A computing power scheduling method and system for ARM cluster servers

Abstract

The present invention relates to a computing power scheduling method and system for an ARM cluster server, and relates to the field of computing power scheduling, including: performing computing power scheduling optimization weight configuration to obtain energy efficiency weight and time weight; constructing a computing power scheduling processing space based on multiple edge data, optimizing multiple computing power scheduling schemes according to energy efficiency weight and time weight to obtain an optimized computing power scheduling scheme library; calculating the scheduling deviations between multiple optimized computing power scheduling schemes and historical computing power scheduling schemes in the optimized computing power scheduling scheme library, and outputting the scheme with the smallest scheduling deviation as the optimal computing power scheduling scheme. Through this application, the technical problem that the traditional computing power scheduling method cannot dynamically adapt to the current state and task requirements of the server, resulting in the lack of flexibility and accuracy of the computing power scheduling scheme, and causing uneven distribution of computing power resources can be solved; the overall resource utilization efficiency can be optimized to achieve the effect of reducing the system operation energy consumption and improving the overall task processing efficiency.

Description

Technical Field

[0001] The present invention relates to the field of computing power scheduling, and in particular to a computing power scheduling method and system for an ARM cluster server. Background Art

[0002] In an edge computing environment, computing power scheduling is a key link to ensure efficient resource utilization and timely task processing, especially in an environment using ARM cluster servers. Due to its advantages in energy efficiency and cost, more and more applications choose to deploy ARM architecture-based servers in edge computing scenarios.

[0003] However, traditional computing power scheduling methods are usually based on fixed rules and preset scheduling strategies, and cannot respond to changes in server status and fluctuations in task requirements in a timely manner. This static scheduling method can easily lead to scheduling solutions that are not suitable for actual conditions when the load changes rapidly or the task requests are diversified. On the other hand, due to the lack of real-time monitoring and intelligent analysis, traditional computing power scheduling methods are difficult to effectively balance the load between servers, causing some servers to be short of resources while other servers are idle. This imbalance not only affects the overall resource utilization efficiency, but also leads to a decline in system performance and an increase in response time. Summary of the invention

[0004] In view of the technical problem that the existing computing power scheduling method for edge computing cannot dynamically adapt to the current status of the server and task requirements, resulting in a lack of flexibility and accuracy in the computing power scheduling scheme and uneven distribution of computing power resources, the present invention provides a computing power scheduling method and system for an ARM cluster server to solve the problem.

[0005] The technical solution of the present invention to solve the above technical problems is as follows:

[0006] In a first aspect, the present invention provides a computing power scheduling method for an ARM cluster server, which is applied to an ARM cluster server, wherein the ARM cluster server includes multiple ARM servers that are communicatively connected to multiple edge nodes, and comprises: monitoring multiple edge data currently collected by the multiple edge nodes, monitoring multiple current occupancy information of the multiple ARM servers, and performing computing power scheduling optimization weight configuration to obtain energy efficiency weight and timeliness weight; constructing a computing power scheduling processing space based on the multiple edge data, and optimizing the computing power scheduling scheme for scheduling the multiple edge data to the multiple ARM servers for processing according to the energy efficiency weight and timeliness weight to obtain an optimized computing power scheduling scheme library; obtaining a historical computing power scheduling scheme for processing the edge data of the multiple edge nodes by the multiple ARM servers at the previous time point; calculating the scheduling deviations between the multiple optimized computing power scheduling schemes in the optimized computing power scheduling scheme library and the historical computing power scheduling schemes, and outputting the optimized computing power scheduling scheme with the smallest scheduling deviation as the optimal computing power scheduling scheme, and according to the optimal computing power scheduling scheme, controlling the multiple edge nodes to transmit the multiple edge data to the corresponding multiple ARM servers for processing.

[0007] In a second aspect, the present invention provides a computing power scheduling system for an ARM cluster server, the system comprising an ARM cluster server, the ARM cluster server comprising a plurality of ARM servers in communication connection with a plurality of edge nodes, and further comprising: a scheduling weight optimization module, for monitoring a plurality of edge data currently collected by the plurality of edge nodes, monitoring a plurality of current occupancy information of the plurality of ARM servers, and performing computing power scheduling optimization weight configuration to obtain energy efficiency weight and timeliness weight; a scheduling scheme optimization module, for constructing a computing power scheduling processing space based on the plurality of edge data, and scheduling the plurality of edge data to the plurality of A nodes according to the energy efficiency weight and timeliness weight. The computing power scheduling scheme processed by the RM server is optimized to obtain an optimized computing power scheduling scheme library; a historical scheduling scheme acquisition module is used to obtain the historical computing power scheduling schemes of the edge data of the multiple edge nodes processed by the multiple ARM servers at the previous time point; an optimal scheduling scheme acquisition module is used to calculate the scheduling deviations between the multiple optimized computing power scheduling schemes in the optimized computing power scheduling scheme library and the historical computing power scheduling schemes, and output the optimized computing power scheduling scheme with the smallest scheduling deviation as the optimal computing power scheduling scheme. According to the optimal computing power scheduling scheme, the multiple edge nodes are controlled to transmit the multiple edge data to the corresponding multiple ARM servers for processing.

