates the actual business volume of the work task according to the user application and allocates cloud resources corresponding to the actual business volume.

[0016] Preferably, the performance optimization unit prioritizes resource nodes according to their computing capabilities and delays, and allocates work tasks to resource nodes whose priorities correspond to the business levels of the work tasks.

[0017] Preferably, the host scheduling unit allocates the work tasks to designated cloud hosts or cloud areas.

[0018] Preferably, the protocol scheduling unit allocates cloud resources that meet the service level agreement to the work task according to the service level agreement requirements.

[0019] Preferably, the processor resource scheduling unit schedules processor resources according to business requirements, including:

[0020] The core-level allocation subunit is used to allocate work tasks to corresponding processor cores according to business needs;

[0021] A hyperthreading scheduling subunit, connected to the statistical analysis module, for allocating parallelizable work tasks to the hyperthreading processors according to work task characteristics;

[0022] The processor affinity subunit is connected to the statistical analysis module and is used to allocate specific work tasks to corresponding processor cores.

[0023] Preferably, the processor load scheduling unit dynamically allocates work tasks according to the load conditions of each resource node, and the processor load scheduling unit includes:

[0024] The load monitoring subunit is used to monitor the resource usage of all resource nodes;

[0025] A dynamic migration subunit, connected to the load monitoring subunit, for dynamically migrating work tasks between the resource nodes, and controlling the load of the resource nodes to be an average;

[0026] The resource adjustment subunit is connected to the load monitoring subunit and is used to adjust the processor quota or instance quantity of each tenant according to the processor load of each tenant.

[0027] A method for adaptive resource management based on a multi-cloud environment, applied to the system for adaptive resource management based on a multi-cloud environment, comprising:

[0028] Step S1, collecting dynamic data from multiple cloud platforms;

[0029] Step S2: Statistically analyze the current status and trend of cloud resource usage based on business needs and load changes, output resource analysis results, and predict future demand for cloud resource expansion;

[0030] Step S3, determine whether the current resource configuration is appropriate; if appropriate, keep the current resource configuration; if not appropriate, execute step S4;

[0031] Step S4, dynamically configure, optimize and adjust the row resources. The adjustment methods include at least cost optimization, performance optimization, designated host scheduling, task scheduling according to the service level agreement, scheduling processor resources and scheduling processor load. After the adjustment, continue to analyze the latest collected dynamic data and return to execute step S3.

[0032] Beneficial effects of the present invention: Due to the adoption of the above technical scheme, the present invention automatically adjusts resource allocation according to the results of resource monitoring and analysis, dynamically allocates and optimizes the management of multi-cloud resources in an automated and intelligent manner, and can adjust the granularity size as needed, thereby significantly improving the resource utilization efficiency and management flexibility of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 This is an architecture diagram of a system for adaptive resource management in a multi-cloud environment in an embodiment of the present invention;

[0034] Figure 2 is an architecture diagram of a processor resource scheduling unit in an embodiment of the present invention;

[0035] Figure 3 is an architecture diagram of a processor load scheduling unit in an embodiment of the present invention;

[0036] Figure 4 A step diagram of a method for adaptive resource management based on a multi-cloud environment in an embodiment of the present invention;

[0037] Figure 5 This is a flowchart of a method for adaptive resource management based on a multi-cloud environment in an embodiment of the present invention. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0039] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0040] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0041] A system based on adaptive resource management in a multi-cloud environment, such as Figure 1 As shown, including

[0042] Resource collection module 1, used to collect dynamic data from multiple cloud platforms;

[0043] The statistical analysis module 2 is connected to the resource collection module 1 to receive dynamic data, and is used to statistically analyze the current status and trend of cloud resource usage according to business needs and load changes, output resource analysis results, and predict future demand for cloud resource expansion;

[0044] The judgment module 3 is connected to the statistical analysis module 2 and is used to judge whether the current resource configuration is appropriate;

[0045] The scheduling optimization module 4 is connected to the judgment module 3, and is used to dynamically configure, optimize and adjust resources according to the judgment result of the judgment module 3. The scheduling optimization module 4 includes a cost optimization unit 41 for performing cost optimization, a performance optimization unit 42 for performing performance optimization, a host scheduling unit 43 for performing designated host scheduling, a protocol scheduling unit 44 for performing task scheduling according to a service level agreement, a processor resource scheduling unit 45 for scheduling processor resources, and a processor load scheduling unit 46 for scheduling processor load, which are respectively connected to the statistical analysis module 2.

