Calculation power optimization distribution system based on remote scheduling

Abstract

The invention relates to the technical field of computing power distribution, in particular to a remote scheduling-based computing power optimal distribution system, which comprises a task association analysis module, a remote scheduling optimization module, a micro-service monitoring module, a dynamic resource adjustment module, a computing power monitoring module and a resource synchronous integration module. According to the method, the task dependency relationship is accurately mastered through refined data flow tracking and task interaction mode analysis, so that task scheduling better meets actual requirements, the accuracy and response speed of resource configuration are improved, and the task scheduling efficiency is improved through real-time monitoring and dynamic adjustment, task queue management optimization and resource adjustment according to execution conditions. Resource idleness and congestion are effectively reduced, resource load changes can be quickly responded through real-time resource monitoring and dynamic adjustment, resource allocation is automatically adjusted, the system operation efficiency and the computing power utilization rate are remarkably improved, configuration is optimized through assessment of the resource use condition through computing power monitoring, it is ensured that resources are most effectively utilized when needed, and the cost efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of computing power allocation, and particularly to a computing power optimization allocation system based on remote scheduling. Background Art

[0002] Computing power allocation, also known as computing resource allocation or resource scheduling, refers to the process of rationally configuring and optimizing computing resources among multiple computing tasks and applications. It mainly involves resource management and scheduling strategies, with the aim of improving the utilization efficiency of computing resources, reducing energy consumption, and ensuring that tasks are met according to priorities and requirements. The core technologies include virtualization technology, cloud computing resource management, load balancing, and prediction algorithms, etc. Through efficient computing power allocation, the system can dynamically adjust resource allocation according to the computing requirements of tasks, the availability of resources, and policy constraints, thus supporting large-scale and high-concurrency computing tasks.

[0003] Among them, the computing power optimization allocation system based on remote scheduling is an efficient computing resource management system, aiming to remotely and dynamically schedule and optimize the allocation of computing resources. Its main purpose is to optimize resource allocation through remote monitoring and scheduling algorithms to maximize computing efficiency and reduce resource idleness. It is particularly suitable for cloud computing environments and large data centers, where resource requirements change frequently and rapid response is needed. By real-time monitoring resource usage and computing requirements, it automatically adjusts resource allocation to support the performance requirements of various compute-intensive applications. In addition, it can also ensure that the performance of critical applications is not affected according to preset business priorities and policies, while optimizing energy consumption and cost efficiency.

[0004] Existing technologies usually show static and inflexible resource allocation when dealing with large-scale high-concurrency computing tasks. Conventional resource management and scheduling strategies fail to respond in a timely manner to the real-time changes of tasks, resulting in the failure to optimize resource allocation in real time, which affects computing efficiency and resource utilization rate. For example, in the face of sudden loads or irregular task requirements, virtualization technology has a slow response in resource reallocation, easily causing resource waste or shortage. In addition, the limitations of load balancing and resource prediction technologies, such as inaccurate prediction algorithms, may lead to unstable performance of critical applications or resource overabundance. These problems are particularly prominent in cloud computing environments and big data centers. Unreasonable resource allocation increases energy consumption, may cause computing delays, and reduces service quality and user experience. Summary of the Invention

[0005] The purpose of the present invention is to solve the drawbacks existing in the prior art, and propose a computing power optimization allocation system based on remote scheduling.

[0006] To achieve the above purpose, the present invention adopts the following technical solutions: The computing power optimization allocation system based on remote scheduling includes: The task correlation analysis module collects the interaction data between computing tasks, conducts data flow tracking, analyzes the task interaction patterns, records the data transmission time and frequency between tasks, updates the influence relationship between tasks, and generates a task dependency list; The remote scheduling optimization module schedules the task execution order according to the task dependency list, monitors the task execution status in real time, adjusts the priorities of the queued tasks, responds to task completion and delay events, dynamically rearranges the task queue, and generates an optimized task execution plan; The microservice monitoring module tracks the resource usage of each microservice unit based on the optimized task execution plan, counts the CPU occupancy, memory usage, and network input / output, updates the resource consumption record table according to the data, and generates a real-time resource monitoring log; The dynamic resource adjustment module compares according to the judgment threshold based on the resource monitoring log, identifies the high-load areas and low-load areas of resources, synchronously implements the preset resource allocation, dynamically transfers resources from the low-load areas to the high-load areas, and generates an adjusted resource distribution state; The computing power monitoring module records and analyzes the efficiency of each computing node according to the adjusted resource distribution state, counts the resource consumption of the nodes, evaluates the adaptability of the computing power usage, updates the computing power usage record, and generates a computing power usage efficiency report; The resource synchronization and integration module adjusts the resource configuration of each computing node in combination with the computing power usage efficiency report, performs resource configuration synchronization processing, checks the consistency and uniformity of the resource configuration, and generates a resource synchronization status record.

[0007] As a further solution of the present invention, the task dependency list includes task types, interaction frequencies, and data flow directions; the optimized task execution plan includes task order, priority setting records, and scheduling response times; the real-time resource monitoring log includes CPU usage records, memory usage records, and network IO data; the adjusted resource distribution state includes a resource load map, high and low load areas, and resource flow paths; the computing power usage efficiency report includes resource consumption data, efficiency scores, and node adaptability analysis results; the resource synchronization status record includes a configuration consistency report, resource uniformity analysis records, and synchronization processing logs.