[0008] In a third aspect, the present invention further provides an electronic device, comprising:

[0009] At least one processor; a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor so that the at least one processor can perform the steps of any one of the methods described in the first aspect above.

[0010] In a fourth aspect, a computer-readable storage medium is provided, wherein a computer program is stored on the computer-readable storage medium, and when the computer program is executed, the steps of the method described in any one of the first aspects are implemented.

[0011] The beneficial effects of the present invention are as follows: by monitoring multiple edge data currently collected by multiple edge nodes, monitoring multiple current occupancy information of multiple ARM servers, and performing computing power scheduling optimization weight configuration according to multiple occupancy information, energy efficiency weight and time weight are obtained; then based on the multiple edge data, a computing power scheduling processing space is constructed, and according to the energy efficiency weight and time weight, the computing power scheduling scheme for scheduling the multiple edge data to the multiple ARM servers for processing is optimized to obtain an optimized computing power scheduling scheme library; on the other hand, the edge data of the multiple edge nodes processed by the multiple ARM servers at the previous time point is obtained. Historical computing power scheduling plan; then calculate the scheduling deviations between multiple optimized computing power scheduling plans in the optimized computing power scheduling plan library and the historical computing power scheduling plan, and output the optimized computing power scheduling plan with the smallest scheduling deviation as the optimal computing power scheduling plan; finally, according to the optimal computing power scheduling plan, control the multiple edge nodes to transmit the multiple edge data to the corresponding multiple ARM servers for processing; the accuracy, scientificity and practicality of the computing power scheduling plan setting can be improved, the computing power load can be evenly distributed among multiple servers, and the overall resource utilization efficiency can be optimized, thereby achieving the technical effect of reducing the system operation energy consumption and improving the overall task processing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 A schematic diagram of a flow chart of a computing power scheduling method for an ARM cluster server provided by the present invention;

[0013] Figure 2 A schematic diagram of the structure of a computing power scheduling system for an ARM cluster server provided by the present invention;

[0014] Figure 3 A schematic diagram of the structure of an electronic device provided by the present invention;

[0015] Figure 4 A schematic diagram of the structure of a computer-readable storage medium provided by the present invention.

[0016] In the accompanying drawings, the components represented by the reference numerals are described as follows:

[0017] Scheduling weight optimization module 01, scheduling scheme optimization module 02, historical scheduling scheme acquisition module 03, optimal scheduling scheme acquisition module 04, electronic device 500, memory 510, processor 520, first computer program 511, computer readable storage medium 600, second computer program 611. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0019] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0020] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described as "for example" in the present invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present invention.

[0021] Embodiment 1, as Figure 1 As shown, an embodiment of the present invention provides a computing power scheduling method for an ARM cluster server, and the method is applied to an ARM cluster server, wherein the ARM cluster server includes multiple ARM servers that are communicatively connected to multiple edge nodes, specifically including:

[0022] S100: Monitor the multiple edge data currently collected by the multiple edge nodes, monitor the multiple occupancy information of the multiple ARM servers, and perform weight configuration for computing power scheduling optimization to obtain energy efficiency weight and time efficiency weight.

[0023] In one embodiment, step S100 of the present application further includes:

[0024] S110: monitoring a plurality of edge data currently collected by the plurality of edge nodes, and monitoring a plurality of current occupancy information of the plurality of ARM servers.

[0025] Specifically, the method provided by the present invention aims to optimize the resource allocation and scheduling efficiency in the edge computing environment, wherein the method is applied to a cluster composed of multiple ARM servers, and multiple ARM servers and multiple edge nodes can realize data interaction based on communication connections, that is, the edge nodes send data requests, and the ARM servers respond and process these requests according to their processing capabilities and current load conditions. The collaborative work of edge nodes and ARM cluster servers can achieve more efficient data processing and resource utilization, and meet diverse industry needs, such as smart traffic management, industrial automation and other scenarios.

[0026] First, monitor the multiple edge data currently collected by the multiple edge nodes. For example, the edge nodes can come from different devices, that is, collect the operation monitoring data of multiple devices, such as sensors, cameras or IoT devices, etc., where the size and type of each edge data are different; on the other hand, at the same time node, monitor the occupancy information of multiple ARM servers in real time, such as CPU utilization, where the occupancy information can reflect the current processing capacity and load status of each server, and provide a basis for subsequent scheduling scheme optimization analysis.

