Intelligent scheduling method and system for distributed computing power scheduling strategy optimization

Abstract

The invention discloses an intelligent scheduling method and system for distributed computing power scheduling strategy optimization, and relates to the technical field of intelligent scheduling. The method comprises the following steps: self-checking distributed computing power nodes; carrying out distributed computing power first evaluation analysis on the distributed computing power nodes; performing first intelligent scheduling according to the first comparative analysis result; carrying out distributed computing power second evaluation analysis on the distributed computing power nodes; performing second intelligent scheduling according to the second comparative analysis result; performing third evaluation analysis on the distributed computing power of the distributed computing power nodes; and performing third intelligent scheduling according to the third comparative analysis result. Through self-inspection, first evaluation analysis, second evaluation analysis and third evaluation analysis and corresponding hierarchical optimization scheduling, the effect of improving the accuracy of distributed computing power resource scheduling is achieved, and the problem that in the prior art, the accuracy of distributed computing power resource scheduling corresponding to distributed power resources is insufficient is solved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent scheduling technology, and in particular to an intelligent scheduling method and system for optimizing distributed computing power scheduling strategies. Background Art

[0002] With the vigorous development of the digital economy, data has become the new core means of production, computing power has become the key productive force, and the construction of a new generation of information infrastructure, artificial intelligence, industrial Internet, Internet of Things, etc., has led to a sharp increase in demand for computing resources; at the same time, the rise of generative artificial intelligence and the need for big data processing have promoted the development of distributed computing systems. These technologies require efficient computing power support to ensure the efficient operation of algorithms.

[0003] The existing intelligent scheduling methods for optimizing distributed computing power scheduling strategies are implemented through the following technologies, including: a multi-level computing power network system framework, which proposes a multi-level computing power network system framework to address challenges such as wide distribution heterogeneity, strong uncertainty, and high constraint complexity; a computing power network resource collaborative scheduling platform to form a computing power resource scheduling strategy between multiple public clouds and private clouds to achieve automatic optimal allocation of resources; cloud-edge collaborative scheduling technology to integrate multi-level ubiquitous computing network resources of the cloud, edge, and end through a computing power network; and cloud-native ubiquitous computing power scheduling technology. With the advancement of cloud computing technology, cloud-native ubiquitous computing power scheduling technology has become the key for enterprises to achieve cost reduction and efficiency improvement by moving to the cloud.

[0004] For example, the invention patent with publication number: CN117032968A discloses an intelligent scheduling method based on a distributed architecture, including: evaluating and analyzing the scale complexity of the distributed architecture system and the load status of each computing node, and expanding or reducing the computing nodes through intelligent decision-making, to achieve elastic expansion and structural optimization of the distributed architecture system, and to monitor the load status of each computing node in real time, and thereby intelligently match the main line path and multiple alternative paths for task processing, and allocate tasks or resources to different computing nodes to fully utilize the computing power and storage resource capabilities of each node.

[0005] For example, the invention patent with announcement number: CN117453398B announces an intelligent optimization method and system for computing power scheduling to improve power supply reliability, including: obtaining distributed configuration data of the cloud computing system, performing digital twin modeling to obtain a cloud computing twin model, deploying intelligent sensors to collect operation data, obtaining equipment operation data, including historical operation data and real-time operation data, updating the model, performing power fault prediction to obtain power fault prediction results, obtaining optimization goals, allocating computing resources, generating a resource allocation plan, allocating computing nodes to each data center, generating a computing power allocation plan, and scheduling computing power for the cloud computing system.

[0006] However, in the process of implementing the technical solution of the invention in the embodiments of the present application, the present application found that the above technology has at least the following technical problems:

[0007] In the prior art, the intelligent scheduling method for optimizing the distributed computing power scheduling strategy for distributed power resources takes many factors into consideration in a specific environment. Distributed power resources themselves have a large volatility problem, and the corresponding distributed computing power nodes cannot be simply scheduled based on a single factor. There is a problem of insufficient accuracy in the scheduling of distributed computing power resources corresponding to distributed power resources. Summary of the invention

[0008] The embodiments of the present application solve the problem of insufficient accuracy in the scheduling of distributed computing power resources corresponding to distributed power resources in the prior art by providing an intelligent scheduling method and system for optimizing distributed computing power scheduling strategies, thereby achieving the effect of improving the accuracy of distributed computing power resource scheduling.

[0009] The embodiment of the present application provides an intelligent scheduling method for optimizing a distributed computing power scheduling strategy, comprising the following steps: self-checking distributed computing power nodes, and performing distributed computing power self-checking adjustment according to the self-checking of the distributed computing power nodes;

[0010] Perform a first evaluation and analysis on the distributed computing power nodes to obtain an evaluation value of the schedulable resources of the distributed computing power nodes; perform a first comparative analysis on the evaluation value of the schedulable resources of the distributed computing power nodes and the corresponding thresholds, and perform a first intelligent scheduling according to the first comparative analysis results; perform a second evaluation and analysis on the distributed computing power nodes to obtain a risk assessment value of the distributed computing power node scheduling computing power resources; perform a second comparative analysis on the distributed computing power risk assessment value of the distributed computing power node scheduling computing power resources and the corresponding thresholds, and perform a second intelligent scheduling according to the second comparative analysis results; perform a third evaluation and analysis on the distributed computing power nodes to obtain a comprehensive evaluation value of the distributed computing power node scheduling computing power resource call; perform a third comparative analysis on the comprehensive evaluation value of the distributed computing power node scheduling computing power resource call and the corresponding thresholds, and perform a third intelligent scheduling according to the third comparative analysis results.

[0011] Furthermore, the specific process of performing distributed computing power self-inspection adjustment according to the distributed computing power node self-inspection is: the distributed computing power node performs a predefined self-inspection scheme at a predefined self-inspection time, and the predefined self-inspection scheme is used for self-inspection of the computing power resources of the distributed computing power node to obtain the self-inspection parameters of the distributed computing power node; if the self-inspection parameter of the distributed computing power node is greater than or equal to the self-inspection parameter threshold of the distributed computing power node, the corresponding distributed computing power node is recorded as a faulty node, and a notification of the faulty node is sent to relevant personnel; if the self-inspection parameter of the distributed computing power node is less than the self-inspection parameter threshold of the distributed computing power node, the corresponding distributed computing power node is recorded as a faulty node to be detected and evaluated.