[0008] As a further solution of the present invention, the task correlation analysis module includes a data collection sub-module, an interaction analysis sub-module, and a dependency list sub-module; The data collection sub-module, based on the interaction data between tasks, uses timestamp recording and event log collection, conducts data flow marking and continuous tracking, and confirms data integrity through network traffic analysis, generating an interaction data set; The interaction analysis sub-module analyzes the timestamps and frequencies of data transmission through the interaction dataset, compares the packet sizes and transmission rates, identifies key interaction patterns, and obtains the interaction pattern analysis results; The dependency list sub-module gradually updates and optimizes the interdependency relationships between tasks based on the interaction pattern analysis results, constructs a task dependency relationship graph through dependency path analysis, and generates a task dependency relationship list.

[0009] As a further solution of the present invention, the remote scheduling optimization module includes a task scheduling sub-module, a status monitoring sub-module, and a dynamic adjustment sub-module; The task scheduling sub-module sorts each task through dependency analysis and priority evaluation based on the task dependency relationship list, and at the same time assigns the optimal start time to each task in combination with resource availability, generating a scheduled execution order; The status monitoring sub-module collects the execution data of each task in real time through sensors based on the scheduled execution order, analyzes the data and monitors the task progress in real time, identifies and discovers potential delays, and generates a task progress monitoring result; The dynamic adjustment sub-module analyzes the current execution situation and the remaining task queue based on the task progress monitoring result, adjusts the priorities and scheduling strategies of future tasks, matches the real-time changes, and generates an optimized task execution plan.

[0010] As a further solution of the present invention, the microservice monitoring module includes a resource usage tracking sub-module, a resource record update sub-module, and a monitoring log generation sub-module; The resource usage tracking sub-module extracts the task allocation information of the microservice unit based on the optimized task execution plan, analyzes the resource call parameters of each microservice unit, analyzes the CPU occupancy periodicity, statistically accumulates the memory consumption, retrieves the network input and output data and uploads and downloads the traffic, and obtains the resource usage statistical data; The resource record update sub-module extracts the single values of resource usage based on the resource usage statistical data, separates the data fields of different resource categories and classifies the values, matches the field structure of the existing resource consumption record table and supplements and updates the data, and establishes the resource consumption update data; The real-time monitoring log generation sub-module reads the updated resource record content based on the resource consumption update data, extracts the dynamic change values of CPU, memory, and network traffic, performs time series arrangement and formatting conversion, and generates a real-time resource monitoring log.

[0011] As a further solution of the present invention, the dynamic resource adjustment module includes a resource load identification sub-module, a resource dynamic transfer sub-module, and a resource distribution status generation sub-module; The resource load identification sub-module parses the load metrics of resource nodes based on the real-time resource monitoring log, performs comparative calculations on resource load thresholds, marks the resource node areas that exceed the high load threshold and are lower than the low load threshold, and generates a resource load status identifier. The resource dynamic transfer sub-module matches the preset resource allocation rules based on the resource load status identifier, extracts the available resource quantity in the low load area and records the corresponding nodes, identifies the resource gap in the high load area and compares the resource data in the low load area, and performs resource transfer and allocation operations according to the resource requirements and allocation rules, generating resource transfer execution data. The resource distribution status generation sub-module statistically analyzes the current resource allocation situation of each node based on the resource transfer execution data, merges the node allocation status, and adjusts the relevant records in the resource distribution table to generate an adjusted resource distribution status.

[0012] As a further solution of the present invention, the computing power monitoring module includes a computing power efficiency recording sub-module, a computing power adaptation evaluation sub-module, and a computing power usage report generation sub-module. The computing power efficiency recording sub-module parses the task completion time and resource consumption of nodes based on the adjusted resource distribution status, calculates the task execution efficiency of each node, conducts a comparative analysis of task efficiency and resource consumption, and generates node efficiency record data. The computing power adaptation evaluation sub-module extracts the matching values of task efficiency and resource consumption based on the node efficiency record data, compares the differences between node computing power allocation and task requirements, and conducts item-by-item evaluations on the computing power usage adaptation of the different nodes, generating a computing power adaptation evaluation result. The computing power usage report generation sub-module summarizes the adaptation situation data of nodes, statistically analyzes the usage situation of the overall computing power distribution, integrates the computing power efficiency and adaptation evaluation content into a structured text, and generates a computing power usage efficiency report.

[0013] As a further solution of the present invention, the task execution efficiency of each node is calculated according to the formula: ; wherein, represents the task execution efficiency of the th node, represents the resource consumption of the th node, represents the time required for the th node to complete the task, represents the total amount of resources allocated to the th node, represents the resource requirement difference of the th node.

[0014] As a further solution of the present invention, the resource synchronization and integration module includes a resource configuration adjustment sub-module, a resource synchronization processing sub-module, and a resource status verification sub-module; Based on the computing power utilization efficiency report, the resource configuration adjustment sub-module extracts the resource allocation data of the nodes, counts the redundant resource amounts of the nodes with low resource utilization and the resource requirements of the nodes with high resource utilization, matches the resource allocation requirements of the nodes and adjusts the resource configuration, and generates adjusted resource configuration data; Based on the adjusted resource configuration data, the resource synchronization processing sub-module extracts the allocation parameters of the node resources and matches them item by item, calls the synchronization mechanism to confirm that the resource allocation operations are consistent, records the resource status of the nodes after the allocation is completed, and generates a resource synchronization configuration record; Based on the resource synchronization configuration record, the resource status verification sub-module extracts the resource allocation status of each node, calculates the uniformity index of the resource allocation of each node, analyzes and marks the abnormal allocation status in the node resource distribution, and verifies the resource synchronization completion status of all nodes, and generates a resource synchronization status record.