[0027] S120: Calculate an energy efficiency weight correction coefficient according to the multiple occupancy rate information.

[0028] In one embodiment, step S120 of the present application further includes:

[0029] S121: Calculate the average of the plurality of occupancy information to obtain average occupancy information; S122: Calculate the ratio of the average occupancy information to the occupancy threshold value to obtain an energy efficiency weight correction coefficient.

[0030] Specifically, first, the average of the multiple occupancy information is calculated, and the average calculation result is used as the average occupancy information. For example, the CPU utilization is used as the occupancy. Assuming that the occupancy rates of the three ARM servers are 40%, 30%, and 50% respectively, the average occupancy is (40%+30%+50%) / 3, that is, 40%; then the occupancy threshold value is obtained, where the occupancy threshold value is used to evaluate whether the current load of the system is normal. Technical personnel in this field can set it according to actual operating requirements, such as usually setting the occupancy threshold value to 50%; then calculate the ratio of the average occupancy information and the occupancy threshold value, and set the ratio of the two as the energy efficiency weight correction coefficient. For example, assuming that the average occupancy information is 40% and the occupancy threshold value is 50%, the energy efficiency weight correction coefficient is 0.8, wherein, when the energy efficiency weight correction coefficient is larger, it indicates that the current load is higher, and energy efficiency balance needs to be given priority, that is, the higher the requirement for energy efficiency balance.

[0031] S130: Using the energy efficiency weight correction coefficient, perform correction calculation on the preset energy efficiency weight to obtain the energy efficiency weight, and calculate the time efficiency weight based on the energy efficiency weight, wherein the sum of the energy efficiency weight and the time efficiency weight is 1.

[0032] Specifically, first, an initial energy efficiency weight value is set, that is, a preset energy efficiency weight, such as 0.7. This weight value reflects the degree of importance the system places on energy efficiency under normal conditions. The higher the degree of importance, the greater the preset energy efficiency weight. Then, the preset energy efficiency weight is corrected and calculated using the energy efficiency weight correction coefficient to obtain the energy efficiency weight, that is, the energy efficiency weight correction coefficient is multiplied by the preset energy efficiency weight, and the product of the two is used as the energy efficiency weight. For example, assuming that the preset energy efficiency weight is 0.7 and the energy efficiency weight correction coefficient is 0.8, the corrected energy efficiency weight is 0.56. Then, the energy efficiency weight is subtracted from 1, and the difference between the two is used as the time efficiency weight. For example, if the energy efficiency weight is 0.56, the time efficiency weight is 0.44.

[0033] By setting the energy efficiency weight correction coefficient according to the current status of multiple ARM servers and optimizing the energy efficiency weight and time efficiency weight according to the correction coefficient, the computing power scheduling solution can dynamically adapt to different workloads, further enhancing the flexibility and efficiency of the edge computing environment.

[0034] S200: Based on the multiple edge data, a computing power scheduling processing space is constructed, and according to the energy efficiency weight and the time efficiency weight, a computing power scheduling scheme for scheduling the multiple edge data to the multiple ARM servers for processing is optimized to obtain an optimized computing power scheduling scheme library.

[0035] In one embodiment, step S200 of the present application further includes:

[0036] S210: Randomly distribute the multiple edge data collected by the multiple edge nodes to the multiple ARM servers for processing, obtain multiple sample computing power scheduling plans, and construct a computing power scheduling processing space; S220: Randomly generate a first computing power scheduling plan in the computing power scheduling processing space.

[0037] Specifically, the multiple edge data collected by the multiple edge nodes are randomly assigned to the multiple ARM servers for processing, wherein each ARM processor is responsible for processing one edge data, and between different sample computing power scheduling schemes, the same edge data can be assigned to the same or different ARM servers for processing. Through the random assignment process, multiple sample computing power scheduling schemes are generated, and each sample computing power scheduling scheme includes the correspondence between different edge data and ARM servers. Finally, a computing power scheduling processing space is constructed according to multiple sample computing power scheduling schemes, wherein the computing power scheduling processing space represents all possible allocation combinations between edge data and ARM servers in the current edge computing environment. Then, any computing power scheduling scheme is randomly selected in the computing power scheduling processing space as the first computing power scheduling scheme.

[0038] S230: Calculate and obtain the first computing power scheduling fitness based on the edge data processed by the multiple ARM servers in the first computing power scheduling scheme in combination with the energy efficiency weight and the timeliness weight.