[0012] Furthermore, the specific process of obtaining the evaluation value of the schedulable resources of the distributed computing power node is as follows: the number of CPU callable cores, the CPU callable operating frequency and the CPU running memory callable capacity are collected through the CPU information query tool of the server of the distributed computing power node; the average CPU running failure rate and the average CPU running recovery time are collected through the fault log monitoring tool of the server of the distributed computing power node; the predefined number of CPU cores, the predefined operating frequency, the predefined CPU running memory callable capacity, the standard value of the predefined CPU running failure rate and the predefined CPU running average recovery time are extracted from the factory log of the server of the distributed computing power node. Recovery time; obtain the weight factor of the average CPU operation failure rate for the distributed computing power node schedulable resource evaluation value, the weight factor of the average CPU operation recovery time for the distributed computing power node schedulable resource evaluation value, the weight factor of the CPU callable core number for the distributed computing power node schedulable resource evaluation value, the weight factor of the CPU callable operating frequency for the distributed computing power node schedulable resource evaluation value and the weight factor of the CPU running memory callable capacity for the distributed computing power node schedulable resource evaluation value through the pre-set mapping relationship in the distributed computing power central database; thereby obtaining the distributed computing power node schedulable resource evaluation value of the distributed computing power node.

[0013] Furthermore, the first intelligent scheduling is performed according to the first comparative analysis result, specifically including: if the distributed computing power node schedulable resource evaluation value of the distributed computing power node is less than or equal to the distributed computing power node schedulable resource evaluation value, the distributed computing power node is recorded as an uncallable distributed resource computing power node, and the node fault recovery plan is enabled for the uncallable distributed resource computing power node; the uncallable distributed resource computing power node is used to describe the distributed computing power node in the uncallable distributed computing power resource when the distributed computing power central database calls the distributed computing power resource; if the distributed computing power node schedulable resource evaluation value of the distributed computing power node is greater than the distributed computing power node schedulable resource evaluation value, the distributed computing power node is recorded as a callable distributed resource computing power node; the callable distributed resource computing power node is used to describe the distributed computing power node in the callable distributed computing power resource when the distributed computing power central database calls the distributed computing power resource.

[0014] Furthermore, the specific process of obtaining the distributed computing power node scheduling computing power resource risk assessment value is as follows: directly obtaining the distributed computing power node power task volume and the distributed computing power node renewable energy task volume through the computing task scheduling tool of the distributed computing power node; directly extracting the distributed computing power power demand forecast correction factor, the light intensity correction factor for the distributed computing power node renewable energy task volume, the distributed computing power node historical power task average volume and the distributed computing power node historical renewable energy task average volume through the distributed computing power central database; thereby analyzing and obtaining the distributed computing power node scheduling computing power resource risk assessment value of the callable distributed resource computing power node.

[0015] Furthermore, the second intelligent scheduling is performed according to the second comparative analysis result, specifically including: if the distributed computing power node scheduling computing power resource risk assessment value of the callable distributed resource computing power node is greater than or equal to the distributed computing power node scheduling computing power resource risk assessment threshold, the corresponding callable distributed resource computing power node is recorded as an uncallable distributed resource computing power node, and a distributed resource computing power node is added to the power resource area corresponding to the uncallable distributed resource computing power node; if the distributed computing power node scheduling computing power resource risk assessment value of the callable distributed resource computing power node is less than the distributed computing power node scheduling computing power resource risk assessment threshold, the corresponding callable distributed resource computing power node is recorded as a first-level callable distributed resource computing power node.

[0016] Furthermore, the distributed computing power node scheduling computing power resource call comprehensive evaluation value obtained specifically includes: performing a first evaluation analysis and a second evaluation analysis on the first-level distributed resource computing power node that can be called, to obtain the distributed computing power node scheduling resource evaluation value of the first-level distributed resource computing power node and the distributed computing power node scheduling computing power resource risk assessment value of the first-level distributed resource computing power node that can be called; obtaining the distributed computing power node network delay through network monitoring equipment; obtaining the server operating temperature and server ambient temperature through server temperature monitoring equipment; thereby analyzing and obtaining the distributed computing power node scheduling computing power resource call comprehensive evaluation value of the first-level distributed resource computing power node that can be called.

[0017] Furthermore, the third intelligent scheduling is performed according to the third comparative analysis result, specifically including: if the distributed computing power node scheduling computing power resource call comprehensive evaluation value of the first-level distributed resource computing power node that can be called is greater than or equal to the distributed computing power node scheduling computing power resource call comprehensive evaluation threshold, then the first-level distributed resource computing power node is included in the list of callable distributed resource computing power nodes, and the distributed resource computing power nodes in the list of callable distributed resource computing power nodes are scheduled for computing power resources according to the predefined node resource scheduling scheme; if the distributed computing power node scheduling computing power resource call comprehensive evaluation value of the first-level distributed resource computing power node that can be called is less than the distributed computing power node scheduling computing power resource call comprehensive evaluation threshold, then the first-level distributed resource computing power node is recorded as a second-level distributed resource computing power node, and the node server corresponding to the second-level distributed resource computing power node is virtualized. Technology, data compression technology and upgraded hardware capacity are used.

[0018] Furthermore, the third intelligent scheduling based on the third comparative analysis result also includes: after the predefined intelligent scheduling distributed computing power detection time, performing a third evaluation analysis on the secondary-callable distributed resource computing power node to obtain the distributed computing power node scheduling computing power resource call comprehensive evaluation value of the secondary-callable distributed resource computing power node; if the distributed computing power node scheduling computing power resource call comprehensive evaluation value of the secondary-callable distributed resource computing power node is greater than or equal to the distributed computing power node scheduling computing power resource call comprehensive evaluation threshold, no adjustment is made; if the distributed computing power node scheduling computing power resource call comprehensive evaluation value of the secondary-callable distributed resource computing power node is less than the distributed computing power node scheduling computing power resource call comprehensive evaluation threshold, then notifying relevant personnel to add distributed computing power nodes in the corresponding area.

[0019] In this embodiment, if after the predefined intelligent scheduling distributed computing power detection time, if the distributed computing power node scheduling computing power resource call comprehensive evaluation value of the secondary distributed resource computing power node that can be called, is less than the distributed computing power node scheduling computing power resource call comprehensive evaluation threshold, then it means that the computing power resources in this area are not sufficient as a whole, and it is necessary to notify relevant personnel to add more nodes to the distributed system to disperse the load and improve the overall computing power.

