CN117762644A - Resource dynamic scheduling technology for distributed cloud computing systems - Google Patents

Abstract

The invention relates to a resource dynamic scheduling technology of a distributed cloud computing system, which comprises the following steps: s1, demand analysis: determining the service requirement and technical requirement of the system, evaluating the existing infrastructure and technology, and setting targets and performance indexes; s2, designing a system architecture: the overall structure of the system is designed based on the results of the demand analysis, and high-level components of the system such as databases, application servers, load balancers, and the like are determined. The invention remarkably improves the utilization rate of resources, reduces the cost, can optimize the system performance, accelerates the task execution speed and improves the overall reliability and usability of the system by intelligently analyzing and predicting the resource demand and adopting a high-efficiency scheduling strategy. In addition, it also supports extensibility and flexibility, enabling the system to accommodate future changes in demand. By constant monitoring and self-tuning, the system ensures high efficiency and stability of long-term operation, thereby creating greater commercial value for the enterprise or organization.

Description

Resource dynamic scheduling technology for distributed cloud computing systems

Technical field

The present invention relates to the technical field of dynamic resource scheduling, and in particular to the dynamic resource scheduling technology of distributed cloud computing systems.

Background technique

The dynamic resource scheduling technology of distributed cloud computing systems is a key technology designed to handle and optimize resource allocation and scheduling in distributed cloud computing environments. It focuses on how to reasonably allocate computing power, storage resources and network bandwidth in a distributed cloud architecture to ensure efficient system performance and response speed.

After the applicant’s research on the scheduling technology required by the current distributed cloud computing environment, it was found that the scheduling technology required by the current distributed cloud computing environment has the following disadvantages:

Existing systems are often unable to respond quickly to sudden changes in workloads, resulting in insufficient resource allocation during peak hours and waste of resources during off-peak hours. In addition, due to the lack of optimization strategies from a global perspective, existing systems are often unable to optimize the entire distributed environment. Optimization of resource allocation is achieved in Ensure high availability and fast recovery, affecting service quality.

Therefore, those skilled in the art have proposed a dynamic resource scheduling technology for distributed cloud computing systems.

Contents of the invention

In view of the above-mentioned problems existing in the prior art, the main purpose of the present invention is to provide a dynamic resource scheduling technology for a distributed cloud computing system.

The technical solution of the present invention is as follows: dynamic resource scheduling technology for distributed cloud computing systems, including the following steps:

S1. Requirements analysis: Determine the business needs and technical requirements of the system, evaluate existing infrastructure and technology, and set goals and performance indicators;

S2. Design system architecture: Design the overall structure of the system based on the results of demand analysis, determine the high-level components of the system, such as databases, application servers, load balancers, etc., as well as the interaction between these components, in order to ensure system scalability and flexibility, which requires designing and planning network architecture and data storage strategies;

S3. Workload analysis: The process of understanding the type, size and frequency of tasks that the system will handle, predicting resource requirements by analyzing historical data and expected workloads, and optimizing resource allocation and scheduling strategies based on these data;

S4. Scheduling algorithm development: Based on the results of workload analysis, develop resource scheduling algorithms to optimize task running efficiency and resource utilization, including the implementation of priority queues, rule-based engines, and adaptive scheduling using machine learning technology algorithm;

S5. Resource management: Realize the monitoring, allocation, optimization and control of physical and virtual resources, including the development of functions to automatically manage the resource life cycle, such as allocation, release, and expansion and migration of resources according to demand;

S6. System integration verification: All independently developed modules and components are combined together and then tested as a complete system, which includes verifying the function and performance of the system and ensuring that all parts can work together to meet the design specifications;

S7. Monitoring strategy implementation: Focus on implementing monitoring and alarm systems to ensure that system health, performance changes and potential problems can be monitored in real time in the production environment;

S8. Continuous system optimization: In the later stage of system deployment, it is necessary to continuously monitor system performance and optimize the system based on feedback. This includes upgrading hardware, updating software, improving scheduling strategies, optimizing resource allocation, etc., to ensure that the system can adapt to constant changes. needs and the latest technology trends.

As a preferred implementation, step S1 can be refined into:

S11. Evaluate hardware resources: Evaluate existing physical machines, virtual machines, storage solutions and network resources. The focus of this step is to understand the performance characteristics, expansion capabilities and limitations of the existing infrastructure. The evaluation includes checking the age and maintenance of the hardware. Records, past performance history and failure rates to facilitate reasonable inferences and decisions for future scale expansion and resource allocation;

S12. Define the workload model: Determine the types of businesses and tasks that the system needs to support. Based on the characteristics of the application, such as transactional systems, data processing, or real-time analysis, analyze the load that may be exerted on the resources. In this step, you need to The expected peak value, average value, steady state and changing trend of the load are defined in the model;

S13. Set scheduling goals: Based on business needs and expected system usage, set the goals to be achieved by the scheduler. More specifically, if the system needs to respond quickly to user requests, it may give priority to minimizing the response time of the task; or if it saves costs, is the main goal and may be optimized to increase the efficiency of energy and resource use;

S14. Determine service level agreement requirements: SLAS defines the quality of service and the performance indicators promised to users, including system availability, service response time, fault repair time, etc.;

S15. Establish monitoring requirements: Monitoring requirements and indicators should match the defined SLAS and performance goals, and determine monitoring indicators, including CPU usage, memory usage, network bandwidth, storage IOPS, etc., as well as monitoring intervals and historical data. Retention policy, this step can help system administrators understand the status of the system and respond promptly when problems occur;

S16. Standardized resource description: Create a unified description and definition for all physical and virtual resources in the system, including CPU model and core number, memory size, storage capacity, network bandwidth, etc. Standardization is the key basis for resource allocation and scheduling decisions. , ensuring that the system can manage and use these resources efficiently and consistently.

Through the above technical means, at this stage, the aim is to identify system requirements and available resources. We evaluate existing hardware and software resources, determine the type and size of workloads, develop scheduling goals, and determine necessary service level objectives (SLAs). Next, we also need to decide on a monitoring solution to ensure that key performance indicators can be tracked, and finally develop a standard description format for resources.

As a preferred implementation, the step S2 can be refined into:

S21. Determine the system architecture model: Choose an appropriate architecture model, such as microservice architecture, service-oriented architecture (SOA), event-driven architecture or traditional single application architecture. The selection of this architecture model needs to be based on demand analysis and maintainability should be considered. , scalability, performance and team experience and other factors;

S22. Design a fault-tolerant mechanism: Plan the redundancy and fault-tolerance mechanism of the system to ensure high availability. This design needs to include data backup solutions, multi-region deployment, failover and recovery strategies, and the implementation of stateless components to support horizontal expansion;

S23. Plan data management strategy: Determine data storage requirements, select an appropriate database system, consider data consistency, integrity, and accessibility, and plan data backup, recovery, and archiving strategies;

S24. Design a system monitoring solution: Based on the monitoring requirements set in the requirements analysis, design a complete monitoring solution, how to collect, store and analyze monitoring data, how to set alarm thresholds, and select monitoring tools and platforms to provide real-time monitoring. and performance analysis;

S25. Plan the network architecture: Refining the network architecture includes internal network isolation, the setting of public and private subnets, the use of load balancers, and security strategies to ensure network communication such as firewalls, intrusion detection systems, and encrypted transmission;

S26. Define system components and interfaces: clarify responsibilities and define interfaces for each component in the system. The API design that is exposed to the outside world should be easy to understand and use, and maintain consistency and version control. Communication between internal components also needs to be defined, including Messaging protocols and data formats.