[0039] In one embodiment, step S230 of the present application further includes:

[0040] S231: According to the edge data processed by the multiple ARM servers in the first computing power scheduling scheme, multiple processing occupancy rates are obtained respectively, and multiple scheduling occupancy rates are calculated by combining the multiple occupancy rate information; S232: According to the edge nodes to which the multiple edge data belong, multiple transmission delays of the multiple edge data in the first computing power scheduling scheme to the corresponding multiple ARM servers for processing are tested and the total transmission delay is calculated; S233: According to the multiple scheduling occupancy rates, total transmission delay, energy efficiency weight and time efficiency weight, the first computing power scheduling fitness of the first computing power scheduling scheme is calculated as follows: ;in, is the computing power scheduling fitness, and are energy efficiency weight and time efficiency weight, M is the number of multiple ARM servers, is the scheduling occupancy rate of the i-th ARM server in the computing power scheduling scheme, D is the variance of the M scheduling occupancy rates, T is the total transmission delay, The preset transmission delay.

[0041] Specifically, first, a historical data query is performed based on the edge data processed by the multiple ARM servers in the first computing power scheduling scheme to obtain historical processing records of multiple ARM servers, including the performance and resource occupancy of each server when processing the same type of edge data in the past, such as CPU utilization percentage; then, multiple historical occupancy rate sets of multiple ARM servers corresponding to the multiple edge data are obtained based on the historical processing records, and the multiple historical occupancy rate sets are averaged, and the average calculation result is used as the expected occupancy rate of the server for processing the corresponding edge data, and multiple processing occupancy rates corresponding to the multiple edge data are obtained; then, for each ARM server in the first computing power scheduling scheme, the expected occupancy rate and the occupancy rate information of the current server are summed up, and the summed calculation result is used as the scheduling occupancy rate to obtain multiple scheduling occupancy rates of multiple ARM servers in the first computing power scheduling scheme.

[0042] Next, for each edge node and its corresponding ARM server, a transmission delay test mechanism is designed, which can measure the transmission delay by sending a test data packet and recording the response time. Then, according to the edge nodes to which the multiple edge data belong, a transmission delay test is performed on each edge data in the first computing power scheduling scheme, and the transmission delay from each edge node to its corresponding ARM server is recorded, and the transmission delays of all edge data are added together to obtain the total transmission delay.

[0043] Construct a computing power scheduling fitness evaluation function: ; The fitness evaluation function is used to evaluate the advantages and disadvantages of different computing power scheduling schemes. This function comprehensively considers factors such as energy efficiency, timeliness, occupancy rate and transmission delay, aiming to provide a quantitative basis for the optimization of computing power scheduling schemes; In the fitness evaluation function, The computing power scheduling fitness. The greater the fitness, the better the overall computing power scheduling solution. and are energy efficiency weight and time efficiency weight, where the energy efficiency weight reflects the emphasis of the solution on energy efficiency, and the time efficiency weight reflects the emphasis of the solution on time efficiency; M is the number of multiple ARM servers, is the scheduling occupancy rate of the i-th ARM server in the computing power scheduling scheme, where the i-th ARM server is any one of the M ARM servers; D is the variance of the M scheduling occupancy rates, where the closer the scheduling occupancy rates of each server are, the smaller the variance of the M scheduling occupancy rates is, which indicates that the energy efficiency balance of the server cluster is higher, and the scheduling fitness is higher; T is the total transmission delay, where the smaller the total transmission delay is, the higher the scheduling fitness is; The preset transmission delay is the shortest total transmission delay required for multiple edge nodes to transmit data to their closest corresponding ARM server. By constructing a scheduling fitness evaluation function, various factors can be comprehensively considered, and a quantitative indicator is provided for each computing power scheduling scheme, thereby providing a more scientific and reasonable basis for the optimization analysis of subsequent computing power scheduling schemes.

[0044] S240: Continue to optimize the computing power scheduling scheme in the computing power scheduling processing space until the optimization converges, output N sample computing power scheduling schemes with the largest computing power scheduling fitness, and build an optimized computing power scheduling scheme library, where N is an integer greater than 1.

[0045] Specifically, a second computing power scheduling scheme is then randomly selected in the computing power scheduling processing space, wherein the second computing power scheduling scheme is different from the first computing power scheduling scheme; then the scheduling fitness evaluation function is used to calculate the second computing power scheduling fitness of the second computing power scheduling scheme; further using the same method, the computing power scheduling scheme is iteratively selected and fitness evaluation is performed in the computing power scheduling processing space until a predetermined number of iterations is reached, wherein the predetermined number of iterations can be set according to actual needs, for example, the predetermined number of iterations is set to 50% or all of the number of sample computing power scheduling schemes in the computing power scheduling processing space; after reaching the predetermined number of iterations, the N sample computing power scheduling schemes with the largest computing power scheduling fitness are output, wherein N is an integer greater than 1, and the value can be taken according to demand, such as setting N to 5, that is, selecting the top 5 sample computing power scheduling schemes with the largest computing power scheduling fitness; and forming an optimized computing power scheduling scheme library based on the N sample computing power scheduling schemes.