[0015] As a further solution of the present invention, the uniformity index of the resource allocation is calculated according to the formula: ; wherein, represents the resource allocation uniformity index of the th node, represents the allocation amount of the th node for the th type of resource, represents the average allocation amount of all nodes for the th type of resource, represents the standard deviation of the th type of resource, is the total number of resource types.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In the present invention, through refined data flow tracking and task interaction mode analysis, an accurate grasp of the task dependency relationship is achieved, so that the task scheduling is more in line with the actual requirements, the accuracy and response speed of resource configuration are improved. Through real-time monitoring and dynamic adjustment, the task queue management is optimized, allowing resources to be adjusted according to the execution situation, effectively reducing resource idleness and congestion. The real-time resource monitoring and dynamic adjustment can quickly respond to resource load changes and automatically adjust resource allocation, significantly improving the system operation efficiency and computing power utilization rate. The computing power monitoring optimizes the configuration by evaluating the resource usage situation, ensures that resources are most effectively utilized when needed, and improves the cost efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is the system flow chart of the present invention; Figure 2 is the system module diagram of the present invention; Figure 3 is the flow chart of the task association analysis module of the present invention; Figure 4 is the flow chart of the remote scheduling optimization module of the present invention; Figure 5 is the flow chart of the microservice monitoring module of the present invention; Figure 6 is the flow chart of the dynamic resource adjustment module of the present invention; Figure 7 is the flow chart of the computing power monitoring module of the present invention; Figure 8 is the flow chart of the resource synchronization and integration module of the present invention. Detailed implementation manners

[0018] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0019] In the description of the present invention, it should be understood that the orientation or positional relationships indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. are based on the orientation or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as limiting the present invention. In addition, in the description of the present invention, "a plurality of" means two or more, unless otherwise specifically defined.

[0020] Please refer to Figure 1 , the computing power optimization allocation system based on remote scheduling includes: The task association analysis module collects the interaction data between computing tasks, performs data flow tracking, analyzes the task interaction mode, records the data transmission time and frequency between tasks, updates the influence relationship between tasks, and generates a task dependency list; The remote scheduling optimization module schedules the task execution order according to the task dependency list, monitors the task execution status in real time, adjusts the priority of the queued tasks, responds to task completion and delay events, dynamically rearranges the task queue, and generates an optimized task execution plan; The microservice monitoring module tracks the resource usage of each microservice unit based on the optimized task execution plan, statistics CPU occupancy, memory usage, and network input / output, updates the resource consumption record table according to the data, and generates real-time resource monitoring logs; The dynamic resource adjustment module compares according to the judgment threshold based on the resource monitoring logs, identifies high-load and low-load resource areas, synchronously implements the preset resource allocation, dynamically transfers resources from the low-load area to the high-load area, and generates the adjusted resource distribution status; The computing power monitoring module records and analyzes the efficiency of each computing node according to the adjusted resource distribution status, statistics the resource consumption of the nodes, evaluates the adaptability of computing power usage, updates the computing power usage record, and generates a computing power usage efficiency report; The resource synchronization and integration module adjusts the resource configuration of each computing node in combination with the computing power usage efficiency report, performs resource configuration synchronization processing, checks the consistency and uniformity of the resource configuration, and generates a resource synchronization status record.

[0021] The task dependency list includes task types, interaction frequencies, and data flow directions; the optimized task execution plan includes task sequences, priority setting records, and scheduling response times; the real-time resource monitoring logs include CPU usage records, memory usage records, and network IO data; the adjusted resource distribution status includes resource load diagrams, high- and low-load areas, and resource flow paths; the computing power usage efficiency report includes resource consumption data, efficiency scores, and node adaptability analysis results; the resource synchronization status record includes configuration consistency reports, resource uniformity analysis records, and synchronization processing logs.

[0022] Please refer to Figure 2 and Figure 3 , the task association analysis module includes a data collection sub-module, an interaction analysis sub-module, and a dependency list sub-module; The data collection sub-module, based on the interaction data between tasks, uses timestamp records and event log collection, performs data flow marking and continuous tracking, and confirms data integrity through network traffic analysis, generating an interaction data set; Based on the interaction data between tasks, the data collection sub-module first sorts the data stream through the timestamp records of the data, determines the time points and order of each data packet, and sequentially records the interaction logs generated by the tasks in chronological order. Each log record contains information such as the identifier, time point, size, and transmission node of the data packet. By comparing the timestamps, the transmission link of the data packet is gradually established. To ensure data integrity, a packet-by-packet comparison method is used to verify the validity of the data. By reading the content bytes of the data packet and performing segment-by-segment verification with the preset header information, the comparison content includes the header length field, data check field, and total number of bytes of the actual data, etc. For lost or abnormally content data packets, the transmission request is directly re-initiated using the timestamp information, and the complete transmission chain is repaired by re-acquiring the lost data blocks. All the verification data and re-collected data are merged to form a unified data stream file. After sorting the generated file and confirming the time sequence of data transmission, a complete interaction data set is output.

[0023] The interaction analysis sub-module analyzes the timestamps and frequencies of data transmission through the interaction data set, and compares the data packet sizes and transmission rates to identify key interaction patterns and obtain the interaction pattern analysis results; The interaction analysis sub-module conducts a detailed analysis of the timestamps and transmission frequencies in the interaction data set. First, it calculates the time interval of data transmission based on the timestamp information of each data packet, that is, the time difference between two data transmissions. After the time interval calculation is completed, the data packet sizes are grouped corresponding to the time interval values, and the data transmission rate is analyzed by the ratio of the data packet size to the time. During this process, the time intervals need to be recorded one by one, and the transmission time sequence is divided into several time slices, and the average transmission rate is extracted in each time slice. At the same time, the change situation of the data packet sizes is also recorded in segments. After statistically classifying the sizes of all data packets, the interaction characteristics are analyzed by comparing the changes in the transmission rates within adjacent time slices. The identification of key interaction patterns is based on the stability of the rate and the abnormality of the transmission frequency. By observing the distribution characteristics of the data packets segment by segment, the areas that do not conform to the typical patterns are found, and the interaction patterns of the areas are separately extracted and marked, and finally summarized as the interaction pattern analysis results.