[0046] Specifically, the present invention collects various resource usage conditions in real time from multiple cloud environments and other infrastructure systems through the resource collection module 1, including the utilization rate and over-allocation of resources such as the CPU, memory, storage, and network bandwidth of the cloud host;

[0047] Statistical analysis module 2 analyzes resource usage status, trends, idle percentage, network access volume, peak fluctuation rate, etc. according to business needs and load changes;

[0048] The judgment module 3 judges whether the resource configuration is appropriate, redundant, resource-constrained, or has limited scalability based on the analysis results of the statistical analysis module 2.

[0049] The scheduling optimization module 4 automatically (or manually) adjusts resource allocation based on the results of resource monitoring and analysis, such as starting or shutting down virtual machine instances, adjusting storage capacity, adjusting bandwidth upper and lower limits, and drifting away from high-load hosts.

[0050] In a preferred embodiment, the resource collection module 1 collects dynamic data such as performance data, resource utilization and operation status information collected from the cloud platform, and the information sources of the dynamic data include performance data, alarm data, logs, network data and environmental information.

[0051] Specifically, the present invention collects dynamic data through the resource collection module 1, and the information sources include performance data, alarm data, logs, network data, environmental information, etc.;

[0052] The resource collection module 1 is responsible for collecting resource usage, performance data and related operating status from multiple cloud platforms.

[0053] The resource collection module 1 is called through the API provided by the cloud platform to collect multiple cloud platform performance data, resource utilization and operation status information in real time, such as central processing unit (CPU), graphics processing unit (GPU), memory, virtual hard disk, virtual router, virtual firewall, virtual load balancing, node health check information, device system log, etc.;

[0054] Collect traffic data from network load balancing devices and firewalls in real time to ensure bandwidth utilization and link quality information;

[0055] Collect disk read / write I / O, cache utilization, directory space utilization, etc. from storage devices in real time;

[0056] Collect cloud host security alerts, attack events, threat analysis, and vulnerability scans in real time from security devices.

[0057] To be more specific, the resource acquisition module 1 not only supports real-time data acquisition, but also supports historical data storage and analysis, and combines time series data for trend prediction. The resource acquisition module 1 stores the collected dynamic data in real time through the time series database, and uses the time series prediction models ARIMA and LSTM to use historical data for trend prediction. The real-time data and future trend analysis results are displayed through the graphical interface provided by the interactive module 5.

[0058] In a preferred embodiment, the statistical analysis module 2 obtains dynamic data, and a resource analysis unit for analyzing resource usage is provided in the statistical analysis module 2. The resource analysis unit uses a machine learning model or a rule engine to statistically analyze the current status and trend of cloud resource usage, and outputs resource analysis results. The business demand and load changes based on which the statistical analysis module 2 predicts future cloud resource expansion needs include at least application response time, business traffic changes, storage space changes, link quality changes, host load changes, system log alarms, and user access details;

[0059] The judgment module 3 is used to judge whether the current resource configuration meets the business needs and load changes. If the current resource configuration is met, the current resource configuration is maintained. If the current resource configuration is not met, the scheduling optimization module 4 performs dynamic resource configuration, optimization and adjustment.

[0060] It also includes an interaction module 5 connected to the scheduling optimization module 4, and the interaction module 5 has a graphical interface and an interaction interface for interaction.

[0061] Specifically, according to business needs and load changes, the current status and trends of resource usage are statistically analyzed to predict future demand for cloud resource expansion; the statistical analysis module 2 performs intelligent analysis based on the collected data, and makes resource adjustment decisions based on the scheduling strategies and rules preset in the cost optimization unit 41, performance optimization unit 42, host scheduling unit 43, protocol scheduling unit 44, processor resource scheduling unit 45 and processor load scheduling unit 46.

[0062] The collected dynamic data is transmitted to the resource analysis module, and the system uses machine learning models or rule engines to analyze resource usage. Machine learning models use supervised learning such as regression analysis to predict future resource requirements, or use unsupervised learning such as cluster analysis to identify anomalies. The system analyzes trend information based on business needs, such as application response time, business traffic changes, storage space changes, link quality changes, host load changes, system log alarms, user access details, etc.

[0063] The system analyzes the workload patterns in historical data to identify the workload characteristics corresponding to the changed business needs, that is, to obtain future resource requirements; or based on historical resource usage and future workload change forecasts, regression analysis and time series analysis are used to build resource growth curves to predict the amount of resources required in the future, thereby predicting changes in resource requirements;

[0064] The present invention can predict resource requirements according to changes in workload and resource growth curves, and perform resource scheduling in advance to stabilize system performance. For example, if it is found that the preset network guaranteed bandwidth is exceeded, the flow control (QOS) policy on the border firewall is adjusted through the API interface to limit the flow.