[0020] The embodiment of the present application provides an intelligent scheduling system for optimizing distributed computing power scheduling strategies, including a distributed computing power node self-check module, a distributed computing power first evaluation and analysis module, a distributed computing power first intelligent scheduling module, a distributed computing power second evaluation and analysis module, a distributed computing power second intelligent scheduling module, a distributed computing power third evaluation and analysis module and a distributed computing power third intelligent scheduling module: the distributed computing power node self-check module: used for distributed computing power node self-check, and performing distributed computing power self-check adjustment according to the distributed computing power node self-check; the distributed computing power first evaluation and analysis module: used for performing a first evaluation and analysis of the distributed computing power on the distributed computing power node, and obtaining an evaluation value of the schedulable resources of the distributed computing power node; the distributed computing power first intelligent scheduling module: used for performing a first comparative analysis of the schedulable resource evaluation value of the distributed computing power node with the corresponding threshold value, based on The first intelligent scheduling is performed according to the first comparative analysis result; the distributed computing power second evaluation and analysis module is used to perform a second evaluation and analysis on the distributed computing power nodes to obtain the risk assessment value of the distributed computing power node scheduling computing power resources; the distributed computing power second intelligent scheduling module is used to perform a second comparative analysis on the risk assessment value of the distributed computing power node scheduling computing power resources with the corresponding threshold, and perform a second intelligent scheduling according to the second comparative analysis result; the distributed computing power third evaluation and analysis module is used to perform a third evaluation and analysis on the distributed computing power nodes to obtain the comprehensive evaluation value of the distributed computing power node scheduling computing power resource call; the distributed computing power third intelligent scheduling module is used to perform a third comparative analysis on the comprehensive evaluation value of the distributed computing power node scheduling computing power resource call with the corresponding threshold, and perform a third intelligent scheduling according to the third comparative analysis result.

[0021] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0022] 1. Perform a first evaluation and analysis on the distributed computing power of the distributed computing power nodes; perform a first intelligent scheduling based on the first comparative analysis results; perform a second evaluation and analysis on the distributed computing power of the distributed computing power nodes; perform a second intelligent scheduling based on the second comparative analysis results; perform a third evaluation and analysis on the distributed computing power of the distributed computing power nodes; perform a third intelligent scheduling based on the third comparative analysis results, thereby achieving optimized scheduling of computing power resources of hierarchical distributed nodes, thereby achieving the effect of improving the accuracy of distributed computing power resource scheduling, and solving the problem of insufficient accuracy in the scheduling of distributed computing power resources corresponding to distributed power resources in the prior art.

[0023] 2. Perform a second comparative analysis on the risk assessment value of the distributed computing power node scheduling computing power resources and the corresponding threshold value, and perform a second intelligent scheduling based on the second comparative analysis results. By identifying the uncallable distributed resource computing power nodes and adding distributed resource computing power nodes, the overall computing power resource utilization rate is improved. By adding distributed resource computing power nodes, the redundancy of the system is increased, thereby improving the reliability of the distributed computing power scheduling method.

[0024] 3. Perform a third comparative analysis on the comprehensive evaluation value of the distributed computing power node scheduling computing power resource call and the corresponding threshold, perform a third intelligent scheduling based on the third comparative analysis results, dynamically manage the distributed resource computing power nodes, adjust resource allocation according to actual needs, thereby identifying potential risk points, and taking corresponding measures, thereby improving the security of the distributed computing power scheduling method.

[0025] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 A flow chart of an intelligent scheduling method for optimizing distributed computing power scheduling strategy provided in an embodiment of the present application;

[0027] Figure 2 A schematic diagram of the structure of an intelligent scheduling system for optimizing distributed computing power scheduling strategies provided in an embodiment of the present application. DETAILED DESCRIPTION

[0028] The embodiments of the present application solve the problem of insufficient accuracy in the scheduling of distributed computing resources corresponding to distributed power resources in the prior art by providing an intelligent scheduling method and system for optimizing distributed computing power scheduling strategies. The accuracy of distributed computing power resource scheduling is improved through self-checking, first evaluation analysis, second evaluation analysis and third evaluation analysis, and corresponding hierarchical optimization scheduling.

[0029] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0030] like Figure 1As shown, it is a flow chart of an intelligent scheduling method for optimizing distributed computing power scheduling strategies provided by an embodiment of the present application. The method is applied to an intelligent scheduling system for optimizing distributed computing power scheduling strategies. The method includes the following steps: self-checking distributed computing power nodes, and adjusting distributed computing power self-checking according to the self-checking of distributed computing power nodes; performing a first evaluation and analysis on the distributed computing power of the distributed computing power nodes to obtain an evaluation value of schedulable resources of the distributed computing power nodes; performing a first comparative analysis on the evaluation value of the schedulable resources of the distributed computing power nodes and the corresponding thresholds, and performing a first intelligent scheduling according to the first comparative analysis result; performing a second evaluation and analysis on the distributed computing power of the distributed computing power nodes to obtain a risk evaluation value of distributed computing power node scheduling computing power resources; performing a second comparative analysis on the risk evaluation value of distributed computing power node scheduling computing power resources and the corresponding thresholds, and performing a second intelligent scheduling according to the second comparative analysis result; performing a third evaluation and analysis on the distributed computing power of the distributed computing power nodes to obtain a comprehensive evaluation value of distributed computing power node scheduling computing power resource calls; performing a third comparative analysis on the comprehensive evaluation value of distributed computing power node scheduling computing power resource calls and the corresponding thresholds, and performing a third intelligent scheduling according to the third comparative analysis result.

[0031] Furthermore, the specific process of performing distributed computing power self-inspection adjustment according to the distributed computing power node self-inspection is as follows: the distributed computing power node performs a predefined self-inspection scheme at a predefined self-inspection time, and the predefined self-inspection scheme is used for self-inspection of the computing power resources of the distributed computing power node to obtain the self-inspection parameters of the distributed computing power node; if the self-inspection parameter of the distributed computing power node is greater than or equal to the self-inspection parameter threshold of the distributed computing power node, the corresponding distributed computing power node is recorded as a faulty node, and a notification of the faulty node is sent to relevant personnel; if the self-inspection parameter of the distributed computing power node is less than the self-inspection parameter threshold of the distributed computing power node, the corresponding distributed computing power node is recorded as a faulty node to be detected and evaluated.

[0032] In this embodiment, the predefined self-test scheme is collected and stored locally at the node or sent to the central computing power scheduling center. The self-test parameter threshold is directly extracted from the distributed computing power central database. The predefined self-test scheme of the distributed computing power node can be automatically triggered at predefined self-test times, or it can be manually triggered after specific events (such as system startup, configuration changes, fault recovery, etc.); the node executes the self-test program and collects hardware status information. Self-test parameters include CPU temperature, memory usage, hard disk usage time, software version information, system configuration information, and network connection status.