[0024] Based on the interaction pattern analysis results, the dependency list sub-module gradually updates and optimizes the mutual dependency relationships between tasks, constructs a task dependency relationship graph through dependency path analysis, and generates a task dependency relationship list; Based on the results of the interactive mode analysis, the dependency list sub-module gradually optimizes the dependency relationships between tasks. First, it is necessary to sort out the interconnections between all tasks, sort the tasks according to their priorities, and determine the main task nodes in the dependency path. The analysis of task nodes is based on the actual relationships between tasks, and the strength of the dependency relationships is gradually defined. For each task node, the relationship chain between its input and output is extracted, and the dependency strength values in the path are recorded. The dependency paths are numbered according to the strength of the paths. After the processing is completed, the length, the number of branches, and the complexity of the intersections of each path are calculated. According to the path optimization principle, the task relationships are rearranged. By adjusting the path length and the node order, an optimized dependency path graph is constructed. Finally, combined with the actual associations between task nodes in the path graph, a task dependency relationship list is generated, and each list record is marked according to the optimization results, finally forming a clear and optimized dependency structure for the execution coordination of subsequent tasks.

[0025] Please refer to Figure 2 and Figure 4 , the remote scheduling optimization module includes a task scheduling sub-module, a status monitoring sub-module, and a dynamic adjustment sub-module; Based on the task dependency relationship list, the task scheduling sub-module sorts each task through dependency analysis and priority evaluation. At the same time, it assigns the optimal start time for each task in combination with resource availability, and generates a scheduled execution order; Based on the task dependency relationship list, the task scheduling sub-module determines the priorities of tasks according to the hierarchical relationships of the dependency paths by analyzing the dependency paths between tasks item by item. At the same time, it counts the input and output resource requirements of each task, and records the resource types and resource occupation durations required by the tasks. On the basis of evaluating the task priorities, combined with the actual availability of resources, the optimal start time of each task is gradually calculated. The allocation of the start time is based on the results of resource conflict detection. For tasks that may cause resource competition, the conflicts are resolved by adjusting the start time or the resource allocation order, and it is ensured that high-priority tasks obtain the resource allocation right first. After the start times and orders of all tasks are determined, a scheduled execution order of the tasks is generated according to the dependency levels and resource constraint conditions of the tasks, and it is ensured that the execution order of the tasks meets the requirements of the dependency relationships. Finally, a complete sorted list of tasks is output for subsequent scheduling and execution.

[0026] Based on the scheduled execution order, the status monitoring sub-module collects the execution data of each task in real time through sensors, analyzes the data and monitors the task progress in real time, identifies and discovers potential delays, and generates task progress monitoring results; The status monitoring sub-module, based on the scheduled execution order, collects the execution data of each task in real time through sensors deployed in the execution environment. The data collected by the sensors includes the current status, completion progress, execution time consumed, and resource usage of the task. After processing the collected raw data, real-time status information of the task is generated. Gradually compare and analyze the execution status of each task, including comparing whether the planned progress is consistent with the actual completion progress, calculating the delay time of the task, and the difference value of the resources actually consumed by the task. Based on the progress status of the task and the execution order among multiple tasks, identify potential delays in advance, including situations where the task waiting time is too long, resources are insufficient, or the task execution efficiency decreases. Record these potential problems item by item and generate the task progress monitoring result, providing accurate status information and data support for subsequent dynamic adjustment.

[0027] Based on the task progress monitoring result, the dynamic adjustment sub-module analyzes the current execution situation and the remaining task queue, adjusts the priorities and scheduling strategies of future tasks, matches the real-time changes, and generates an optimized task execution plan; Based on the task progress monitoring result, the dynamic adjustment sub-module reorders the priorities of future tasks by analyzing the completion status of the currently executing tasks and the priorities and dependency paths of the tasks in the remaining task queue. First, classify the delayed tasks in the monitoring result, including delays caused by insufficient resources and delays caused by decreased task efficiency, and adopt different adjustment strategies for different types of delays. For the situation of insufficient resources, reallocate the resource occupation order of the tasks, and give priority to ensuring the tasks on the critical path; for the situation of decreased task efficiency, adjust the start time and execution order of the remaining tasks. Combining the real-time status of the current task execution, gradually optimize the scheduling order of future tasks, dynamically sort the uncompleted task queue, and generate an optimized task execution plan based on the adjusted task priorities for real-time updating of the task scheduling strategy.

[0028] Please refer to Figure 2 and Figure 5 , the microservice monitoring module includes a resource usage tracking sub-module, a resource record update sub-module, and a monitoring log generation sub-module; Based on the optimized task execution plan, the resource usage tracking sub-module extracts the task allocation information of the microservice unit, parses the resource call parameters of each microservice unit, analyzes the CPU occupancy periodicity, counts the cumulative memory consumption, retrieves the network input and output data and uploads and downloads the traffic, and obtains the resource usage statistical data; Based on the optimized task execution plan, the resource usage tracking sub-module extracts the task assignment information related to each microservice unit, and parses each resource call parameter involved in the assigned tasks one by one, including the number of CPU cores, call frequency, memory allocation amount, and the traffic size of uploaded and downloaded data. By parsing the microservice call cycle, the usage cycle and idle cycle of each CPU are recorded, the occupation time of each CPU during the call process is accumulated, and the total occupation duration is statistically calculated by microservice unit. For memory consumption, the memory size allocated for each call is compared with the released amount after the call ends, and the total accumulated memory consumption is recorded. The network input and output data traffic are collected in real time, and the resource consumption is statistically calculated by distinguishing the byte sizes of uploaded and downloaded traffic respectively. Finally, based on all the parsed and statistical results, complete resource usage statistical data is obtained for subsequent resource recording and optimization.