Through the above technical means, the design phase involves the overall planning of the system structure. This includes structuring the resource hierarchy, setting up core system components, defining scheduling processes, building internal communication protocols, and planning fault handling and data persistence strategies. The goal of this phase is to create a solid blueprint that guides the detailed implementation of the system.

As a preferred implementation, step S3 can be refined into:

S31. Data collection and cleaning: Design an efficient data collection mechanism, which involves automatically collecting resource utilization data from different services and applications in a distributed environment, deploying log aggregation tools such as ELK (Elasticsearch, Logstash, and Kibana) or Fluentd, and monitoring systems Such as Prometheus or Datadog to collect data in a standardized and centralized manner and formulate cleaning rules, including anomaly detection and correction, time zone synchronization, and unified data format;

S32. Feature selection: Use data analysis and visualization tools (such as Pandas and Seaborn in Python) to evaluate resource usage patterns, identify key features through statistical testing and correlation analysis, and implement machine learning feature selection techniques (such as recursive feature elimination), Supplemented by professional knowledge to optimize the input feature set of the model;

S33. Algorithm development: Based on the selected feature set, explore various prediction models, including traditional statistical methods (such as ARIMA) and modern machine learning techniques (such as gradient boosting machines and neural networks), to ensure that the algorithm development environment has the necessary computing resources , and use an appropriate ML framework (such as TensorFlow or scikit-learn);

S34. Model training: Based on historical data, use machine learning workflow management technology (such as MLflow or Kubeflow) to ensure the traceability and consistency of the model training process, and use parameter search (such as grid search or random search) to optimize model and use cross-validation to evaluate the generalization ability of the model;

S35. Model verification: Verify the model performance on an independent test data set, use suitable evaluation indicators (such as MAE, RMSE) to measure prediction accuracy, and use the verification results to further optimize model parameters, and review the model regularly according to changes in application scenarios. Representation, updating the model to adapt to new data patterns;

S36. Integration and deployment: Design and implement the automated deployment process of the model, including CI/CD pipeline (continuous integration/continuous deployment), so that the predictive model can be smoothly integrated into the existing scheduling system to ensure adequate monitoring and alarming Mechanisms to track model performance and manage model versions in production environments.

Through the above technical means, workload prediction is to be able to foresee future resource requirements. It starts with the collection and cleaning of data, then feature selection to determine which data points are most relevant, followed by algorithm development and model training, and finally validating and optimizing the accuracy of the model and its effectiveness in actual operations, and integrate it into the scheduler.

As a preferred implementation, step S4 can be refined into:

S41. Policy framework design: Build a flexible and scalable scheduling framework that supports plug-in policy exchange. The framework should be able to be seamlessly integrated with the system architecture, configurable and easy to maintain. Use industry standard design patterns (such as strategy, observer pattern) to assist code separation and ensure system modularity;

S42. Static scheduling algorithm implementation: Write a series of basic static scheduling algorithms, such as first come first served (FCFS), round-robin (Round-Robin) and fixed priority (Fixed Priority) scheduling. More specifically, the algorithm should be able to Make decisions quickly and easily taking into account external changes and create standard test cases for these algorithms to verify their performance and correctness;

S43. Dynamic scheduling mechanism: Develop dynamic scheduling algorithms that can respond to changes in real-time system status. These algorithms consider factors such as current resource utilization, historical load data, service level agreements (SLA), and priorities, using bionic algorithms (such as genetic algorithms). ), heuristic methods or machine learning techniques to achieve dynamic response and adaptive scheduling decisions;

S44. Algorithm optimization: Continuously conduct performance analysis and optimization of existing algorithms, and use micro and macro optimization technologies, such as algorithm complexity reduction, multi-thread parallel processing and resource reservation mechanisms. At the same time, consider implementing a simulation environment for evaluation and optimization Algorithm performance under different loads and conditions;

S45. Multi-objective scheduling support: Develop scheduling algorithms that support multi-objective decision-making. The algorithm can balance multiple factors such as response time, resource usage, energy efficiency, and cost. It uses multi-objective optimization theory and technology, such as Pareto optimization, to develop Scheduling strategies for different business and technical needs;

S46. Isolation and security: Consider inter-tenant isolation and application-level security in scheduling decisions, implement role-based access control (RBAC) and hardware resource isolation between tenant workloads, and add security verification and monitoring steps to the policy. , protect the system from malicious behavior.

Through the above technical means, this stage focuses on developing scheduling algorithms that can effectively allocate resources. It includes designing a flexible policy framework, implementing static and dynamic scheduling algorithms, and optimizing algorithms to ensure that scheduling results meet business requirements. At the same time, isolation and security factors must also be taken into consideration.

As a preferred implementation, step S5 can be refined into:

S51. Define resource management strategies: analyze the characteristics and business requirements of different types of resources, formulate corresponding resource management strategies, formulate core binding, priority and time slice quotas for CPU scheduling, configure priority and recycling policies for memory, and provide storage Set read and write rate limits, allocate maximum and minimum bandwidth limits for network bandwidth, and ensure efficient and fair resource use while meeting SLA requirements;

S52. Resource allocation algorithm: Develop a finely controlled resource allocation algorithm, which can dynamically allocate resources according to the current system resource usage and workload requirements. The algorithm considers the interdependence and constraints of resources to achieve rapid response and flexible adjustment of resource allocation. ;

S53. Resource optimization cycle: Implement regular resource optimization cycles, identify resource usage patterns and efficiency bottlenecks by analyzing historical data, regularly adjust resource allocation, merge memory fragments, rebalance distribution, etc., to promote resource efficiency and system stability;

S54. Elastic scaling logic: Implement elastic scaling logic, allowing the system to automatically adjust resource allocation based on real-time monitoring data and prediction results. When the load changes, it can increase or decrease computing nodes or service instances in a timely manner to ensure that applications always maintain optimal performance and cost-effectiveness;

S55. Energy efficiency optimization: Develop an energy optimization subsystem to monitor the energy usage of the data center, analyze energy consumption patterns, and reduce energy consumption by adjusting resource usage strategies and scheduling strategies to ensure the green operation of the system and help reduce costs. ;

S56. Persistence and state recovery: Build a state persistence mechanism to regularly save the current state of the system to persistent storage, such as a database or distributed storage system. After a system failure, these persistent state data can be used to quickly restore system work. , minimizing the impact of failures on business.