[0046] By constructing a scheduling fitness evaluation function to optimize the computing power scheduling plan, we can balance the reduction of the occupancy rate of multiple ARM servers and the total delay of multiple edge data transmission to the ARM server, ensure the reasonable distribution of edge data among various ARM servers, and reduce the data transmission delay between edge nodes and ARM servers, thereby significantly improving the overall performance and energy efficiency of ARM cluster servers.

[0047] S300: Obtain a historical computing power scheduling plan for processing edge data of the multiple edge nodes by the multiple ARM servers at a previous time point.

[0048] In one embodiment, step S300 of the present application further includes:

[0049] S310: Obtain a historical computing power scheduling plan for allocating a plurality of historical edge data collected by the plurality of edge nodes at a previous moment to the plurality of ARM servers for processing.

[0050] Specifically, each time computing power scheduling is performed, the edge data allocation of each edge node is recorded in real time, including the specific data allocated to each ARM server, and a timestamp is added to each scheduling record so that the scheduling information of the previous moment can be accurately obtained during subsequent queries, and these historical scheduling information are stored in the scheduling database. Then, the scheduling database is used to obtain the historical computing power scheduling plan of the multiple historical edge data collected by the multiple edge nodes at the previous moment and allocated to the multiple ARM servers for processing.

[0051] S400: Calculate the scheduling deviations between the multiple optimized computing power scheduling schemes in the optimized computing power scheduling scheme library and the historical computing power scheduling scheme, output the optimized computing power scheduling scheme with the smallest scheduling deviation as the optimal computing power scheduling scheme, and control the multiple edge nodes to transmit the multiple edge data to the corresponding multiple ARM servers for processing according to the optimal computing power scheduling scheme.

[0052] In one embodiment, step S400 of the present application further includes:

[0053] S410: Extract the first optimized computing power scheduling scheme from the optimized computing power scheduling scheme library, and obtain the first processing mapping relationship between the multiple edge data collected by the multiple edge nodes in the first optimized computing power scheduling scheme and the multiple ARM servers, wherein the first processing mapping relationship includes multiple first mapping tuples of edge nodes and ARM servers; S420: Obtain the historical processing mapping relationship between the multiple edge nodes and the multiple ARM servers in the historical computing power scheduling scheme, wherein the historical processing mapping relationship includes multiple historical mapping tuples of edge nodes and ARM servers; S430: Calculate the proportion of different mapping tuples in the multiple first mapping tuples and the multiple historical mapping tuples to obtain a first scheduling deviation; S440: Continue to calculate the scheduling deviations of other optimized computing power scheduling schemes and the historical computing power scheduling scheme, and output the optimized computing power scheduling scheme with the smallest scheduling deviation as the optimal computing power scheduling scheme.

[0054] Specifically, first, any one optimized computing power scheduling scheme is randomly selected from the optimized computing power scheduling scheme library and set as the first optimized computing power scheduling scheme; then, a first processing mapping relationship between multiple edge data collected by the multiple edge nodes in the first optimized computing power scheduling scheme and the multiple ARM servers is obtained, wherein the first processing mapping relationship includes multiple first mapping tuples of edge nodes and ARM servers. For example, each edge node and ARM server is numbered, such as edge node A, edge node B, edge node C, ARM server 1, ARM server 2, ARM server 3, then the mapping tuples are (edge ​​node A, ARM server 2), (edge ​​node B, ARM server 1), etc.

[0055] Then, the historical processing mapping relationship between the multiple edge nodes and the multiple ARM servers in the historical computing power scheduling scheme is obtained, wherein the historical processing mapping relationship includes multiple historical mapping tuples of edge nodes and ARM servers, wherein the numbering rules of the historical mapping tuples and the mapping tuples are the same, such as (historical edge node A, historical ARM server 3). Then, the multiple first mapping tuples and the multiple historical mapping tuples are traversed and compared and analyzed, and the proportions of different mapping tuples in the multiple first mapping tuples and the multiple historical mapping tuples are counted, and the proportions are set as the first scheduling deviation to obtain the first scheduling deviation, wherein the larger the scheduling deviation, the greater the difference between the current scheduling scheme and the historical scheme.

[0056] Then continue to calculate the scheduling deviations of other optimized computing power scheduling plans and the historical computing power scheduling plan, obtain multiple scheduling deviations of multiple optimized computing power scheduling plans, and select the optimized computing power scheduling plan with the smallest scheduling deviation as the optimal computing power scheduling plan. By quickly calculating the scheduling deviation and selecting the optimal plan with the smallest deviation, the time for computing power scheduling updates can be significantly reduced, and the system response speed can be improved.