[0065] Scheduling Optimization Module 4 implements decision-making, dynamically configures, optimizes and adjusts resources to ensure that the system continues to operate efficiently in the following ways:

[0066] Automatically adjust resource allocation and usage policies based on the output of the resource analysis module, for example, automatically expand computing resources when resources are overloaded, or reduce resources when utilization is low;

[0067] Realize intelligent resource expansion, load balancing, resource recycling and other operations. When the resource analysis module determines that resource adjustment is needed, it will determine the best resource configuration method based on the preset strategy and actual situation. The resource configuration operations include resource expansion, contraction, load balancing or node migration.

[0068] It allows users to independently adjust business scenario strategies based on business usage, such as elastic expansion of cloud hosts, load balancing, VPC protection, etc. Users can quickly set corresponding management policies based on preset policy templates to implement user-defined functions.

[0069] It supports dynamic adjustment of heterogeneous cloud resources, can connect to a variety of mainstream cloud platform products, and can adapt to the API differences of different cloud platforms to provide a consistent resource management interface. Benfu Buying uses an adaptive algorithm model to adjust resources, which can optimize resource allocation between different cloud providers, thereby avoiding excessive dependence on a single cloud platform and improving system elasticity.

[0070] The interactive module 5 has a graphical interface and an interactive interface for interaction. The interactive interface implements visual management, report generation, and user policy configuration. The graphical interface (GUI) helps users view resource status, management policies, and performance reports in real time. It supports automatic report generation to help users understand the resource usage and performance of the system.

[0071] In a preferred embodiment, the cost optimization unit 41 evaluates the actual business volume of the work task according to the user application and allocates cloud resources corresponding to the actual business volume.

[0072] Specifically, based on the evaluation of user applications and actual business volume, the actual business volume can be learned based on the actual amount of resources used by the same type of user applications in historical data, and different cloud resource types, such as bandwidth resources, memory resources, and hard disk resources, can be reasonably allocated to maximize the utilization rate of cloud resources. Since users usually apply for resources that exceed the actual business volume when applying for resources, the cost optimization unit 41 can reduce unnecessary resource usage caused by inaccurate business usage assessment when users apply for resources, and solve the problem that users always apply for larger amounts.

[0073] In a preferred embodiment, the performance optimization unit 42 prioritizes resource nodes according to their computing capabilities and delays, and allocates work tasks to resource nodes whose priorities correspond to the business levels of the work tasks.

[0074] Specifically, the present invention allocates work tasks to resource nodes that provide high computing power and low latency, and gives priority to high-performance computing nodes with stronger processors, more memory, and low network latency according to the service level. During the scheduling process, load balancing is considered, and the over-use of various resources is evaluated to avoid excessive resource load on a single node, thereby optimizing overall performance.

[0075] In a preferred embodiment, the host scheduling unit 43 allocates the work tasks to the designated cloud hosts or cloud regions.

[0076] Specifically, workloads are scheduled to designated cloud hosts or designated cloud platforms according to user or business needs. Users can add specific tags to resources, such as security level, business level, and allowed access range, and the present invention assigns tasks to designated cloud hosts or cloud regions based on the tags. Users can directly bind workloads that need to run stably in a specific environment for a long time to specific cloud hosts.

[0077] In a preferred embodiment, the protocol scheduling unit 44 allocates cloud resources that meet the service level agreement to the work task according to the service level agreement requirements.

[0078] Specifically, according to the requirements of the service level agreement (SLA), ensure that the cloud service availability, performance and response time meet the predetermined goals. According to the availability requirements defined in the SLA, the present invention dispatches key tasks to cloud hosts with high availability guarantees using strategies such as multi-copy deployment and cross-region redundant deployment. To meet the high availability SLA requirements, the present invention supports a failover strategy, that is, when a cloud host fails, the task is automatically migrated to other available resource nodes. By real-time monitoring of various performance indicators, such as latency, bandwidth, response time, etc., resource allocation is dynamically adjusted to ensure that various SLA indicators are met.

[0079] In a preferred embodiment, the processor resource scheduling unit 45 schedules processor resources according to business requirements, such as Figure 2 As shown, including

[0080] The core level allocation subunit 451 is used to allocate work tasks to corresponding processor cores according to business requirements;

[0081] The hyperthreading scheduling subunit 452 is connected to the statistical analysis module 2 and is used to allocate parallelizable work tasks to the hyperthreading processors according to the work task characteristics;

[0082] The processor affinity subunit 453 is connected to the statistical analysis module 2 and is used to allocate specific work tasks to corresponding processor cores.