Through the above technical means, in the resource management stage, resource management strategies are formulated, resource allocation algorithms are developed, and resource optimization cycles are introduced. In addition, elastic scaling logic is implemented to adapt to load changes, optimize energy utilization, and ensure system durability and resilience in the face of failures.

As a preferred implementation, step S6 can be refined into:

S61. Module integration: Integrate the developed modules (such as scheduler, resource manager, monitoring system) into the main system framework to ensure that the interfaces between modules are compatible, and complete necessary integration tests to check whether the data flow and function calls are as required. expected work;

S62. System-wide testing: Conduct end-to-end testing of the entire system, including functional testing, integration testing, system testing and acceptance testing, to verify the core functions, user stories and boundary conditions of the system. In addition, it is necessary to test the system under abnormal conditions. behavior to ensure stability;

S63. Performance evaluation: Evaluate the performance of the system under typical and extreme workloads, including response time, throughput and resource usage efficiency, use standardized performance evaluation tools and methods, and compare it with the performance requirements in the business goals to ensure that it is met predetermined indicators;

S64. Isolation verification: Conduct a security audit on the system to check whether there are code vulnerabilities, misconfigurations and potential security risks, verify the multi-tenant isolation capabilities in the system, and ensure the safe isolation of data and operations of different users or organizations;

S65. User case simulation: Based on typical user scenarios, simulate user operations to verify system functions, ensure the correctness of user processes and business processes, monitor the performance of the system under simulated operations, and adjust system configuration to better serve user needs. ;

S66. Load stress test: By simulating high load conditions and artificially created pressure points, the performance bottlenecks and potential problems of the system in the extreme state are identified, and the system architecture and resource allocation are adjusted based on the test results to ensure the stability and stability of the system in actual operation. reliability.

Using the techniques described above, this phase involves integrating previously independently developed modules into a complete system, conducting system-wide testing to ensure the modules work together correctly, and then evaluating the performance of the system to ensure it meets established performance standards. . Security and isolation assessments are also conducted and tested by simulating user actions.

As a preferred implementation, step S7 can be refined into:

S71. Deployment planning: Create a detailed deployment plan and consider various deployment scenarios, such as blue-green deployment to achieve zero downtime updates, or rolling updates to gradually replace older version instances. The plan should include risk assessment and prior notification. , rollback strategy, deployment schedule, and consideration of critical business periods;

S72. Automated deployment process: Develop an automated deployment process, use continuous integration and continuous deployment (CI/CD) pipelines to automatically test, package and deploy new versions, ensuring that this process covers code submission, build, automated testing, deployment and notification , to reduce human error and improve efficiency;

S73. Monitoring strategy implementation: Select monitoring tools based on system key performance indicators (KPIs), establish monitoring strategies, realize real-time resource usage, service operating status and abnormal event monitoring, and build dashboards and alarm systems to facilitate operation and maintenance personnel quickly Respond to any questions;

S74. Fault drills: Regularly arrange fault drill activities to simulate fault scenarios and verify the system's recovery process and backup strategy. These drills help ensure that when a real emergency occurs, the team can take quick and accurate actions to reduce possible business impacts;

S75. Performance tuning: Continuously collect performance data, and perform regular performance tuning of the system based on these data, identify performance bottlenecks, optimize configurations, apply best practices, and update hardware to address performance challenges. At the same time, cost-effectiveness must be considered to find The best balance between resource usage and performance;

S76. Documentation and logs: Maintain detailed system documentation, including design documents, user manuals, operation guides and frequently asked questions (FAQs), so that users and managers can understand and use the system, implement log management policies, and record system operation in detail. Combine with log analysis tools to gain insight into system behavior and quickly diagnose problems.

Through the above technical means, during the deployment phase, we need to formulate a detailed deployment plan, realize the automated deployment process, and deploy monitoring strategies to track resource and application-level performance. Failure drills and backup verification are to establish preparations for possible failures and perform performance tuning based on actual system operation.

As a preferred implementation, step S8 can be refined into:

S81. User feedback collection: Establish a comprehensive user feedback collection mechanism, including online surveys, user forums, direct visits or customer support feedback and other channels to ensure that system usage, performance bottlenecks, possible improvement suggestions and User satisfaction data, regularly analyze these data, and identify user experience improvement points;

S82. Issue tracking and repair: Build an issue tracking system to continuously monitor, record and classify issues that arise in the system, assign a priority to each issue, and carry out issue and repair work within the corresponding time frame to ensure smooth cross-team communication. , so that relevant departments can work closely to solve problems quickly;

S83. Performance monitoring and optimization: Deploy advanced performance monitoring tools to track system performance indicators in real time, use data analysis and machine learning technology to identify performance trends and potential problems, and based on the analysis results, make timely system adjustments or optimize configurations to maintain optimal performance. ;

S84. Update iteration: Implement a structured update and iteration strategy to enable the system to integrate new features, performance improvements, and security patches, ensure that the update process minimizes disruption to on-site operations, and ensure stability and stability before and after updates through automated testing. compatibility;

S85. Compliance assurance: Regularly check and implement the latest security updates, monitor the security vulnerability database, ensure that the system is updated in a timely manner to prevent potential attacks, and at the same time, maintain the system in compliance with all relevant industry standards and regulatory requirements, and conduct regular compliance audits and evaluations;

S86. Technology trend assessment: Continuously monitor and evaluate emerging technologies and industry trends, examine the benefits they may bring to the system, organize regular internal knowledge sharing meetings and technical lectures, encourage team members to learn new technologies and consider integrating them into system to stay ahead of the curve.

Through the above technical means, the system will enter the maintenance phase. At this time, user feedback is collected, problems are tracked and fixed, system performance is continuously monitored, and updates and iterations are performed regularly. Security and compliance are one of the focuses of this phase. At the same time, new technology trends need to be constantly evaluated to decide whether to incorporate them into the system.

Compared with the existing technology, the advantages and positive effects of the present invention are:

The invention assesses resource needs by analyzing historical and real-time data in detail, and then designs a structure to meet those needs while being flexible enough to handle future changes. The system then uses predictive models to estimate future resource requests to accurately allocate resources to maximize performance and efficiency. Ultimately, scheduling algorithms are designed to ensure that resources are allocated correctly and efficiently among different tasks and services, so that the system can maintain high performance and high availability. This entire process needs to be continuously optimized and adjusted through feedback and monitoring. This scheduling technology significantly improves resource utilization and reduces overhead through intelligent analysis and prediction of resource needs and the adoption of efficient scheduling strategies. It can optimize system performance. It speeds up task execution and improves the overall reliability and availability of the system. Additionally, it supports scalability and flexibility, allowing the system to adapt to future changes in requirements. Through continuous monitoring and self-adjustment, the system ensures high efficiency and stability in long-term operation, thereby creating greater business value for the enterprise or organization.