[0033] Furthermore, the specific process of obtaining the evaluation value of the schedulable resources of the distributed computing power node is as follows: the number of CPU callable cores, the CPU callable operating frequency, and the CPU running memory callable capacity are collected through the CPU information query tool of the server of the distributed computing power node; the average CPU running failure rate and the average CPU running recovery time are collected through the fault log monitoring tool of the server of the distributed computing power node; the predefined number of CPU cores, the predefined operating frequency, the predefined CPU running memory callable capacity, the standard value of the predefined CPU running failure rate, and the predefined CPU running average recovery time are extracted from the factory log of the server of the distributed computing power node. Complex time; obtain the weight factor of the average CPU operation failure rate for the distributed computing power node schedulable resource evaluation value, the weight factor of the average CPU operation recovery time for the distributed computing power node schedulable resource evaluation value, the weight factor of the CPU callable core number for the distributed computing power node schedulable resource evaluation value, the weight factor of the CPU callable operation frequency for the distributed computing power node schedulable resource evaluation value and the weight factor of the CPU operation memory callable capacity for the distributed computing power node schedulable resource evaluation value through the pre-set mapping relationship in the distributed computing power central database; thereby obtaining the distributed computing power node schedulable resource evaluation value of the distributed computing power node.

[0034] In this embodiment, the CPU information query tool, for example, l scpu, cat, proc, and pu info. The fault log monitoring tool, for example, Nagi os and Zabbix.

[0035] The distributed computing nodes of the distributed computing scheduling system are numbered, JD0 represents the number of the distributed computing node, and JD represents the total number of distributed computing nodes.

[0036] The distributed computing power node detection time is divided into different distributed computing power node detection time periods according to the size of the predefined computing power node detection time window, and the distributed computing power node detection time periods are numbered, JC0 represents the number of the distributed computing power node detection time period, and JC represents the total number of the distributed computing power node detection time periods.

[0037] The schedulable resource evaluation value of a distributed computing node is used to describe the relative basic level of software computing resources that can be called by a distributed computing node.

[0038] Represents the evaluation value of the schedulable resources of the distributed computing power node in the JC0th distributed computing power node detection time period of the JD0th distributed computing power node.

[0039]

[0040] Indicates the number of CPU callable cores in the detection time period of the JC0th distributed computing node of the JD0th distributed computing node.

[0041] Indicates the number of predefined CPU cores in the detection time period of the JC0th distributed computing node of the JD0th distributed computing node.

[0042] Indicates the CPU callable operating frequency of the JC0th distributed computing node of the JD0th distributed computing node in the detection time period.

[0043] Indicates the CPU predefined operating frequency of the JC0th distributed computing node in the detection time period of the JD0th distributed computing node.

[0044] Indicates the CPU running memory available capacity of the JC0th distributed computing node of the JD0th distributed computing node during the detection time period.

[0045] Indicates the CPU predefined running memory callable capacity of the JC0th distributed computing node in the detection time period of the JD0th distributed computing node.

[0046] Indicates the average CPU failure rate of the JD0th distributed computing node.

[0047] Indicates the predefined failure rate standard value of the CPU running of the JD0th distributed computing power node. Indicates the average CPU operation recovery time of the JD0th distributed computing node.

[0048] Indicates the predefined average recovery time of the CPU operation of the JD0th distributed computing power node.

[0049] σ 1 Indicates the weight factor of the number of CPU callable cores for the evaluation value of the schedulable resources of the distributed computing power node.

[0050] σ 2 Indicates the weight factor of the CPU callable running frequency for the distributed computing power node schedulable resource evaluation value.

[0051] σ 3 Indicates the weight factor of the CPU running memory available capacity for the distributed computing power node schedulable resource evaluation value.

[0052] The weight factor of the number of CPU callable cores for the evaluation value of the schedulable resources of the distributed computing power node, the weight factor of the CPU callable operating frequency for the evaluation value of the schedulable resources of the distributed computing power node, and the weight factor of the CPU running memory callable capacity for the evaluation value of the schedulable resources of the distributed computing power node respectively represent the numerical values ​​of the influence of the number of CPU callable cores, the CPU callable operating frequency, and the CPU running memory callable capacity on the evaluation value of the schedulable resources of the distributed computing power node.

[0053] The weight factor of the number of CPU callable cores for the distributed computing node schedulable resource evaluation value, the weight factor of the CPU callable running frequency for the distributed computing node schedulable resource evaluation value, and the weight factor of the CPU running memory callable capacity for the distributed computing node schedulable resource evaluation value can be directly obtained from the distributed computing power central database through a pre-set mapping relationship. For example, a mapping set of the real-time CPU utilization of the distributed computing power node and its corresponding weight factor of the number of CPU callable cores for the distributed computing node schedulable resource evaluation value, the weight factor of the CPU callable running frequency for the distributed computing node schedulable resource evaluation value, and the weight factor of the CPU running memory callable capacity for the distributed computing node schedulable resource evaluation value is constructed, and the real-time CPU utilization of the distributed computing power node is input into the mapping set to obtain the weight factor of the number of CPU callable cores for the distributed computing node schedulable resource evaluation value, the weight factor of the CPU callable running frequency for the distributed computing node schedulable resource evaluation value, and the weight factor of the CPU running memory callable capacity for the distributed computing node schedulable resource evaluation value, wherein the mapping relationship can be one-to-one or many-to-one.

[0054] ψ 1 Indicates the weight factor of the average CPU operation failure rate for the evaluation value of the schedulable resources of the distributed computing power node.

[0055] ψ 2 Indicates the weight factor of the average CPU operation recovery time for the evaluation value of the schedulable resources of the distributed computing power node.

[0056] The weight factor of the average CPU operation failure rate on the evaluation value of the schedulable resources of the distributed computing power nodes and the weight factor of the average CPU operation recovery time on the evaluation value of the schedulable resources of the distributed computing power nodes respectively represent the numerical values ​​of the influence of the average CPU operation failure rate and the average CPU operation recovery time on the evaluation value of the schedulable resources of the distributed computing power nodes.

[0057] The weight factor of the average CPU operation failure rate for the distributed computing node schedulable resource evaluation value and the weight factor of the average CPU operation recovery time for the distributed computing node schedulable resource evaluation value can be directly obtained from the distributed computing power central database through a pre-set mapping relationship. For example, a mapping set of the real-time distributed computing node computing resource call frequency and its corresponding average CPU operation failure rate for the distributed computing node schedulable resource evaluation value and the average CPU operation recovery time for the distributed computing node schedulable resource evaluation value is constructed, and the real-time distributed computing node computing resource call frequency is input into the mapping set to obtain the weight factor of the average CPU operation failure rate for the distributed computing node schedulable resource evaluation value and the weight factor of the average CPU operation recovery time for the distributed computing node schedulable resource evaluation value, wherein the mapping relationship can be one-to-one or many-to-one.