[0029] Based on the resource usage statistical data, the resource record update sub-module extracts the single-item values of resource usage, separates the data fields of different resource categories and classifies the values, matches the field structure of the existing resource consumption record table and supplements and updates the data to establish resource consumption update data; Based on the resource usage statistical data, the resource record update sub-module gradually extracts each resource consumption value recorded in the statistical data, separates various resource data, including fields such as CPU usage time, memory consumption size, network upload traffic, and download traffic. The separated data fields are classified and grouped, and the consumption values of the same type of resources are stored in groups. At the same time, the field structure in the existing resource consumption record table is matched, the consistency of the field names and unit formats is checked, and the new data is supplemented into the corresponding fields of the record table to keep the table structure complete. For the fields that already exist in the original record table, each item is updated by comparing the difference values between the new data and the original data to ensure that the consumption data can reflect the actual usage of resources in real time. Finally, updated data on resource consumption is generated to provide the latest record for subsequent monitoring.

[0030] Based on the resource consumption update data, the real-time monitoring log generation sub-module reads the updated resource record content, extracts the dynamic change values of CPU, memory, and network traffic, performs time series arrangement and formatting conversion, and generates real-time resource monitoring logs; The real-time monitoring log generation sub-module updates data based on resource consumption, reads the updated resource record data content line by line, dynamically extracts the CPU usage rate, memory consumption, and network upload and download traffic in the records, and arranges the dynamic change values of each resource in chronological order by analyzing the timestamp and data change value of each record. During the arrangement process, the resource change data within the time interval is formatted and transformed, including converting the numerical values to standard units and adjusting the time fields to a unified format, ensuring that the data can be further processed and parsed. The data after arrangement and transformation is recorded line by line into the real-time monitoring log. The log contains the CPU usage, memory occupancy, and network traffic changes at each time point. The finally generated real-time resource monitoring log can comprehensively reflect the dynamic changes of resources and provide a reliable basis for the real-time analysis of resource management.

[0031] Please refer to Figure 2 and Figure 6 , the dynamic resource adjustment module includes a resource load identification sub-module, a resource dynamic transfer sub-module, and a resource distribution status generation sub-module; The resource load identification sub-module, based on the real-time resource monitoring log, analyzes the load metrics of resource nodes, performs comparison calculations on resource load thresholds, and marks the resource node areas that exceed the high load threshold and are below the low load threshold, generating a resource load status identifier; The resource load identification sub-module, based on the real-time resource monitoring log, by analyzing the load metrics such as the CPU utilization rate, memory occupancy rate, and network traffic transmission volume of resource nodes line by line, compares the real-time load situation of each node with the preset load threshold. For high load situations, it extracts the resource nodes that exceed the high load threshold and records their relevant resource usage information, including specific load parameters, duration, and node locations; for low load situations, it identifies the resource nodes that are below the low load threshold and also records their resource remaining situations and node areas. All the marked high load and low load nodes are divided into areas according to the load status, generating a resource load status identifier for subsequent dynamic resource allocation. The identifier content includes the load type, node location, and its load status to clearly show the abnormal status of resource distribution.

[0032] The resource dynamic transfer sub-module, based on the resource load status identifier, matches the preset resource allocation rules, extracts the available resource quantity in the low load area and records the corresponding nodes, identifies the resource gap in the high load area and compares it with the resource data in the low load area, and performs resource transfer and allocation operations according to the resource requirements and allocation rules, generating resource transfer execution data; Based on the resource load status identifier, the resource dynamic transfer sub-module analyzes the available resources in the low-load area, extracts the remaining CPU, memory, and network bandwidth in the node, and records the node location and remaining resource parameters. Analyze the resource requirements recorded in the high-load area, including the required CPU quantity, memory size, and network traffic gap. By comparing the resource requirements in the high-load area with the remaining resources in the low-load area one by one, prioritize the high-load nodes according to the preset resource allocation rules, and retrieve resources from the low-load area according to the priority to perform the dynamic transfer and allocation operation of resources. Record the specific parameters of each resource allocation during the transfer process, including the type, quantity, source node, and target node of the transferred resources. After all transfer operations are completed, generate resource transfer execution data to reflect the specific execution situation of the current resource transfer.

[0033] Based on the resource transfer execution data, the resource distribution status generation sub-module counts the current resource allocation of each node, combines the node allocation status, and adjusts the relevant records in the resource distribution table to generate the adjusted resource distribution status; Based on the resource transfer execution data, the resource distribution status generation sub-module counts the resource allocation of each node, checks the current CPU, memory, and network bandwidth usage of each node one by one, and updates the resource status after transfer to the node allocation record. By combining the resource allocation status of all nodes, analyze the balance of the overall resource distribution, and adjust and correct the relevant fields in the resource distribution table to ensure the accuracy and consistency of the resource allocation record. Finally, generate the adjusted resource distribution status, record the resource allocation of each node according to the time series to reflect the current resource distribution status and node load level, and provide a reliable data basis for subsequent resource scheduling and optimization. The adjusted resource distribution status covers the allocation information of all nodes and presents the distribution in a unified format.