Description of the drawings

Figure 1 is a large step flow chart of the present invention;

Figure 2 is a flow chart of S1 steps of the present invention;

Figure 3 is a flow chart of S2 steps of the present invention;

Figure 4 is a flow chart of S3 steps of the present invention;

Figure 5 is a flow chart of S4 steps of the present invention;

Figure 6 is a flow chart of S5 steps of the present invention;

Figure 7 is a flow chart of S6 steps of the present invention;

Figure 8 is a flow chart of steps S7 of the present invention;

Figure 9 is a flow chart of step S8 of the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments These are some embodiments of the present invention, rather than all embodiments.

The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

Example

As shown in Figures 1-8, the present invention provides resource dynamic scheduling technology for distributed cloud computing systems: including the following steps:

S1. Requirements analysis: Determine the business needs and technical requirements of the system, evaluate existing infrastructure and technology, and set goals and performance indicators;

S2. Design system architecture: Design the overall structure of the system based on the results of demand analysis, determine the high-level components of the system, such as databases, application servers, load balancers, etc., as well as the interaction between these components, in order to ensure system scalability and flexibility, which requires designing and planning network architecture and data storage strategies;

S3. Workload analysis: The process of understanding the type, size and frequency of tasks that the system will handle, predicting resource requirements by analyzing historical data and expected workloads, and optimizing resource allocation and scheduling strategies based on these data;

S4. Scheduling algorithm development: Based on the results of workload analysis, develop resource scheduling algorithms to optimize task running efficiency and resource utilization, including the implementation of priority queues, rule-based engines, and adaptive scheduling using machine learning technology algorithm;

S5. Resource management: Realize the monitoring, allocation, optimization and control of physical and virtual resources, including the development of functions to automatically manage the resource life cycle, such as allocation, release, and expansion and migration of resources according to demand;

S6. System integration verification: All independently developed modules and components are combined together and then tested as a complete system, which includes verifying the function and performance of the system and ensuring that all parts can work together to meet the design specifications;

S7. Monitoring strategy implementation: Focus on implementing monitoring and alarm systems to ensure that system health, performance changes and potential problems can be monitored in real time in the production environment;

S8. Continuous system optimization: In the later stage of system deployment, it is necessary to continuously monitor system performance and optimize the system based on feedback. This includes upgrading hardware, updating software, improving scheduling strategies, optimizing resource allocation, etc., to ensure that the system can adapt to constant changes. needs and the latest technology trends;

The step S1 can be refined into:

S11. Evaluate hardware resources: Evaluate existing physical machines, virtual machines, storage solutions and network resources. The focus of this step is to understand the performance characteristics, expansion capabilities and limitations of the existing infrastructure. The evaluation includes checking the age and maintenance of the hardware. Records, past performance history and failure rates to facilitate reasonable inferences and decisions for future scale expansion and resource allocation;

S12. Define the workload model: Determine the types of businesses and tasks that the system needs to support. Based on the characteristics of the application, such as transactional systems, data processing, or real-time analysis, analyze the load that may be exerted on the resources. In this step, you need to The expected peak value, average value, steady state and changing trend of the load are defined in the model;

S13. Set scheduling goals: Based on business needs and expected system usage, set the goals to be achieved by the scheduler. More specifically, if the system needs to respond quickly to user requests, it may give priority to minimizing the response time of the task; or if it saves costs, is the main goal and may be optimized to increase the efficiency of energy and resource use;

S14. Determine service level agreement requirements: SLAS defines the quality of service and the performance indicators promised to users, including system availability, service response time, fault repair time, etc.;

S15. Establish monitoring requirements: Monitoring requirements and indicators should match the defined SLAS and performance goals, and determine monitoring indicators, including CPU usage, memory usage, network bandwidth, storage IOPS, etc., as well as monitoring intervals and historical data. Retention policy, this step can help system administrators understand the status of the system and respond promptly when problems occur;

S16. Standardized resource description: Create a unified description and definition for all physical and virtual resources in the system, including CPU model and core number, memory size, storage capacity, network bandwidth, etc. Standardization is the key basis for resource allocation and scheduling decisions. , ensure that the system can manage and use these resources efficiently and consistently;

To summarize: In this phase, the aim is to identify system requirements and available resources. We evaluate existing hardware and software resources, determine the type and size of workloads, develop scheduling goals, and determine necessary service level objectives (SLAs). Next, we also need to decide on a monitoring solution to ensure that key performance indicators can be tracked, and finally develop a standard description format for resources.

The step S2 can be refined into:

S21. Determine the system architecture model: Choose an appropriate architecture model, such as microservice architecture, service-oriented architecture (SOA), event-driven architecture or traditional single application architecture. The selection of this architecture model needs to be based on demand analysis and maintainability should be considered. , scalability, performance and team experience and other factors;

S22. Design a fault-tolerant mechanism: Plan the redundancy and fault-tolerance mechanism of the system to ensure high availability. This design needs to include data backup solutions, multi-region deployment, failover and recovery strategies, and the implementation of stateless components to support horizontal expansion;

S23. Plan data management strategy: Determine data storage requirements, select an appropriate database system, consider data consistency, integrity, and accessibility, and plan data backup, recovery, and archiving strategies;

S24. Design a system monitoring solution: Based on the monitoring requirements set in the requirements analysis, design a complete monitoring solution, how to collect, store and analyze monitoring data, how to set alarm thresholds, and select monitoring tools and platforms to provide real-time monitoring. and performance analysis;

S25. Plan the network architecture: Refining the network architecture includes internal network isolation, the setting of public and private subnets, the use of load balancers, and security strategies to ensure network communication such as firewalls, intrusion detection systems, and encrypted transmission;

S26. Define system components and interfaces: clarify responsibilities and define interfaces for each component in the system. The API design that is exposed to the outside world should be easy to understand and use, and maintain consistency and version control. Communication between internal components also needs to be defined, including Messaging protocols and data formats;

To sum up: the design phase involves the overall planning of the system structure. This includes structuring the resource hierarchy, setting up core system components, defining scheduling processes, building internal communication protocols, and planning fault handling and data persistence strategies. The goal of this phase is to create a solid blueprint that guides the detailed implementation of the system.