[0058] Furthermore, a first intelligent scheduling is performed based on the first comparative analysis result, specifically including: if the distributed computing power node schedulable resource evaluation value of the distributed computing power node is less than or equal to the distributed computing power node schedulable resource evaluation value, the distributed computing power node is recorded as an uncallable distributed resource computing power node, and a node fault recovery plan is enabled for the uncallable distributed resource computing power node; the uncallable distributed resource computing power node is used to describe the distributed computing power node in the uncallable distributed computing power resource when the distributed computing power central database calls the distributed computing power resource; if the distributed computing power node schedulable resource evaluation value of the distributed computing power node is greater than the distributed computing power node schedulable resource evaluation value, the distributed computing power node is recorded as a callable distributed resource computing power node; the callable distributed resource computing power node is used to describe the distributed computing power node in the callable distributed computing power resource when the distributed computing power central database calls the distributed computing power resource.

[0059] In this embodiment, node fault recovery solutions, such as automatic restart and failover, can improve the availability and stability of distributed computing nodes.

[0060] Automatic restart: If the distributed computing node schedulable resource evaluation value of the distributed computing node is less than or equal to the distributed computing node schedulable resource evaluation value, it may be due to hardware failure or other problems. Set the judgment condition for automatic restart. For example, if the distributed computing node schedulable resource evaluation value of the distributed computing node is less than or equal to the distributed computing node schedulable resource evaluation value and the CPU utilization is less than the predefined CPU utilization threshold, the distributed computing node will automatically restart;

[0061] Failover: If the distributed computing node's schedulable resource assessment value is less than or equal to the distributed computing node's schedulable resource assessment value, it may be due to hardware failure or other problems. The intelligent scheduling system will trigger the failover mechanism and transfer the affected tasks to other nodes to ensure task continuity and system stability.

[0062] If the distributed computing node's schedulable resource assessment value is less than or equal to the distributed computing node's schedulable resource assessment value, it means that the node's resources are already relatively tight and there is no computing power margin to unify the computing power resource calls.

[0063] Furthermore, the specific process of obtaining the risk assessment value of distributed computing power node scheduling computing power resources is as follows: directly obtain the distributed computing power node power task volume and the distributed computing power node renewable energy task volume through the computing task scheduling tool of the distributed computing power node; directly extract the distributed computing power power demand forecast correction factor, the light intensity correction factor for the distributed computing power node renewable energy task volume, the distributed computing power node historical power task average volume and the distributed computing power node historical renewable energy task average volume through the distributed computing power central database; thereby analyzing and obtaining the distributed computing power node scheduling computing power resource risk assessment value of the callable distributed resource computing power node.

[0064] In this embodiment, the distributed computing power node scheduling computing power resource risk assessment value is used to describe the relative risk level of the software computing power resources called by the distributed computing power node.

[0065] Represents the distributed computing power node scheduling computing power resource risk assessment value of the JC0th distributed computing power node detection time period of the JD0th distributed computing power node.

[0066]

[0067] e represents a natural constant;

[0068] Represents the power task volume of the distributed computing power node in the JC0th distributed computing power node detection time period of the JD0th distributed computing power node.

[0069] Represents the average historical power task of the JD0th distributed computing node.

[0070] The power task volume of distributed computing nodes refers to the number of computing tasks that need to be processed in the power system. Changes in the task volume will affect the demand for computing resources and thus affect the scheduling strategy.

[0071] It represents the maximum value of the renewable energy task volume of the distributed computing power node in the detection time period of the JC0th distributed computing power node of the JD0th distributed computing power node.

[0072] It represents the minimum value of the renewable energy task amount of the distributed computing power node in the detection time period of the JC0th distributed computing power node of the JD0th distributed computing power node.

[0073] Represents the historical average value of renewable energy tasks of the distributed computing power node JD0th distributed computing power node.

[0074] The amount of renewable energy tasks, such as the output of photovoltaic and wind power, is affected by environmental factors and has a large volatility. This volatility of power resources will further affect the stability of distributed node computing resource calls.

[0075] The distributed computing power power demand forecast correction factor representing the detection time period of the JC0th distributed computing power node of the JD0th distributed computing power node, with a value range of (1, 2);

[0076] The distributed computing power power demand forecast correction factor represents the numerical value of the impact of power demand forecast fluctuations on the increased power task volume on the risk assessment value of distributed computing power node scheduling computing power resources.

[0077] The power demand forecast fluctuations and increased power tasks can be directly extracted from the distributed computing power central database.

[0078] The distributed computing power power demand forecast correction factor can be directly obtained from the distributed computing power central database through a pre-set mapping relationship. For example, a mapping set of real-time power demand forecast fluctuations, increased power task volume and its corresponding distributed computing power power demand forecast correction factor is constructed, and the real-time power demand forecast fluctuations, increased power task volume are input into the mapping set to obtain the distributed computing power power demand forecast correction factor, wherein the mapping relationship can be one-to-one or many-to-one.

[0079] It represents the correction factor of the light intensity of the JC0th distributed computing node in the detection time period of the JD0th distributed computing node for the renewable energy task amount of the distributed computing node, and the value range is (0, 2);

[0080] The correction factor of light intensity on the renewable energy task volume of distributed computing power nodes represents the numerical value of the influence of light intensity on the renewable energy task volume of distributed computing power nodes.

[0081] The light intensity is directly collected and extracted from the light intensity measuring instrument.

[0082] The correction factor of light intensity for the renewable energy task amount of distributed computing power nodes can be directly obtained from the distributed computing power central database through a pre-set mapping relationship. For example, a mapping set of real-time light intensity and its corresponding correction factor of light intensity for the renewable energy task amount of distributed computing power nodes is constructed, and the real-time light intensity is input into the mapping set to obtain the correction factor of light intensity for the renewable energy task amount of distributed computing power nodes. The mapping relationship can be one-to-one or many-to-one.

[0083] In the actual distributed power resource system, the volatility of the power resources themselves is closely related to the computing tasks of the corresponding distributed nodes, and the computing tasks of the distributed nodes are related to the distributed computing resources that can be called. Therefore, if the distributed power resources in a region are themselves in a period of high volatility, even if the distributed computing resources that can be called are monitored at a certain moment, the actual call may collide with the next peak of distributed power resources volatility, resulting in the failure of the distributed node computing resources to call. Therefore, for the distributed computing nodes corresponding to the distributed power resource nodes with greater volatility, the risk of calling computing resources should be quantitatively evaluated. On the contrary, for nodes with relatively stable power resources and at the same time with a surplus of callable computing resources, the computing resources can be called with confidence.