[0034] Please refer to Figure 2 and Figure 7 , the computing power monitoring module includes a computing power efficiency recording sub-module, a computing power adaptation evaluation sub-module, and a computing power usage report generation sub-module; Based on the adjusted resource distribution status, the computing power efficiency recording sub-module analyzes the node task completion time and resource consumption, calculates the task execution efficiency of each node, conducts a comparative analysis of task efficiency and resource consumption, and generates node efficiency record data; The task execution efficiency of each node is calculated according to the formula: ; For calculation, where represents the task execution efficiency of the th node, indicating the effectiveness of the node in completing its assigned tasks within the given resources and time, Represents the resource consumption of the th node, specifically including but not limited to CPU usage and memory usage. Represents the time required for the th node to complete the task, and calculates the total time from task assignment to task completion. Represents the total amount of resources allocated to the th node, covering all computing and storage resources allocated to this node. Represents the resource requirement difference of the th node, that is, the difference between the actual resources required by the node and the allocated resources.

[0035] : Resource consumption of node . This parameter is directly obtained through the resource monitoring system and reflects the total CPU and memory resources consumed by the node during the execution of a specific task. For example, node consumed a total of 150 units of CPU time and 1024MB of memory during the task execution.

[0036] : Time required for node to complete the task. This value is recorded through the task management system, which is the time interval from task startup to completion. Assume that node i takes 240 seconds to complete the task.

[0037] : Total amount of resources allocated to node . This is the data obtained from the resource allocation log, including all computing and storage resources allocated to this node. Assume that node is allocated 200 units of CPU time and 2048MB of memory.

[0038] : Resource requirement difference of node , that is, the difference between the actual resources required and the allocated resources. This value is calculated through a resource requirement analysis tool. For example, the actual requirements of node are 180 units of CPU time and 1500MB of memory, thus units, and the calculation method is: ; Substitute the specific values into the formula for calculation: ; First, calculate : ; Next, find the absolute value inside the square root and its square root: ; Substitute these values into the formula: ; The result shows that the task execution efficiency of node under the given resource and time conditions is 3020.13, which reflects the resource utilization and time management efficiency of the node when processing assigned tasks. A high value indicates that the node can complete more work with fewer resources and time, which is crucial for optimizing resource allocation and improving the performance of the overall computing network.

[0039] Based on the node efficiency record data, the computing power adaptation evaluation sub-module extracts the matching values of task efficiency and resource consumption, compares the differences between the computing power allocation of the node and the task requirements, and evaluates item by item the adaptation of the computing power usage of the different nodes, generating a computing power adaptation evaluation result; Based on the node efficiency record data, the computing power adaptation evaluation sub-module extracts the matching values of the task execution efficiency and resource consumption of each node. By comparing the efficiency values with the node resource allocation situation, it analyzes one by one whether the computing power allocation of the node meets the actual task requirements. For nodes with excessive computing power allocation but low task efficiency, it records the specific task types, resource consumption amounts, and efficiency matching situations; for nodes with insufficient computing power allocation but high task efficiency, it analyzes the upper limit of resource consumption and records the bottleneck points of resource allocation. The computing power usage situations of all different nodes will be evaluated item by item, marking the adaptation status of their computing power usage, including labels such as good adaptation, resource waste, or resource shortage, and at the same time recording the main reasons affecting the adaptation. Finally, a computing power adaptation evaluation result is generated to comprehensively evaluate the rationality of the node's computing power allocation.

[0040] Based on the computing power adaptation evaluation result, the computing power usage report generation sub-module summarizes the adaptation situation data of the nodes, statistics the usage situation of the overall computing power distribution, integrates the computing power efficiency and adaptation evaluation content into structured text, and generates a computing power usage efficiency report; Based on the computing power adaptation evaluation result, the computing power usage report generation sub-module summarizes the computing power adaptation status and efficiency matching situations of all nodes, classifies the nodes into three categories: good adaptation, resource waste, and resource shortage, and statistics the number and distribution ratio of nodes in each category. It summarizes the usage situation of the overall computing power distribution, analyzes the balance and overall efficiency of resource usage, and calculates the utilization rate of the overall resources. By integrating the detailed content of the node task execution efficiency, resource consumption amount, and adaptation situation into structured text, it records the main characteristics, adaptation status, and related data of each type of node. The finally generated computing power usage efficiency report contains comprehensive data on task execution efficiency, analysis results of computing power allocation adaptability, and statistical information on overall computing power usage, providing detailed basis for further optimizing computing power allocation and resource usage.

[0041] Please refer toFigure 2 and Figure 8 The resource synchronization and integration module includes a resource configuration adjustment sub-module, a resource synchronization processing sub-module, and a resource status verification sub-module; Based on the computing power utilization efficiency report, the resource configuration adjustment sub-module extracts the resource allocation data of the nodes, counts the redundant resources of the nodes with low resource utilization and the resource requirements of the nodes with high resource utilization, matches the resource allocation requirements of the nodes and adjusts the resource configuration, and generates the adjusted resource configuration data; Based on the computing power utilization efficiency report, the resource configuration adjustment sub-module extracts the resource usage of each node from the resource allocation data of the nodes, counts the resource redundancy of the nodes with low resource utilization one by one, including the number of idle CPU cores, the amount of unoccupied memory, and the network bandwidth margin, and at the same time records the resource requirements of the nodes with high resource utilization, including the number of missing CPU cores, the amount of memory to be supplemented, and the additional network traffic capacity required. By comparing the redundant resources of the nodes with low resource utilization with the resource requirements of the nodes with high resource utilization, the allocation requirements between the two are matched, the target nodes and allocation parameters for resource reallocation are determined. According to the matching results, the resource configuration is adjusted, and the flow direction, adjustment quantity, and node change situation of the resources are recorded for each allocation adjustment. Finally, the adjusted resource configuration data is generated for further synchronization and verification operations.