The step S3 can be refined into:

S31. Data collection and cleaning: Design an efficient data collection mechanism, which involves automatically collecting resource utilization data from different services and applications in a distributed environment, deploying log aggregation tools such as ELK (Elasticsearch, Logstash, and Kibana) or Fluentd, and monitoring systems Such as Prometheus or Datadog to collect data in a standardized and centralized manner and formulate cleaning rules, including anomaly detection and correction, time zone synchronization, and unified data format;

S32. Feature selection: Use data analysis and visualization tools (such as Pandas and Seaborn in Python) to evaluate resource usage patterns, identify key features through statistical testing and correlation analysis, and implement machine learning feature selection techniques (such as recursive feature elimination), Supplemented by professional knowledge to optimize the input feature set of the model;

S34. Model training: Based on historical data, use machine learning workflow management technology (such as MLflow or Kubeflow) to ensure the traceability and consistency of the model training process, and use parameter search (such as grid search or random search) to optimize model and use cross-validation to evaluate the generalization ability of the model;

S35. Model verification: Verify the model performance on an independent test data set, use suitable evaluation indicators (such as MAE, RMSE) to measure prediction accuracy, and use the verification results to further optimize model parameters, and review the model regularly according to changes in application scenarios. Representation, updating the model to adapt to new data patterns;

S36. Integration and deployment: Design and implement the automated deployment process of the model, including CI/CD pipeline (continuous integration/continuous deployment), so that the predictive model can be smoothly integrated into the existing scheduling system to ensure adequate monitoring and alarming Mechanisms to track model performance and manage model versions in production environments;

To sum up: workload forecasting is about being able to foresee future resource needs. It starts with the collection and cleaning of data, then feature selection to determine which data points are most relevant, followed by algorithm development and model training, and finally validating and optimizing the accuracy of the model and its effectiveness in actual operations, and integrate it into the scheduler.

The step S4 can be refined into:

S41. Policy framework design: Build a flexible and scalable scheduling framework that supports plug-in policy exchange. The framework should be able to be seamlessly integrated with the system architecture, configurable and easy to maintain. Use industry standard design patterns (such as strategy, observer pattern) to assist code separation and ensure system modularity;

S42. Static scheduling algorithm implementation: Write a series of basic static scheduling algorithms, such as first come first served (FCFS), round-robin (Round-Robin) and fixed priority (Fixed Priority) scheduling. More specifically, the algorithm should be able to Make decisions quickly and easily taking into account external changes and create standard test cases for these algorithms to verify their performance and correctness;

S43. Dynamic scheduling mechanism: Develop dynamic scheduling algorithms that can respond to changes in real-time system status. These algorithms consider factors such as current resource utilization, historical load data, service level agreements (SLA), and priorities, using bionic algorithms (such as genetic algorithms). ), heuristic methods or machine learning techniques to achieve dynamic response and adaptive scheduling decisions;

S44. Algorithm optimization: Continuously conduct performance analysis and optimization of existing algorithms, and use micro and macro optimization technologies, such as algorithm complexity reduction, multi-thread parallel processing and resource reservation mechanisms. At the same time, consider implementing a simulation environment for evaluation and optimization Algorithm performance under different loads and conditions;

S45. Multi-objective scheduling support: Develop scheduling algorithms that support multi-objective decision-making. The algorithm can balance multiple factors such as response time, resource usage, energy efficiency, and cost. It uses multi-objective optimization theory and technology, such as Pareto optimization, to develop Scheduling strategies for different business and technical needs;

S46. Isolation and security: Consider inter-tenant isolation and application-level security in scheduling decisions, implement role-based access control (RBAC) and hardware resource isolation between tenant workloads, and add security verification and monitoring steps to the policy. , protect the system from malicious behavior;

To summarize: this phase focuses on developing scheduling algorithms that can allocate resources efficiently. It includes designing a flexible policy framework, implementing static and dynamic scheduling algorithms, and optimizing algorithms to ensure that scheduling results meet business requirements. At the same time, isolation and security factors must also be taken into consideration.

The step S5 can be refined into:

S51. Define resource management strategies: analyze the characteristics and business requirements of different types of resources, formulate corresponding resource management strategies, formulate core binding, priority and time slice quotas for CPU scheduling, configure priority and recycling policies for memory, and provide storage Set read and write rate limits, allocate maximum and minimum bandwidth limits for network bandwidth, and ensure efficient and fair resource use while meeting SLA requirements;

S52. Resource allocation algorithm: Develop a finely controlled resource allocation algorithm, which can dynamically allocate resources according to the current system resource usage and workload requirements. The algorithm considers the interdependence and constraints of resources to achieve rapid response and flexible adjustment of resource allocation. ;

S53. Resource optimization cycle: Implement regular resource optimization cycles, identify resource usage patterns and efficiency bottlenecks by analyzing historical data, regularly adjust resource allocation, merge memory fragments, rebalance distribution, etc., to promote resource efficiency and system stability;

S54. Elastic scaling logic: Implement elastic scaling logic, allowing the system to automatically adjust resource allocation based on real-time monitoring data and prediction results. When the load changes, it can increase or decrease computing nodes or service instances in a timely manner to ensure that applications always maintain optimal performance and cost-effectiveness;

S55. Energy efficiency optimization: Develop an energy optimization subsystem to monitor the energy usage of the data center, analyze energy consumption patterns, and reduce energy consumption by adjusting resource usage strategies and scheduling strategies to ensure the green operation of the system and help reduce costs. ;

S56. Persistence and state recovery: Build a state persistence mechanism to regularly save the current state of the system to persistent storage, such as a database or distributed storage system. After a system failure, these persistent state data can be used to quickly restore system work. , minimize the impact of failures on business;

To summarize: In the resource management stage, resource management strategies are formulated, resource allocation algorithms are developed, and resource optimization cycles are introduced. In addition, elastic scaling logic is implemented to adapt to load changes, optimize energy utilization, and ensure system durability and resilience in the face of failures.

The step S6 can be refined into:

S61. Module integration: Integrate the developed modules (such as scheduler, resource manager, monitoring system) into the main system framework to ensure that the interfaces between modules are compatible, and complete necessary integration tests to check whether the data flow and function calls are as required. expected work;

S62. System-wide testing: Conduct end-to-end testing of the entire system, including functional testing, integration testing, system testing and acceptance testing, to verify the core functions, user stories and boundary conditions of the system. In addition, it is necessary to test the system under abnormal conditions. behavior to ensure stability;

S63. Performance evaluation: Evaluate the performance of the system under typical and extreme workloads, including response time, throughput and resource usage efficiency, use standardized performance evaluation tools and methods, and compare it with the performance requirements in the business goals to ensure that it is met predetermined indicators;

S64. Isolation verification: Conduct a security audit on the system to check whether there are code vulnerabilities, misconfigurations and potential security risks, verify the multi-tenant isolation capabilities in the system, and ensure the safe isolation of data and operations of different users or organizations;

S65. User case simulation: Based on typical user scenarios, simulate user operations to verify system functions, ensure the correctness of user processes and business processes, monitor the performance of the system under simulated operations, and adjust system configuration to better serve user needs. ;

S66. Load stress test: By simulating high load conditions and artificially created pressure points, the performance bottlenecks and potential problems of the system in the extreme state are identified, and the system architecture and resource allocation are adjusted based on the test results to ensure the stability and stability of the system in actual operation. reliability;

To summarize: This phase involves integrating previously independently developed modules into a complete system, conducting system-wide testing to ensure that the modules work together correctly, and then evaluating the performance of the system to ensure that it meets established performance standards. . Security and isolation assessments are also conducted and tested by simulating user actions.