[0084] Furthermore, a second intelligent scheduling is performed based on the second comparative analysis result, specifically including: if the distributed computing power node scheduling computing power resource risk assessment value of the callable distributed resource computing power node is greater than or equal to the distributed computing power node scheduling computing power resource risk assessment threshold, the corresponding callable distributed resource computing power node is recorded as an uncallable distributed resource computing power node, and a distributed resource computing power node is added to the power resource area corresponding to the uncallable distributed resource computing power node; if the distributed computing power node scheduling computing power resource risk assessment value of the callable distributed resource computing power node is less than the distributed computing power node scheduling computing power resource risk assessment threshold, the corresponding callable distributed resource computing power node is recorded as a first-level callable distributed resource computing power node.

[0085] In this embodiment, if the distributed computing power node scheduling computing power resource risk assessment value of the callable distributed resource computing power node is greater than or equal to the distributed computing power node scheduling computing power resource risk assessment threshold, it means that the power resource volatility of the power resource area corresponding to the corresponding callable distributed resource computing power node is large, then additional distributed resource computing power nodes should be added to the corresponding power resource area so as to reduce the callable computing power resource pressure of the distributed resource computing power nodes in the corresponding power resource area.

[0086] Furthermore, a comprehensive evaluation value of the distributed computing power node scheduling computing power resource call is obtained, specifically including: performing a first evaluation analysis and a second evaluation analysis on the first-level callable distributed resource computing power node to obtain the distributed computing power node scheduling resource evaluation value of the first-level callable distributed resource computing power node and the distributed computing power node scheduling computing power resource risk assessment value of the first-level callable distributed resource computing power node; obtaining the distributed computing power node network delay through network monitoring equipment; obtaining the server operating temperature and server ambient temperature through server temperature monitoring equipment; and obtaining the distributed computing power node scheduling computing power resource call comprehensive evaluation value of the first-level callable distributed resource computing power node through analysis.

[0087] In this embodiment, the distributed computing power node scheduling computing power resource call comprehensive evaluation value is used to describe the comprehensive level of software computing power resources that can be called by the distributed computing power node.

[0088] The distributed computing power nodes in the following text include first-level callable distributed resource computing power nodes. If the distributed computing power node scheduling computing power resource call comprehensive evaluation value of the first-level callable distributed resource computing power node is obtained, then the distributed computing power node scheduling computing power resource call comprehensive evaluation value of the corresponding distributed computing power node can be calculated. The same applies to different evaluation values ​​of other different levels of callable distributed resource computing power nodes.

[0089]

[0090] Represents the evaluation value of the schedulable resources of the distributed computing power node in the JC0th distributed computing power node detection time period of the JD0th distributed computing power node.

[0091] Represents the distributed computing power node scheduling computing power resource risk assessment value of the JC0th distributed computing power node detection time period of the JD0th distributed computing power node.

[0092] Represents the comprehensive evaluation value of the distributed computing power node scheduling computing power resource call in the JC0th distributed computing power node detection time period of the JD0th distributed computing power node.

[0093] Indicates the distributed computing node network delay of the JC0th distributed computing node detection time period of the JD0th distributed computing node.

[0094] It represents the standard value of the distributed computing node network historical delay of the JC0th distributed computing node detection time period of the JD0th distributed computing node.

[0095] Network latency affects data transmission and task execution efficiency, and is therefore also a factor that needs to be considered in scheduling.

[0096] Indicates the server operating temperature of the JC0th distributed computing node of the JD0th distributed computing node during the detection time period.

[0097] Indicates the server environment temperature of the JC0th distributed computing node of the JD0th distributed computing node during the detection time period.

[0098] If the server operating temperature exceeds the ambient temperature by too much, it means that the server is in a relatively high-load operating state. There may be distributed computing resources that can be called upon. However, if the computing resources are called upon, the operating pressure of the corresponding node is increased, which may cause damage to the server under higher load.

[0099] Indicates the correction factor of the light intensity of the JC0th distributed computing node of the JD0th distributed computing node to the ambient temperature during the detection period, with a value range of (1, 2);

[0100] The correction factor of light intensity for ambient temperature represents the numerical value of the influence of light intensity on the ambient temperature of the distributed computing power node.

[0101] The correction factor of light intensity for ambient temperature can be directly obtained from the distributed computing power central database through a pre-set mapping relationship. For example, a mapping set of real-time light intensity and its corresponding correction factor of light intensity for ambient temperature is constructed, and the real-time light intensity is input into the mapping set to obtain the correction factor of light intensity for ambient temperature. The mapping relationship can be one-to-one or many-to-one.

[0102] Furthermore, a third intelligent scheduling is performed based on the third comparative analysis result, specifically including: if the distributed computing power node scheduling computing power resource call comprehensive evaluation value of the first-level distributed resource computing power node that can be called is greater than or equal to the distributed computing power node scheduling computing power resource call comprehensive evaluation threshold, then the first-level distributed resource computing power node that can be called is included in the list of callable distributed resource computing power nodes, and the distributed resource computing power nodes in the list of callable distributed resource computing power nodes are scheduled for computing power resources according to the predefined node resource scheduling scheme; if the distributed computing power node scheduling computing power resource call comprehensive evaluation value of the first-level distributed resource computing power node that can be called is less than the distributed computing power node scheduling computing power resource call comprehensive evaluation threshold, then the first-level distributed resource computing power node that can be called is recorded as a second-level distributed resource computing power node, and virtualization technology, data compression technology and hardware capacity upgrade are used for the node server corresponding to the second-level distributed resource computing power node.

[0103] In this embodiment, the distributed computing power node scheduling computing power resource call comprehensive evaluation threshold is directly extracted from the distributed computing power central database. Virtualization technology: Use virtualization technology, such as virtual machines (VM) or containers (such as Docker), to improve resource utilization; data compression and optimization, compress and optimize data to reduce data transmission and storage requirements; cache strategy optimization: optimize cache strategy, reduce access to primary storage, and increase data processing speed. Upgrade hardware capacity, such as upgrading CPU, GPU, memory and other hardware to improve the computing power of the node; increase storage capacity: increase storage capacity to support more data and tasks; use solid-state drives (SSD): replace traditional hard disk drives (HDD) with SSD to increase data access speed; use dedicated hardware accelerators: For specific computing tasks, use dedicated hardware accelerators, such as FPGA or AS IC, to improve performance.

[0104] Furthermore, performing a third intelligent scheduling based on the third comparative analysis result also includes: after a predefined intelligent scheduling distributed computing power detection time, performing a third evaluation analysis on the secondary-callable distributed resource computing power node to obtain a distributed computing power node scheduling computing power resource call comprehensive evaluation value of the secondary-callable distributed resource computing power node; if the distributed computing power node scheduling computing power resource call comprehensive evaluation value of the secondary-callable distributed resource computing power node is greater than or equal to the distributed computing power node scheduling computing power resource call comprehensive evaluation threshold, then no adjustment is made; if the distributed computing power node scheduling computing power resource call comprehensive evaluation value of the secondary-callable distributed resource computing power node is less than the distributed computing power node scheduling computing power resource call comprehensive evaluation threshold, then notifying relevant personnel to add distributed computing power nodes in the corresponding area.