[0057] Finally, according to the optimal computing power scheduling scheme, the multiple edge nodes are controlled to transmit the multiple edge data to the corresponding multiple ARM servers for processing.

[0058] The computing power scheduling method of an ARM cluster server provided by an embodiment of the present invention has at least the following technical effects:

[0059] By monitoring multiple edge data currently collected by multiple edge nodes, monitoring multiple current occupancy information of multiple ARM servers, and configuring computing power scheduling optimization weights according to multiple occupancy information, energy efficiency weights and timeliness weights are obtained; then based on the multiple edge data, a computing power scheduling processing space is constructed, and according to the energy efficiency weights and timeliness weights, the computing power scheduling scheme for scheduling the multiple edge data to the multiple ARM servers for processing is optimized to obtain an optimized computing power scheduling scheme library; on the other hand, a historical computing power scheduling scheme for the multiple ARM servers to process the edge data of the multiple edge nodes at a previous time point is obtained; then the scheduling deviations of the multiple optimized computing power scheduling schemes in the optimized computing power scheduling scheme library and the historical computing power scheduling scheme are calculated, and the optimized computing power scheduling scheme with the smallest scheduling deviation is output as the optimal computing power scheduling scheme; finally, according to the optimal computing power scheduling scheme, the multiple edge nodes are controlled to transmit the multiple edge data to the corresponding multiple ARM servers for processing; the accuracy, scientificity and practicality of the computing power scheduling scheme setting can be improved, the computing power load can be ensured to be evenly distributed among multiple servers, and the overall resource utilization efficiency can be optimized, thereby achieving the technical effect of reducing the energy consumption of system operation and improving the overall task processing efficiency.

[0060] Embodiment 2, as Figure 2 As shown, based on the same inventive concept as the computing power scheduling method of an ARM cluster server provided in Embodiment 1, an embodiment of the present invention further provides a computing power scheduling system for an ARM cluster server, the system includes an ARM cluster server, the ARM cluster server includes multiple ARM servers that are communicatively connected to multiple edge nodes, and also includes:

[0061] The scheduling weight optimization module 01 is used to monitor the multiple edge data currently collected by the multiple edge nodes, monitor the multiple current occupancy information of the multiple ARM servers, and perform computing power scheduling optimization weight configuration to obtain energy efficiency weight and time weight; the scheduling scheme optimization module 02 is used to construct a computing power scheduling processing space based on the multiple edge data, and optimize the computing power scheduling scheme for scheduling the multiple edge data to the multiple ARM servers for processing according to the energy efficiency weight and time weight, and obtain an optimized computing power scheduling scheme library; the historical scheduling scheme acquisition module 03 is used to obtain the historical computing power scheduling scheme for the multiple ARM servers to process the edge data of the multiple edge nodes at the previous time point; the optimal scheduling scheme acquisition module 04 is used to calculate the scheduling deviation between the multiple optimized computing power scheduling schemes in the optimized computing power scheduling scheme library and the historical computing power scheduling scheme, and output the optimized computing power scheduling scheme with the smallest scheduling deviation as the optimal computing power scheduling scheme, and according to the optimal computing power scheduling scheme, control the multiple edge nodes to transmit the multiple edge data to the corresponding multiple ARM servers for processing.

[0062] In one embodiment, the computing power scheduling system of the ARM cluster server is also used for:

[0063] Monitor the multiple edge data currently collected by the multiple edge nodes, and monitor the multiple occupancy information of the multiple ARM servers; calculate the energy efficiency weight correction coefficient based on the multiple occupancy information; use the energy efficiency weight correction coefficient to correct the preset energy efficiency weight to obtain the energy efficiency weight, and calculate the timeliness weight based on the energy efficiency weight, wherein the sum of the energy efficiency weight and the timeliness weight is 1.

[0064] In one embodiment, the computing power scheduling system of the ARM cluster server is also used for:

[0065] The average of the plurality of occupancy rate information is calculated to obtain average occupancy rate information; and the ratio of the average occupancy rate information to the occupancy rate threshold is calculated to obtain an energy efficiency weight correction coefficient.

[0066] In one embodiment, the computing power scheduling system of the ARM cluster server is also used for:

[0067] The multiple edge data collected by the multiple edge nodes are randomly distributed to the multiple ARM servers for processing to obtain multiple sample computing power scheduling schemes, and a computing power scheduling processing space is constructed; a first computing power scheduling scheme is randomly generated in the computing power scheduling processing space; according to the edge data processed by the multiple ARM servers in the first computing power scheduling scheme, combined with the energy efficiency weight and the timeliness weight, a first computing power scheduling fitness is calculated and obtained; the computing power scheduling scheme is continuously optimized in the computing power scheduling processing space until the optimization converges, and N sample computing power scheduling schemes with the largest computing power scheduling fitness are output to construct an optimized computing power scheduling scheme library, where N is an integer greater than 1.