The step S7 can be refined into:

S71. Deployment planning: Create a detailed deployment plan and consider various deployment scenarios, such as blue-green deployment to achieve zero downtime updates, or rolling updates to gradually replace older version instances. The plan should include risk assessment and prior notification. , rollback strategy, deployment schedule, and consideration of critical business periods;

S72. Automated deployment process: Develop an automated deployment process, use continuous integration and continuous deployment (CI/CD) pipelines to automatically test, package and deploy new versions, ensuring that this process covers code submission, build, automated testing, deployment and notification , to reduce human error and improve efficiency;

S73. Monitoring strategy implementation: Select monitoring tools based on system key performance indicators (KPIs), establish monitoring strategies, realize real-time resource usage, service operating status and abnormal event monitoring, and build dashboards and alarm systems to facilitate operation and maintenance personnel quickly Respond to any questions;

S74. Fault drills: Regularly arrange fault drill activities to simulate fault scenarios and verify the system's recovery process and backup strategy. These drills help ensure that when a real emergency occurs, the team can take quick and accurate actions to reduce possible business impacts;

S75. Performance tuning: Continuously collect performance data, and perform regular performance tuning of the system based on these data, identify performance bottlenecks, optimize configurations, apply best practices, and update hardware to address performance challenges. At the same time, cost-effectiveness must be considered to find The best balance between resource usage and performance;

S76. Documentation and logs: Maintain detailed system documentation, including design documents, user manuals, operation guides and frequently asked questions (FAQs), so that users and managers can understand and use the system, implement log management policies, and record system operation in detail. Combined with log analysis tools to gain insight into system behavior and quickly diagnose problems;

To summarize: During the deployment phase, we need to develop a detailed deployment plan, automate the deployment process, and deploy monitoring strategies to track resource and application-level performance. Failure drills and backup verification are to establish preparations for possible failures and perform performance tuning based on actual system operation.

The step S8 can be refined into:

S81. User feedback collection: Establish a comprehensive user feedback collection mechanism, including online surveys, user forums, direct visits or customer support feedback and other channels to ensure that system usage, performance bottlenecks, possible improvement suggestions and User satisfaction data, regularly analyze these data, and identify user experience improvement points;

S82. Issue tracking and repair: Build an issue tracking system to continuously monitor, record and classify issues that arise in the system, assign a priority to each issue, and carry out issue and repair work within the corresponding time frame to ensure smooth cross-team communication. , so that relevant departments can work closely to solve problems quickly;

S83. Performance monitoring and optimization: Deploy advanced performance monitoring tools to track system performance indicators in real time, use data analysis and machine learning technology to identify performance trends and potential problems, and based on the analysis results, make timely system adjustments or optimize configurations to maintain optimal performance. ;

S84. Update iteration: Implement a structured update and iteration strategy to enable the system to integrate new features, performance improvements, and security patches, ensure that the update process minimizes disruption to on-site operations, and ensure stability and stability before and after updates through automated testing. compatibility;

S85. Compliance assurance: Regularly check and implement the latest security updates, monitor the security vulnerability database, ensure that the system is updated in a timely manner to prevent potential attacks, and at the same time, maintain the system in compliance with all relevant industry standards and regulatory requirements, and conduct regular compliance audits and evaluations;

S86. Technology trend assessment: Continuously monitor and evaluate emerging technologies and industry trends, examine the benefits they may bring to the system, organize regular internal knowledge sharing meetings and technical lectures, encourage team members to learn new technologies and consider integrating them into system to stay ahead of the curve;

To sum up: the system needs to enter the maintenance phase. At this time, user feedback is collected, problems are tracked and fixed, system performance is continuously monitored, and updates and iterations are performed regularly. Security and compliance are one of the focuses of this phase. At the same time, new technology trends need to be constantly evaluated to decide whether to incorporate them into the system.

Finally, it should be noted that the above-mentioned embodiments are only used to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art will understand that : It is still possible to modify the technical solutions recorded in the foregoing embodiments, or to equivalently replace some or all of the technical features; and these modifications or substitutions do not deviate from the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention. range.

Claims (9)

1. The resource dynamic scheduling technology of the distributed cloud computing system is characterized by comprising the following steps of:

s1, demand analysis: determining the service requirement and technical requirement of the system, evaluating the existing infrastructure and technology, and setting targets and performance indexes;

s2, designing a system architecture: designing the overall structure of the system based on the result of the demand analysis, determining high-level components of the system such as a database, an application server, a load balancer and the like, and the interaction mode among the components, wherein in order to ensure the expandability and the flexibility of the system, a network architecture and a data storage strategy need to be designed and planned;

s3, workload analysis: the process of understanding the type, size and frequency of tasks to be processed by the system predicts resource demands by analyzing historical data and expected workload, and optimizes resource allocation and scheduling strategies according to the data;

s4, developing a scheduling algorithm: according to the result of workload analysis, developing a resource scheduling algorithm to optimize the efficiency of task operation and the resource utilization rate, wherein the resource scheduling algorithm comprises an implementation priority queue, a rule-based engine and an adaptive scheduling algorithm adopting a machine learning technology;

S5, resource management: the monitoring, allocation, optimization and control of physical and virtual resources are realized, wherein the monitoring, allocation, optimization and control of physical and virtual resources comprise that a development function automatically manages the life cycle of the resources, such as allocation, release and expansion, contraction and migration of the resources according to requirements;

s6, system integration verification: combining all independently developed modules and components together and then testing as a complete system, including verifying the functionality and performance of the system and ensuring that the parts can work together to meet design specifications;

s7, monitoring policy implementation: emphasis is placed on implementing a monitoring and alarm system to ensure that system health, performance changes and potential problems can be monitored in real time in a production environment;

s8, continuously optimizing the system: in the late stages of system deployment, it is necessary to continuously monitor system performance and optimize the system according to feedback, including upgrading hardware, updating software, improving scheduling policies, optimizing resource allocation, etc., to ensure that the system can accommodate changing demands and up-to-date technology trends.