[0083] Specifically, based on the business's demand for computing power, processor resources can be flexibly scheduled to optimize the efficiency of computing resource utilization.

[0084] Core-level allocation: The scheduling platform can accurately allocate tasks to specific processor cores according to the requirements of the tasks. For example, the present invention monitors and analyzes the CPU usage of tasks to determine which work tasks require more computing resources. At runtime, according to the real-time CPU load and task requirements, the allocation of tasks is dynamically adjusted to allocate tasks that require a lot of computing to cores with stronger computing performance.

[0085] Hyper-threading scheduling: On processors that support hyper-threading, the scheduling platform can intelligently allocate tasks and maximize the parallel processing capabilities of the CPU using hyper-threading technology. Hyper-threading technology allows a single physical core to process multiple threads simultaneously. The present invention can maximize the parallel processing capabilities of multiple threads by performing task scheduling through the hyper-threading scheduling subunit 452. When scheduling tasks, parallel tasks are preferentially allocated to different logical cores of the same physical core, thereby improving the efficiency of CPU usage.

[0086] Processor affinity: By setting processor affinity (CPU affinity), tasks are fixed on specific processor cores to reduce the performance overhead caused by context switching and cache invalidation.

[0087] In a preferred embodiment, the processor load scheduling unit 46 dynamically allocates work tasks according to the load conditions of each resource node, such as Figure 3 As shown, the processor load scheduling unit 46 includes,

[0088] The load monitoring subunit 461 is used to monitor the resource usage of all resource nodes;

[0089] The dynamic migration subunit 462 is connected to the load monitoring subunit 461 and is used to dynamically migrate the work tasks between the resource nodes and control the load of each resource node to be an average value;

[0090] The resource adjustment subunit 463 is connected to the load monitoring subunit 461 and is used to adjust the processor quota or instance quantity of each tenant according to the processor load of each tenant.

[0091] Specifically, tasks are dynamically allocated according to the CPU load of each node to avoid overload and balance resource usage.

[0092] The load monitoring subunit 461 monitors the processor usage of all computing nodes in real time to ensure that the workload is evenly distributed and to avoid overloading or idling of some nodes.

[0093] If the processor load of a certain node is too high, the dynamic migration subunit 462 migrates part of the tasks to other nodes with lower loads, thereby balancing the load of the entire computing environment.

[0094] In a multi-tenant environment, the resource adjustment subunit 463 dynamically adjusts the processor quota or instance quantity of each tenant according to the processor load of each tenant to ensure the balance of overall performance and resource utilization. The present invention can dynamically adjust the quota by defining the initial processor quota of each tenant and combining it with real-time monitoring data. For example, if the CPU utilization of a tenant is lower than a certain threshold for a long time, its quota is reduced.

[0095] A method for adaptive resource management in a multi-cloud environment is applied to a system for adaptive resource management in a multi-cloud environment in any embodiment, such as Figure 4 , Figure 5 As shown, including

[0096] Step S1, collecting dynamic data from multiple cloud platforms;

[0097] Step S2: Statistically analyze the current status and trend of cloud resource usage based on business needs and load changes, output resource analysis results, and predict future demand for cloud resource expansion;

[0098] Step S3, determine whether the current resource configuration is appropriate; if appropriate, keep the current resource configuration; if not appropriate, execute step S4;

[0099] Step S4, dynamically configure, optimize and adjust resources. The adjustment methods include at least cost optimization, performance optimization, designated host scheduling, task scheduling according to the service level agreement, scheduling processor resources and scheduling processor load. After the adjustment, continue to analyze the latest collected dynamic data and return to execute step S3.

[0100] In summary, the present invention can dynamically allocate and optimize the management of multi-cloud resources in an automated and intelligent manner, significantly improving the resource utilization efficiency and management flexibility of the system.

[0101] The above description is only a preferred embodiment of the present invention, and does not limit the implementation mode and protection scope of the present invention. For those skilled in the art, it should be aware that all solutions obtained by equivalent substitutions and obvious changes made using the description and illustrations of the present invention should be included in the protection scope of the present invention.

Claims

1. A system based on adaptive resource management in a multi-cloud environment, characterized in that: include, A resource collection module (1), used for collecting dynamic data from multiple cloud platforms; A statistical analysis module (2) is connected to the resource acquisition module (1) to receive the dynamic data, and is used to statistically analyze the current status and trend of cloud resource usage according to business needs and load changes, output resource analysis results, and predict future demand for cloud resource expansion; A judgment module (3), connected to the statistical analysis module (2), is used to judge whether the current resource configuration is appropriate; A scheduling optimization module (4) is connected to the judgment module (3) and is used to dynamically configure, optimize and adjust resources according to the judgment result of the judgment module (3). The scheduling optimization module (4) includes a cost optimization unit (41) for performing cost optimization, a performance optimization unit (42) for performing performance optimization, a host scheduling unit (43) for performing designated host scheduling, a protocol scheduling unit (44) for performing task scheduling according to a service level agreement, a processor resource scheduling unit (45) for scheduling processor resources, and a processor load scheduling unit (46) for scheduling processor load.