[0068] In one embodiment, the computing power scheduling system of the ARM cluster server is also used for:

[0069] According to the edge data processed by the multiple ARM servers in the first computing power scheduling scheme, multiple processing occupancy rates are obtained respectively, and multiple scheduling occupancy rates are calculated in combination with the multiple occupancy rate information; according to the edge nodes to which the multiple edge data belong, multiple transmission delays of the multiple edge data in the first computing power scheduling scheme to the corresponding multiple ARM servers for processing are obtained by testing, and the total transmission delay is calculated; according to the multiple scheduling occupancy rates, the total transmission delay, the energy efficiency weight and the time efficiency weight, the first computing power scheduling fitness of the first computing power scheduling scheme is calculated, as shown in the following formula: ;in, is the computing power scheduling fitness, and are energy efficiency weight and time efficiency weight, M is the number of multiple ARM servers, is the scheduling occupancy rate of the i-th ARM server in the computing power scheduling scheme, D is the variance of the M scheduling occupancy rates, T is the total transmission delay, The preset transmission delay.

[0070] In one embodiment, the computing power scheduling system of the ARM cluster server is also used for:

[0071] A historical computing power scheduling scheme is obtained by allocating multiple historical edge data collected by the multiple edge nodes at a previous moment to the multiple ARM servers for processing.

[0072] In one embodiment, the computing power scheduling system of the ARM cluster server is also used for:

[0073] Extract a first optimized computing power scheduling scheme from the optimized computing power scheduling scheme library, obtain a first processing mapping relationship between multiple edge data collected by the multiple edge nodes in the first optimized computing power scheduling scheme and the multiple ARM servers, the first processing mapping relationship includes multiple first mapping tuples of edge nodes and ARM servers; obtain a historical processing mapping relationship between the multiple edge nodes and the multiple ARM servers in the historical computing power scheduling scheme, the historical processing mapping relationship includes multiple historical mapping tuples of edge nodes and ARM servers; calculate the proportion of different mapping tuples in the multiple first mapping tuples and the multiple historical mapping tuples, and obtain a first scheduling deviation; continue to calculate the scheduling deviations of other optimized computing power scheduling schemes and the historical computing power scheduling scheme, and output the optimized computing power scheduling scheme with the smallest scheduling deviation as the optimal computing power scheduling scheme.

[0074] For example 3, please refer to Figure 3 , Figure 3 Schematic diagram of an electronic device provided by an embodiment of the present invention. Figure 3As shown, an embodiment of the present invention provides an electronic device 500, including a memory 510, a processor 520, and a first computer program 511 stored in the memory 510 and executable on the processor 520. When the processor 520 executes the first computer program 511, the following steps are implemented: monitoring multiple edge data currently collected by multiple edge nodes, monitoring multiple current occupancy information of multiple ARM servers, and performing computing power scheduling optimization weight configuration to obtain energy efficiency weights and time weights; based on the multiple edge data, constructing a computing power scheduling processing space, and optimizing the computing power scheduling scheme for scheduling the multiple edge data to the multiple ARM servers for processing according to the energy efficiency weights and time weights, to obtain an optimized computing power scheduling scheme library; obtaining the historical computing power scheduling schemes for the multiple ARM servers to process the edge data of the multiple edge nodes at the previous time point; calculating the scheduling deviations between the multiple optimized computing power scheduling schemes in the optimized computing power scheduling scheme library and the historical computing power scheduling schemes, and outputting the optimized computing power scheduling scheme with the smallest scheduling deviation as the optimal computing power scheduling scheme, and according to the optimal computing power scheduling scheme, controlling the multiple edge nodes to transmit the multiple edge data to the corresponding multiple ARM servers for processing.

[0075] Example 4, please refer to Figure 4 , Figure 4 A schematic diagram of an embodiment of a computer-readable storage medium provided in an embodiment of the present invention. Figure 4 As shown, this embodiment provides a computer-readable storage medium 600, on which a second computer program 611 is stored. When the second computer program 611 is executed by a processor, the following steps are implemented: monitoring multiple edge data currently collected by multiple edge nodes, monitoring multiple current occupancy information of multiple ARM servers, and performing computing power scheduling optimization weight configuration to obtain energy efficiency weights and time weights; based on the multiple edge data, constructing a computing power scheduling processing space, and optimizing the computing power scheduling scheme for scheduling the multiple edge data to the multiple ARM servers for processing according to the energy efficiency weights and time weights, to obtain an optimized computing power scheduling scheme library; obtaining the historical computing power scheduling schemes for the multiple ARM servers to process the edge data of the multiple edge nodes at the previous time point; calculating the scheduling deviations of multiple optimized computing power scheduling schemes in the optimized computing power scheduling scheme library and the historical computing power scheduling schemes, and outputting the optimized computing power scheduling scheme with the smallest scheduling deviation as the optimal computing power scheduling scheme, and according to the optimal computing power scheduling scheme, controlling the multiple edge nodes to transmit the multiple edge data to the corresponding multiple ARM servers for processing.