2. The resource dynamic scheduling technique of a distributed cloud computing system according to claim 1, wherein: the step S1 may be refined as:

S11, evaluating hardware resources: evaluating existing physical machines, virtual machines, storage solutions and network resources, the focus of this step being on understanding the performance characteristics, expansion capabilities and limitations of existing infrastructure, the evaluation including checking the age of the hardware, maintenance records, past performance history and failure rates to facilitate future scale expansion and resource allocation to make reasonable inferences and decisions;

s12, defining a workload model: determining the type of service and task which the system needs to support, analyzing the load which possibly occurs to the resource according to the characteristics of the application program, such as transaction type system, data processing or real-time analysis, etc., wherein in the step, the expected peak value, average value, steady state and change trend of the load need to be defined in the model;

s13, setting a scheduling target: setting the target to be achieved by the scheduler according to the service requirements and the expected system use, more specifically, if the system needs to respond to the user request quickly, the response time of the minimum task may be prioritized; or if cost savings are a primary goal, may be optimized to increase energy and resource usage efficiency;

s14, determining service level agreement requirements: SLAS defines quality of service and performance metrics promised to users, including, among other things, availability of the system, service response time, time to fail-over, etc.;

S15, establishing monitoring requirements: the monitoring requirements and indexes should be matched with the defined SLAS and performance targets to determine monitoring indexes, wherein the monitoring indexes comprise CPU utilization rate, memory use, network bandwidth, storage IOPS and the like, and monitoring intervals and historical data retention strategies, and the steps can help a system administrator to understand the system state and respond in time when a problem occurs;

s16, standardized resource description: a unified description definition is created for all physical and virtual resources in the system, including the model and core number of the CPU, memory size, storage capacity, network bandwidth, etc., standardization is a key basis for resource allocation and scheduling decisions, ensuring that the system can efficiently and consistently manage and use these resources.

3. The resource dynamic scheduling technique of a distributed cloud computing system according to claim 1, wherein: the step S2 may be refined as:

s21, determining a system architecture mode: selecting a suitable architecture mode, such as a micro-service architecture, a Service Oriented Architecture (SOA), an event driven architecture, or a traditional single application architecture, wherein the selection of the architecture mode is based on demand analysis, and factors such as maintainability, expandability, performance, team experience and the like are considered;

S22, designing a fault-tolerant mechanism: planning redundancy and fault tolerance mechanisms of the system to ensure high availability, this design needs to include data backup schemes, multi-zone deployments, failover and recovery strategies, and implementing stateless components to support horizontal expansion;

s23, planning a data management strategy: determining the storage requirement of data, selecting a proper database system, considering data consistency, integrity and accessibility, and planning data backup, recovery and archiving strategies;

s24, designing a system monitoring solution: based on the monitoring requirements set in the requirement analysis, designing a complete monitoring solution, how to collect, store and analyze monitoring data, how to set an alarm threshold, and selecting monitoring tools and platforms to provide real-time monitoring and performance analysis;

s25, planning a network architecture: the refined network architecture comprises internal network isolation, setting of public and private subnets, use of load balancers and security policies ensuring network communication such as firewalls, intrusion detection systems and encrypted transmissions;

s26, defining a system component and an interface: the API design disclosed herein should facilitate understanding and use of and maintain consistency and version control for each component in the system by defining interfaces and defining communication between internal components, including messaging protocols and data formats.

4. The resource dynamic scheduling technique of a distributed cloud computing system according to claim 1, wherein: the step S3 may be refined as:

s31, data collection and cleaning: designing an efficient data collection mechanism, involving automatically collecting resource utilization data from different services and applications in a distributed environment, deploying a log aggregation tool such as ELK (elastic search, log stack, and Kibana) or Fluentd, and a monitoring system such as promethaus or Datadog, to standardize and centrally collect data, formulate cleaning rules including anomaly detection and correction, time zone synchronization, and data format unification;

s32, feature selection: evaluating the resource usage pattern using data analysis and visualization tools (e.g., pandas and seaban in Python), and determining key features through statistical testing and correlation analysis, implementing machine learning feature selection techniques (e.g., recursive feature elimination), aided with expert knowledge, to optimize the input feature set of the model;

s33, developing an algorithm: based on the selected feature set, various predictive models are explored, including traditional statistical methods (such as ARIMA) and modern machine learning techniques (such as gradient hoisting and neural networks), ensuring that the algorithm development environment is provided with the necessary computing resources, and using an appropriate ML framework (e.g., tensorFlow or scikit-learn);

S34, model training: based on historical data, adopting a machine learning workflow management technology (such as MLflow or Kubeflow) to ensure the traceability and consistency of a model training process, optimizing the model by using parameter searching (such as grid searching or random searching), and evaluating the generalization capability of the model by using cross verification;

s35, model verification: verifying model performance on an independent test data set, measuring prediction accuracy by adopting a proper evaluation index (such as MAE and RMSE), further optimizing model parameters by using a verification result, periodically reviewing model performance according to the change of an application scene, and updating a model to adapt to a new data mode;

s36, integration and deployment: an automated deployment flow for designing and implementing models, including CI/CD pipelining (continuous integration/continuous deployment), so that predictive models can be smoothly integrated into existing dispatch systems, ensuring adequate monitoring and alarm mechanisms to track model performance, and managing versions of models in a production environment.

5. The resource dynamic scheduling technique of a distributed cloud computing system according to claim 1, wherein: the step S4 may be refined as:

s41, strategy framework design: constructing a flexible and extensible scheduling framework supporting plug-in policy exchange, wherein the framework should be capable of being seamlessly integrated with a system architecture, configurable and easy to maintain, and using industry standard design modes (such as policies, observer modes) to assist code separation and ensure modularization of the system;

S42, realizing a static scheduling algorithm: a series of basic static scheduling algorithms, such as First Come First Served (FCFS), round-Robin (Round-Robin) and Fixed Priority (Fixed Priority) scheduling, are written, more specifically, the algorithms should be able to make decisions simply and quickly without taking external changes into account, and standard test cases are created for these algorithms in order to verify their performance and correctness;

s43, a dynamic scheduling mechanism: developing dynamic scheduling algorithms capable of responding to real-time system state changes, wherein the algorithms consider the current resource utilization rate, historical load data, service Level Agreement (SLA), priority and other factors, and a bionic algorithm (such as a genetic algorithm), a heuristic method or a machine learning technology is used for realizing dynamic response and self-adaptive scheduling decision;

s44, algorithm optimization: continuously performing performance analysis and optimization on the existing algorithm, and applying microscopic and macroscopic optimization technologies such as algorithm complexity reduction, multithread parallel processing and resource reservation mechanism, and simultaneously, considering implementation simulation environment to evaluate and optimize the performance of the algorithm under different loads and conditions;

s45, multi-target scheduling support: developing a scheduling algorithm supporting multi-objective decision, wherein the algorithm can balance various factors such as response time, resource utilization rate, energy efficiency, cost and the like, and develop scheduling strategies meeting different business and technical requirements by utilizing multi-objective optimization theory and technology such as Pareto optimization;

S46, isolation and security: taking tenant separation and application level security into consideration in scheduling decision, realizing hardware resource separation between role-based access control (RBAC) and tenant workload, adding security verification and monitoring steps in the policy, and protecting the system from malicious behaviors.