2. The system based on adaptive resource management in a multi-cloud environment according to claim 1, characterized in that: The resource collection module (1) collects the dynamic data, which are performance data, resource utilization and operation status information collected from the cloud platform. The information sources of the dynamic data include performance data, alarm data, logs, network data and environmental information.

3. The system based on adaptive resource management in a multi-cloud environment according to claim 1, characterized in that: The statistical analysis module (2) obtains the dynamic data. The statistical analysis module (2) is provided with a resource analysis unit for analyzing resource usage. The resource analysis unit uses a machine learning model or a rule engine to statistically analyze the current status and trend of cloud resource usage and outputs the resource analysis result. The business demand and load changes based on which the statistical analysis module (2) predicts the future expansion demand of cloud resources include at least application response time, business traffic change, storage space change, link quality change, host load change, system log alarm and user access details. The judgment module (3) is used to judge whether the current resource configuration meets the business requirements and load changes. If the current resource configuration meets the business requirements and load changes, the current resource configuration is maintained. If the current resource configuration does not meet the business requirements and load changes, the scheduling optimization module (4) performs dynamic resource configuration, optimization and adjustment. It also includes an interaction module (5) connected to the scheduling optimization module (4), and the interaction module (5) is provided with a graphical interface and an interaction interface for interaction.

4. The system based on adaptive resource management in a multi-cloud environment according to claim 1, characterized in that: The cost optimization unit (41) evaluates the actual business volume of the work task according to the user application and allocates cloud resources corresponding to the actual business volume.

5. The system based on adaptive resource management in a multi-cloud environment according to claim 1, characterized in that: The performance optimization unit (42) performs priority sorting according to the computing capabilities and delays of the resource nodes, and allocates the work tasks to the resource nodes whose priorities correspond to the business levels of the work tasks.

6. The system based on adaptive resource management in a multi-cloud environment according to claim 1, characterized in that: The host scheduling unit (43) distributes the work tasks to the designated cloud hosts or cloud areas.

7. The system based on adaptive resource management in a multi-cloud environment according to claim 1, characterized in that: The protocol scheduling unit (44) allocates cloud resources that meet the service level agreement to the work task according to the service level agreement requirements.

8. The system based on adaptive resource management in a multi-cloud environment according to claim 1, characterized in that: The processor resource scheduling unit (45) schedules processor resources according to business requirements, including: A core level allocation subunit (451), used to allocate work tasks to corresponding processor cores according to business requirements; A hyperthreading scheduling subunit (452), connected to the statistical analysis module (2), for allocating parallelizable work tasks to the hyperthreading processors according to work task characteristics; The processor affinity subunit (453) is connected to the statistical analysis module (2) and is used to allocate specific work tasks to corresponding processor cores.

9. The system based on adaptive resource management in a multi-cloud environment according to claim 5, characterized in that: The processor load scheduling unit (46) dynamically allocates work tasks according to the load conditions of each resource node, and the processor load scheduling unit (46) includes: A load monitoring subunit (461), used to monitor resource usage of all resource nodes; A dynamic migration subunit (462), connected to the load monitoring subunit (461), is used to dynamically migrate work tasks between the resource nodes and control the load of the resource nodes to be an average value; The resource adjustment subunit (463) is connected to the load monitoring subunit (461) and is used to adjust the processor quota or instance quantity of each tenant according to the processor load of each tenant.

10. A method for adaptive resource management in a multi-cloud environment, characterized in that: A system for adaptive resource management based on a multi-cloud environment as described in any one of claims 1 to 9, comprising: Step S1, collecting dynamic data from multiple cloud platforms; Step S2: Statistically analyze the current status and trend of cloud resource usage based on business needs and load changes, output resource analysis results, and predict future demand for cloud resource expansion; Step S3, determining whether the current resource configuration is appropriate; if appropriate, maintaining the current resource configuration; If not suitable, proceed to step S4; Step S4, dynamically configure, optimize and adjust resources. The adjustment methods include at least cost optimization, performance optimization, designated host scheduling, task scheduling according to the service level agreement, scheduling processor resources and scheduling processor load. After the adjustment, continue to analyze the latest collected dynamic data and return to execute step S3.

Images