[0105] In this embodiment, if after the predefined intelligent scheduling distributed computing power detection time, if the distributed computing power node scheduling computing power resource call comprehensive evaluation value of the secondary distributed resource computing power node that can be called, is less than the distributed computing power node scheduling computing power resource call comprehensive evaluation threshold, then it means that the computing power resources in this area are not sufficient as a whole, and it is necessary to notify relevant personnel to add more nodes to the distributed system to disperse the load and improve the overall computing power.

[0106] like Figure 2As shown, it is a structural schematic diagram of an intelligent scheduling system for optimizing distributed computing power scheduling strategies provided in an embodiment of the present application. The intelligent scheduling system for optimizing distributed computing power scheduling strategies provided in an embodiment of the present application includes a distributed computing power node self-check module, a distributed computing power first evaluation and analysis module, a distributed computing power first intelligent scheduling module, a distributed computing power second evaluation and analysis module, a distributed computing power second intelligent scheduling module, a distributed computing power third evaluation and analysis module and a distributed computing power third intelligent scheduling module: Distributed computing power node self-check module: used for distributed computing power node self-check, and performing distributed computing power self-check adjustment according to the distributed computing power node self-check; Distributed computing power first evaluation and analysis module: used to perform a first evaluation and analysis of the distributed computing power nodes to obtain the evaluation value of the schedulable resources of the distributed computing power nodes; Distributed computing power first intelligent scheduling module: used to adjust the schedulable resources of the distributed computing power nodes A first comparative analysis is performed on the evaluation value and the corresponding threshold value, and a first intelligent scheduling is performed according to the result of the first comparative analysis; a second evaluation and analysis module for distributed computing power: used to perform a second evaluation and analysis on the distributed computing power nodes, and obtain a risk evaluation value of the distributed computing power node scheduling computing power resources; a second intelligent scheduling module for distributed computing power: used to perform a second comparative analysis on the risk evaluation value of the distributed computing power node scheduling computing power resources and the corresponding threshold value, and perform a second intelligent scheduling according to the result of the second comparative analysis; a third evaluation and analysis module for distributed computing power: used to perform a third evaluation and analysis on the distributed computing power nodes, and obtain a comprehensive evaluation value of the distributed computing power node scheduling computing power resource call; a third intelligent scheduling module for distributed computing power: used to perform a third comparative analysis on the comprehensive evaluation value of the distributed computing power node scheduling computing power resource call and the corresponding threshold value, and perform a third intelligent scheduling according to the result of the third comparative analysis.

[0107] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0108] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0109] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0110] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0111] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0112] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. An intelligent scheduling method for optimizing distributed computing power scheduling strategy, characterized in that: The following steps are involved: Distributed computing power node self-inspection, and distributed computing power self-inspection and adjustment based on distributed computing power node self-inspection; Performing a first evaluation and analysis on the distributed computing power nodes to obtain an evaluation value of schedulable resources of the distributed computing power nodes; Perform a first comparative analysis on the schedulable resource evaluation value of the distributed computing power node and the corresponding threshold, and perform a first intelligent scheduling according to the first comparative analysis result; Perform a second evaluation and analysis on the distributed computing power nodes to obtain a risk assessment value of the distributed computing power node scheduling computing power resources; Perform a second comparative analysis on the risk assessment value of the distributed computing power node scheduling computing power resources and the corresponding threshold, and perform a second intelligent scheduling according to the second comparative analysis result; Perform a third evaluation analysis on the distributed computing power nodes to obtain a comprehensive evaluation value of the distributed computing power node scheduling computing power resource call; A third comparative analysis is performed on the distributed computing power node scheduling computing power resource call comprehensive evaluation value and the corresponding threshold value, and a third intelligent scheduling is performed according to the third comparative analysis result.

2. The intelligent scheduling method for optimizing distributed computing power scheduling strategy according to claim 1, characterized in that: The specific process of performing distributed computing power self-checking and adjustment according to the distributed computing power node self-checking is as follows: The distributed computing power node performs a predefined self-check scheme at every predefined self-check time, and the predefined self-check scheme is used for self-checking the computing power resources of the distributed computing power node to obtain the self-check parameters of the distributed computing power node; If the self-check parameter of the distributed computing power node is greater than or equal to the self-check parameter threshold of the distributed computing power node, the corresponding distributed computing power node is recorded as a faulty node, and a notification of the faulty node is sent to relevant personnel; If the self-test parameter of the distributed computing power node is less than the self-test parameter threshold of the distributed computing power node, the corresponding distributed computing power node is recorded as a fault node to be detected and evaluated.

3. The intelligent scheduling method for optimizing distributed computing power scheduling strategy according to claim 1, characterized in that: The specific process of obtaining the evaluation value of the schedulable resources of the distributed computing power node is as follows: The CPU information query tool of the server of the distributed computing power node is used to collect the number of CPU callable cores, CPU callable operating frequency, and CPU running memory callable capacity; The average CPU operation failure rate and average CPU operation recovery time are collected through the fault log monitoring tool of the server of the distributed computing power node; The predefined number of CPU cores, the predefined operating frequency of the CPU, the available capacity of the predefined operating memory of the CPU, the standard value of the predefined failure rate of the CPU, and the predefined average recovery time of the CPU are obtained by extracting the factory logs of the servers of the distributed computing nodes; Obtain, through a pre-set mapping relationship in a distributed computing power central database, a weight factor of the average CPU operation failure rate for the distributed computing power node dispatchable resource evaluation value, a weight factor of the average CPU operation recovery time for the distributed computing power node dispatchable resource evaluation value, a weight factor of the number of CPU callable cores for the distributed computing power node dispatchable resource evaluation value, a weight factor of the CPU callable operation frequency for the distributed computing power node dispatchable resource evaluation value, and a weight factor of the CPU operation memory callable capacity for the distributed computing power node dispatchable resource evaluation value; Thus, the distributed computing power node schedulable resource evaluation value of the distributed computing power node is obtained.