[0042] Based on the adjusted resource configuration data, the resource synchronization processing sub-module extracts the allocation parameters of the node resources and matches them item by item, calls the synchronization mechanism to confirm that the resource allocation operations are consistent, and records the resource status of the nodes after the allocation is completed, generating a resource synchronization configuration record; Based on the adjusted resource configuration data, the resource synchronization processing sub-module extracts the resource allocation parameters of each node item by item from it, including the CPU allocation quantity, the memory allocation size, and the network bandwidth parameters, and checks the resource allocation of each node item by item. Call the synchronization mechanism of resource allocation to ensure the consistency of operation of the allocation parameters among all nodes. Record the start node, target node, and completion time of resource allocation through synchronization operations, and record the resource status of each node after each allocation operation item by item, including the remaining resources of the node and the resource occupation situation of the current task. After the resource synchronization is completed, a resource synchronization configuration record is generated, including the complete information of the resource allocation operation and the updated status of the node resources, providing data support for subsequent resource status verification.

[0043] Based on the resource synchronization configuration record, the resource status verification sub-module extracts the resource allocation status of each node, calculates the uniformity index of the resource allocation of each node, analyzes and marks the abnormal allocation status in the node resource distribution, and verifies the resource synchronization completion situation of all nodes, generating a resource synchronization status record; The uniformity index of resource allocation, according to the formula: ; Perform calculations, where represents the resource allocation uniformity index of the th node, which is used to evaluate the balance degree of the node in resource allocation, represents the allocation amount of the th node for the th type of resource, such as CPU time, memory size, etc., represents the average allocation amount of all nodes for the th type of resource. The calculation method is the arithmetic mean of the allocation amounts of all nodes for the th type of resource, represents the standard deviation of the th type of resource, which is used to measure the dispersion degree of each node in the allocation of the th type of resource, is the total number of resource types.

[0044] : The actual allocation amount of node for resource type . This data is obtained through the resource management system. For example, the allocation of node A for CPU resources is 1800 MHz.

[0045] : The average allocation amount of all nodes for resource type . This is obtained by taking the arithmetic mean of the resource allocation data of all nodes. For example, the average allocation of all nodes for CPU resources is 1700 MHz.

[0046] : The allocation standard deviation of resource type . This is calculated through statistical methods and reflects the allocation fluctuations of all nodes for this resource type. For example, the standard deviation of CPU resources is 120 MHz.

[0047] : The total number of resource types. In this example, it is assumed that CPU and memory resources are monitored, so .

[0048] Set the actual allocations of node A for CPU and memory resources to be 1800 MHz and 8 GB respectively. The average allocation of all nodes for CPU is 1700 MHz, and for memory is 7.5 GB. The standard deviations of CPU and memory are 120 MHz and 0.5 GB respectively.

[0049] Calculate the single - item uniformity index for each resource: For CPU resources: ; For memory resources: ; Calculate the uniformity index : ; The result shows that the resource allocation uniformity index of node A is 0.9165. This value is close to 1, indicating that the resource allocation of node A is relatively uniform, without significant resource allocation deviation, which helps to identify possible problems in resource allocation and ensure that all computing nodes can effectively synchronize and utilize the allocated resources.

[0050] The above are only the preferred embodiments of the present invention, and do not limit the present invention in other forms. Any person skilled in the art may use the technical content disclosed above to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention still belong to the protection scope of the technical solution of the present invention.

Claims

1. A computing power optimization allocation system based on remote scheduling, characterized by: The system comprises: The task association analysis module collects the interaction data between computing tasks, tracks the data flow, analyzes the task interaction mode, records the data transmission time and frequency between tasks, updates the impact relationship between tasks, and generates a task dependency list; The remote scheduling optimization module arranges the order of task execution according to the task dependency list, monitors the task execution status in real time, adjusts the priority of queued tasks, responds to task completion and delay events, dynamically rearranges the task queue, and generates an optimized task execution plan; The microservice monitoring module tracks the resource usage of each microservice unit based on the optimized task execution plan, counts CPU occupancy, memory usage, and network input and output, updates the resource consumption record table according to the data, and generates a real-time resource monitoring log; The dynamic resource adjustment module compares the resource monitoring log with the judgment threshold, identifies the resource high-load area and the low-load area, and simultaneously implements the preset resource allocation to dynamically transfer resources from the low-load area to the high-load area, thereby generating an adjusted resource distribution state; The computing power monitoring module records and analyzes the efficiency of each computing node according to the adjusted resource distribution state, counts the resource consumption of the node, evaluates the adaptability of computing power usage, updates computing power usage records, and generates a computing power usage efficiency report; The resource synchronization integration module combines the computing power utilization efficiency report, adjusts the resource configuration of each computing node, performs resource configuration synchronization processing, verifies the consistency and uniformity of resource configuration, and generates a resource synchronization status record.

2. The computing power optimization allocation system based on remote scheduling according to claim 1 is characterized in that: The task dependency list includes task type, interaction frequency and data flow direction; the optimized task execution plan includes task sequence, priority setting record and scheduling response time; the real-time resource monitoring log includes CPU usage record, memory usage record and network IO data; the adjusted resource distribution status includes resource load diagram, high and low load areas and resource flow path; the computing power utilization efficiency report includes resource consumption data, efficiency score and node adaptability analysis results; the resource synchronization status record includes configuration consistency report, resource uniformity analysis record and synchronization processing log.

3. The computing power optimization allocation system based on remote scheduling according to claim 1 is characterized in that: The task association analysis module includes a data collection submodule, an interaction analysis submodule, and a dependency list submodule; The data collection submodule uses timestamp recording and event log collection based on the interaction data between tasks to mark and continuously track data flows, and confirms data integrity through network traffic analysis to generate interaction data sets; The interaction analysis submodule analyzes the timestamp and frequency of data transmission through the interaction data set, compares the data packet size and transmission rate, identifies key interaction patterns, and obtains interaction pattern analysis results; The dependency list submodule gradually updates and optimizes the interdependencies between tasks based on the interaction mode analysis results, constructs a task dependency graph through dependency path analysis, and generates a task dependency list.