6. The resource dynamic scheduling technique of a distributed cloud computing system according to claim 1, wherein: the step S5 may be refined as:

s51, defining a resource management strategy: analyzing the characteristics and business requirements of different types of resources, formulating corresponding resource management strategies, formulating core binding, priority and time slice quota for CPU scheduling, configuring priority and recovery policies for memory, setting read-write rate limit for storage, allocating maximum and minimum bandwidth limit for network bandwidth, ensuring efficient and fair resource use, and meeting SLA requirements;

s52, a resource allocation algorithm: developing a finely controlled resource allocation algorithm, wherein the resource allocation algorithm can dynamically allocate resources according to the current system resource use condition and the workload demand, and the algorithm considers the interdependence and constraint of the resources to realize quick response and flexibly adjusted resource allocation;

S53, resource optimization cycle: implementing periodic resource optimization cycle, identifying the use mode and efficiency bottleneck of the resource by analyzing the historical data, adjusting the resource allocation at regular time, merging memory fragments, re-balancing distribution and the like so as to promote the resource efficiency and the system stability;

s54, elastic expansion logic: realizing elastic telescopic logic, allowing the system to automatically adjust resource allocation according to real-time monitoring data and a prediction result, and increasing and decreasing computing nodes or service examples timely when the load changes, so as to ensure that the application always keeps optimal performance and cost benefit;

s55, optimizing energy efficiency: the energy optimization subsystem is developed, the energy use condition of the data center is monitored, the energy consumption mode is analyzed, the energy consumption is reduced by adjusting the resource use strategy and the scheduling strategy, the environment-friendly operation of the system is ensured, and the cost is reduced;

s56, persistence and state recovery: and (3) constructing a state persistence mechanism, periodically storing the current state of the system into persistent storage, such as a database or a distributed storage system, and after the system fails, rapidly recovering the system to work by utilizing the persistent state data to minimize the influence of the failure on the service.

7. The resource dynamic scheduling technique of a distributed cloud computing system according to claim 1, wherein: the step S6 may be refined as:

s61, module integration: integrating the developed modules (such as a scheduler, a resource manager and a monitoring system) into a main system framework, ensuring compatibility among the modules, completing necessary integrated test, and checking whether data flow and function call work as expected;

s62, system range test: performing end-to-end tests on the whole system, including functional tests, integration tests, system tests and acceptance tests, verifying the core functions, user stories and boundary conditions of the system, and in addition, requiring testing the behavior of the system under abnormal conditions to ensure stability;

s63, performance evaluation: evaluating the performance of the system under typical and extreme workloads, including response time, throughput, and resource usage efficiency, using standardized performance evaluation tools and methods, and comparing to performance requirements in business objectives, to ensure that predetermined metrics are met;

s64, isolation verification: performing security audit on the system, checking whether code loopholes, error configuration and potential safety hazards exist or not, verifying multi-tenant isolation capability in the system, and ensuring security isolation of data and operations of different users or organizations;

S65, user case simulation: according to typical user scenes, simulating user operation to verify system functions, ensuring correctness of user flows and business flows, monitoring performance of the system under the simulated operation, and adjusting system configuration to better serve user requirements;

s66, load pressure test: by simulating high load conditions and artificially manufacturing pressure points, performance bottlenecks and potential problems of the system in a limit state are identified, system architecture and resource allocation are adjusted according to test results, and stability and reliability of the system in actual operation are ensured.

8. The resource dynamic scheduling technique of a distributed cloud computing system according to claim 1, wherein: the step S7 may be refined as:

s71, deployment planning: creating a detailed deployment plan, considering various deployment scenarios, such as blue-green deployment to realize zero downtime update or rolling update to gradually replace old version examples, wherein the plan should include risk assessment, advance notice, rollback strategy, deployment schedule and consideration of key service period;

s72, an automatic deployment flow: developing an automatic deployment flow, automatically testing, packaging and deploying new versions by utilizing a continuous integration and continuous deployment (CI/CD) pipeline, and ensuring that the process covers code submission, construction, automatic testing, deployment and notification so as to reduce human errors and improve efficiency;

S73, implementing a monitoring strategy: selecting a monitoring tool according to Key Performance Indexes (KPIs) of the system, establishing a monitoring strategy, realizing monitoring of real-time resource use conditions, service running conditions and abnormal events, and constructing an instrument panel and an alarm system so as to facilitate operation and maintenance personnel to quickly respond to any problem;

s74, fault drilling: the fault exercise activities are arranged regularly, fault scenes are simulated, the recovery flow and the backup strategy of the system are verified, the exercises help ensure that when an emergency actually happens, a team can quickly and accurately take action, and possible business influence is reduced;

s75, performance tuning: continuously collecting performance data, performing periodic performance tuning on the system based on the data, identifying performance bottlenecks, optimizing configuration, applying best practices and updating hardware to cope with performance challenges, and simultaneously taking cost benefits into consideration to find the best balance point of resource use and performance;

s76, documents and logs: detailed system documents are maintained, including design documents, user manuals, operation guidelines and common problem Solutions (FAQs), so that users and administrators can understand and use the system, implement log management policies, record system operation conditions in detail, and combine log analysis tools to provide insight into system behavior and quickly diagnose problems.

9. The resource dynamic scheduling technique of a distributed cloud computing system according to claim 1, wherein: the step S8 may be refined as:

s81, user feedback collection: establishing a comprehensive user feedback collection mechanism comprising a plurality of channels such as online investigation, user forum, direct access or customer support feedback and the like, ensuring that data about system use conditions, performance bottlenecks, possible improvement suggestions and user satisfaction can be collected, periodically analyzing the data, and identifying user experience improvement points;

s82, problem tracking and repairing: constructing a problem tracking system to continuously monitor, record and classify problems in the system, distributing priority to each problem, and developing problems and repairing work in a corresponding time frame to ensure smooth inter-team communication, so that related departments can closely cooperate to rapidly solve the problems;

s83, performance monitoring optimization: deploying an advanced performance monitoring tool to track system performance indexes in real time, determining performance trends and potential problems by using data analysis and machine learning technologies, and timely performing system adjustment or optimal configuration to maintain an optimal performance state based on analysis results;

S84, updating iteration: implementing a structured updating and iteration strategy, so that the system can integrate new characteristics, performance improvement and security patches, ensure that the updating process minimizes interference to field operation, and ensure stability and compatibility before and after updating through automatic testing;

s85, compliance assurance: periodically checking and implementing the latest security update, monitoring a security vulnerability database, ensuring that the system is updated in time to prevent potential attacks, and simultaneously keeping the system to meet all relevant industry standards and supervision requirements for periodic compliance audit and evaluation;

s86, technical trend evaluation: emerging technologies and industry trends are continually monitored and assessed, looking at their possible benefits to the system, organizing regular internal knowledge sharing conferences and technical lectures, encouraging team members to learn new technologies, and considering their integration into the system to maintain leading advantages.