4. The intelligent scheduling method for optimizing distributed computing power scheduling strategy according to claim 1, characterized in that: The performing the first intelligent scheduling according to the first comparative analysis result specifically includes: If the distributed computing node dispatchable resource evaluation value of the distributed computing node is less than or equal to the distributed computing node dispatchable resource evaluation value, the distributed computing node is recorded as an unavailable distributed resource computing node, and the node failure recovery plan is enabled for the unavailable distributed resource computing node; The non-callable distributed resource computing power node is used to describe the distributed computing power node in the non-callable distributed computing power resource when the distributed computing power central database calls the distributed computing power resource; If the distributed computing node's schedulable resource evaluation value is greater than the distributed computing node's schedulable resource evaluation value, the distributed computing node is recorded as a callable distributed resource computing node; The callable distributed resource computing power node is used to describe the distributed computing power nodes in the callable distributed computing power resources when the distributed computing power central database calls the distributed computing power resources.

5. The intelligent scheduling method for optimizing distributed computing power scheduling strategy according to claim 1, characterized in that: The specific process of obtaining the risk assessment value of the distributed computing power node scheduling computing power resources is as follows: Directly obtain the power task volume and renewable energy task volume of distributed computing power nodes through the computing task scheduling tool of distributed computing power nodes; Directly extract the distributed computing power central database to obtain the distributed computing power power demand forecast correction factor, the light intensity correction factor for the distributed computing power node renewable energy task amount, the distributed computing power node historical power task average amount and the distributed computing power node historical renewable energy task average amount; This analysis obtains the distributed computing power node scheduling computing power resource risk assessment value of the callable distributed resource computing power node.

6. The intelligent scheduling method for optimizing distributed computing power scheduling strategy according to claim 1, characterized in that: The performing the second intelligent scheduling according to the second comparative analysis result specifically includes: If the distributed computing power node scheduling computing power resource risk assessment value of the callable distributed resource computing power node is greater than or equal to the distributed computing power node scheduling computing power resource risk assessment threshold, the corresponding callable distributed resource computing power node is recorded as an uncallable distributed resource computing power node, and a distributed resource computing power node is added to the power resource area corresponding to the uncallable distributed resource computing power node; If the distributed computing power node scheduling computing power resource risk assessment value of the callable distributed resource computing power node is less than the distributed computing power node scheduling computing power resource risk assessment threshold, the corresponding callable distributed resource computing power node is recorded as a first-level callable distributed resource computing power node.

7. The intelligent scheduling method for optimizing distributed computing power scheduling strategy according to claim 1, characterized in that: The obtaining of the distributed computing power node scheduling computing power resource call comprehensive evaluation value specifically includes: Performing a first evaluation analysis and a second evaluation analysis on the first-level callable distributed resource computing power node to obtain a distributed computing power node dispatchable resource evaluation value of the first-level callable distributed resource computing power node and a distributed computing power node dispatchable computing power resource risk evaluation value of the first-level callable distributed resource computing power node; The network delay of distributed computing nodes is collected through network monitoring equipment; The server operating temperature and server ambient temperature are collected through the server temperature monitoring device; This analysis yields a comprehensive evaluation value for the distributed computing power node scheduling computing power resource call of the first-level callable distributed resource computing power node.

8. The intelligent scheduling method for optimizing distributed computing power scheduling strategy according to claim 1, characterized in that: The performing the third intelligent scheduling according to the third comparative analysis result specifically includes: If the distributed computing power node scheduling computing power resource call comprehensive evaluation value of the first-level callable distributed resource computing power node is greater than or equal to the distributed computing power node scheduling computing power resource call comprehensive evaluation threshold, the first-level callable distributed resource computing power node is included in the callable distributed resource computing power node list, and the distributed resource computing power nodes in the callable distributed resource computing power node list are scheduled for computing power resources according to the predefined node resource scheduling plan; If the comprehensive evaluation value of the distributed computing power node scheduling computing power resource call of the first-level callable distributed resource computing power node is less than the comprehensive evaluation threshold of the distributed computing power node scheduling computing power resource call, the first-level callable distributed resource computing power node will be recorded as a second-level callable distributed resource computing power node, and virtualization technology, data compression technology and upgraded hardware capacity will be used for the node server corresponding to the second-level callable distributed resource computing power node.

9. The intelligent scheduling method for optimizing distributed computing power scheduling strategy according to claim 1, characterized in that: The performing the third intelligent scheduling according to the third comparative analysis result also includes: After the predefined intelligent scheduling distributed computing power detection time, the third evaluation analysis is performed on the secondary callable distributed resource computing power node to obtain the distributed computing power node scheduling computing power resource call comprehensive evaluation value of the secondary callable distributed resource computing power node; If the comprehensive evaluation value of the distributed computing power node scheduling computing power resource call of the second-level callable distributed resource computing power node is greater than or equal to the comprehensive evaluation threshold of the distributed computing power node scheduling computing power resource call, no adjustment will be made; If the comprehensive evaluation value of the distributed computing power node scheduling computing power resource call of the second-level callable distributed resource computing power node is less than the comprehensive evaluation threshold of the distributed computing power node scheduling computing power resource call, the relevant personnel will be notified to add distributed computing power nodes in the corresponding area.

10. An intelligent scheduling system for optimizing distributed computing power scheduling strategy, characterized in that: It includes a distributed computing power node self-check module, a distributed computing power first evaluation and analysis module, a distributed computing power first intelligent scheduling module, a distributed computing power second evaluation and analysis module, a distributed computing power second intelligent scheduling module, a distributed computing power third evaluation and analysis module and a distributed computing power third intelligent scheduling module: The distributed computing power node self-check module is used for distributed computing power node self-check and performs distributed computing power self-check adjustment according to the distributed computing power node self-check; The distributed computing power first evaluation and analysis module is used to perform a first evaluation and analysis on the distributed computing power nodes to obtain the evaluation value of the schedulable resources of the distributed computing power nodes; The distributed computing power first intelligent scheduling module is used to perform a first comparative analysis on the schedulable resource evaluation value of the distributed computing power node and the corresponding threshold value, and perform a first intelligent scheduling according to the first comparative analysis result; The distributed computing power second evaluation and analysis module is used to perform a second evaluation and analysis on the distributed computing power nodes to obtain a risk assessment value of the distributed computing power node scheduling computing power resources; The distributed computing power second intelligent scheduling module is used to perform a second comparative analysis on the risk assessment value of the distributed computing power node scheduling computing power resources and the corresponding threshold value, and perform a second intelligent scheduling according to the second comparative analysis result; The distributed computing power third evaluation and analysis module is used to perform a third evaluation and analysis on the distributed computing power nodes to obtain a comprehensive evaluation value of the distributed computing power node scheduling computing power resource call; The distributed computing power third intelligent scheduling module is used to perform a third comparative analysis on the distributed computing power node scheduling computing power resource call comprehensive evaluation value and the corresponding threshold value, and perform a third intelligent scheduling according to the third comparative analysis result.

Images