4. The computing power optimization allocation system based on remote scheduling according to claim 1 is characterized in that: The remote scheduling optimization module includes a task scheduling submodule, a status monitoring submodule, and a dynamic adjustment submodule; The task scheduling submodule sorts the tasks based on the task dependency list through dependency analysis and priority evaluation, and allocates the optimal start time to each task in combination with resource availability to generate a scheduled execution order; The status monitoring submodule collects the execution data of each task in real time through sensors based on the scheduled execution order, analyzes the data and monitors the task progress in real time, identifies and discovers potential delays, and generates task progress monitoring results; The dynamic adjustment submodule analyzes the current execution status and the remaining task queue based on the task progress monitoring results, adjusts the priority and scheduling strategy of future tasks, matches real-time changes, and generates an optimized task execution plan.

5. The computing power optimization allocation system based on remote scheduling according to claim 1 is characterized in that: The microservice monitoring module includes a resource usage tracking submodule, a resource record update submodule, and a monitoring log generation submodule; The resource usage tracking submodule extracts the task allocation information of the microservice unit based on the optimized task execution plan, parses the resource call parameters of each microservice unit, analyzes the CPU occupancy periodicity, counts the cumulative memory consumption, retrieves the network input and output data and performs upload and download traffic, and obtains resource usage statistics; The resource record update submodule extracts the single value of resource usage based on the resource usage statistical data, separates the data fields of different resource categories and classifies the values, matches the field structure of the existing resource consumption record table and supplements the update data to establish resource consumption update data; The real-time monitoring log generation submodule reads the updated resource record content based on the resource consumption update data, extracts the dynamic change values ​​of CPU, memory and network traffic, arranges and formats the time series, and generates a real-time resource monitoring log.

6. The computing power optimization allocation system based on remote scheduling according to claim 1 is characterized in that: The dynamic resource adjustment module includes a resource load identification submodule, a resource dynamic transfer submodule, and a resource distribution status generation submodule; The resource load identification submodule analyzes the load indicators of resource nodes based on the real-time resource monitoring log, performs comparative calculation of resource load thresholds, marks resource node areas exceeding the high load threshold and below the low load threshold, and generates a resource load status identifier; The resource dynamic transfer submodule is based on the resource load state identifier, matches the preset resource allocation rules, extracts the number of available resources in the low-load area and records the corresponding nodes, identifies the resource gap in the high-load area and compares the resource data in the low-load area, performs resource transfer allocation operations according to resource requirements and allocation rules, and generates resource transfer execution data; The resource distribution status generating submodule counts the current resource allocation of each node based on the resource transfer execution data, merges the node allocation status and adjusts the relevant records of the resource distribution table to generate the adjusted resource distribution status.

7. The computing power optimization allocation system based on remote scheduling according to claim 1 is characterized in that: The computing power monitoring module includes a computing power efficiency recording submodule, a computing power adaptation evaluation submodule, and a computing power usage report generation submodule; The computing power efficiency recording submodule analyzes the node task completion time and resource consumption based on the adjusted resource distribution state, calculates the task execution efficiency of each node, performs a comparative analysis of task efficiency and resource consumption, and generates node efficiency recording data; The computing power adaptation evaluation submodule extracts the matching value of task efficiency and resource consumption based on the node efficiency record data, compares the difference between node computing power allocation and task requirements, and evaluates the computing power usage adaptation of the different nodes item by item to generate a computing power adaptation evaluation result; The computing power usage report generation submodule summarizes the adaptation data of the nodes based on the computing power adaptation evaluation results, counts the usage of the overall computing power distribution, integrates the computing power efficiency and adaptation evaluation content into a structured text, and generates a computing power usage efficiency report.

8. The computing power optimization allocation system based on remote scheduling according to claim 7 is characterized in that: The task execution efficiency of each node is calculated according to the formula: ； Calculate, where represents the task execution efficiency of the i-th node, represents the resource consumption of the i-th node, Representative The time required for a node to complete the task, Representative The total amount of resources allocated to each node, Representative The resource requirements of the nodes are different.

9. The computing power optimization allocation system based on remote scheduling according to claim 1 is characterized in that: The resource synchronization integration module includes a resource configuration adjustment submodule, a resource synchronization processing submodule, and a resource status inspection submodule; The resource configuration adjustment submodule extracts the resource allocation data of the node based on the computing power utilization efficiency report, counts the redundant resources of the low resource utilization node and the resource demand of the high resource utilization node, matches the resource allocation demand of the node and adjusts the resource configuration to generate adjusted resource configuration data; The resource synchronization processing submodule extracts the allocation parameters of the node resources based on the adjusted resource configuration data and matches them item by item, calls the synchronization mechanism to confirm that the resource allocation operation is consistent, records the node resource status after the allocation is completed, and generates a resource synchronization configuration record; The resource status verification submodule extracts the resource allocation status of each node based on the resource synchronization configuration record, calculates the uniformity index of resource allocation of each node, analyzes and marks the abnormal allocation status in the node resource distribution, verifies the completion of resource synchronization of all nodes, and generates a resource synchronization status record.

10. The computing power optimization allocation system based on remote scheduling according to claim 9 is characterized in that: The uniformity index of resource allocation is according to the formula: ； Calculate, where Representative The resource allocation uniformity index of each node, Representative The node in The amount of resources allocated, Represents all nodes in The average distribution of resources, Representative The standard deviation of the resource, is the total number of resource types.

Images