[0076] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0077] It will be appreciated by those skilled in the art 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. Furthermore, 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.

[0078] 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 computer, 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.

[0079] 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.

[0080] 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 in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0081] Although preferred embodiments of the present invention have been described, additional changes and modifications may occur to these embodiments once those skilled in the art understand the basic inventive concepts.

[0082] Obviously, those skilled in the art can 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 belong to the scope of the present invention and its equivalent technologies, the present invention is also intended to include these changes and variations.

Claims

1. A computing power scheduling method for an ARM cluster server, characterized in that: The method is applied to an ARM cluster server, wherein the ARM cluster server includes multiple ARM servers that are communicatively connected to multiple edge nodes, and the method includes: Monitor the multiple edge data currently collected by the multiple edge nodes, monitor the multiple occupancy information of the multiple ARM servers, and perform weight configuration for computing power scheduling optimization to obtain energy efficiency weight and timeliness weight; Based on the multiple edge data, a computing power scheduling processing space is constructed, and according to the energy efficiency weight and the time efficiency weight, a computing power scheduling scheme for scheduling the multiple edge data to the multiple ARM servers for processing is optimized to obtain an optimized computing power scheduling scheme library; Obtain a historical computing power scheduling plan for processing edge data of the multiple edge nodes by the multiple ARM servers at a previous time point; Calculate the scheduling deviations of multiple optimized computing power scheduling schemes in the optimized computing power scheduling scheme library and the historical computing power scheduling scheme, output the optimized computing power scheduling scheme with the smallest scheduling deviation as the optimal computing power scheduling scheme, and control the multiple edge nodes to transmit the multiple edge data to the corresponding multiple ARM servers for processing according to the optimal computing power scheduling scheme; Wherein, based on the plurality of edge data, a computing power scheduling processing space is constructed, and according to the energy efficiency weight and the timeliness weight, a computing power scheduling scheme for scheduling the plurality of edge data to the plurality of ARM servers for processing is optimized, including: Randomly distribute the multiple edge data collected by the multiple edge nodes to the multiple ARM servers for processing, obtain multiple sample computing power scheduling solutions, and construct a computing power scheduling processing space; Randomly generating a first computing power scheduling solution in the computing power scheduling processing space; According to the edge data processed by the multiple ARM servers in the first computing power scheduling scheme, combined with the energy efficiency weight and the timeliness weight, a first computing power scheduling fitness is calculated; Continue to optimize the computing power scheduling scheme in the computing power scheduling processing space until the optimization converges, output N sample computing power scheduling schemes with the largest computing power scheduling fitness, and build an optimized computing power scheduling scheme library, where N is an integer greater than 1; Among them, according to the edge data processed by the multiple ARM servers in the first computing power scheduling scheme, combined with the energy efficiency weight and the timeliness weight, the first computing power scheduling fitness is calculated, including: According to the edge data processed by the multiple ARM servers in the first computing power scheduling solution, multiple processing occupancy rates are obtained respectively, and multiple scheduling occupancy rates are calculated by combining the multiple occupancy rate information; According to the edge nodes to which the multiple edge data belong, test and obtain multiple transmission delays of the multiple edge data in the first computing power scheduling scheme to the corresponding multiple ARM servers for processing, and calculate and obtain the total transmission delay; According to the multiple scheduling occupancy rates, total transmission delays, energy efficiency weights, and time efficiency weights, the first computing power scheduling fitness of the first computing power scheduling scheme is calculated and obtained as follows: ； in, is the computing power scheduling fitness, and are energy efficiency weight and time efficiency weight, M is the number of multiple ARM servers, is the scheduling occupancy rate of the i-th ARM server in the computing power scheduling scheme, D is the variance of the M scheduling occupancy rates, T is the total transmission delay, The preset transmission delay.

2. The computing power scheduling method of an ARM cluster server according to claim 1, characterized in that: Monitoring the multiple edge data currently collected by the multiple edge nodes, monitoring the multiple occupancy information of the multiple ARM servers, and performing weight configuration for computing power scheduling optimization, including: Monitoring a plurality of edge data currently collected by the plurality of edge nodes, and monitoring a plurality of current occupancy information of the plurality of ARM servers; Calculating an energy efficiency weight correction coefficient according to the plurality of occupancy rate information; The energy efficiency weight correction coefficient is used to correct the preset energy efficiency weight to obtain the energy efficiency weight, and the time efficiency weight is calculated based on the energy efficiency weight, wherein the sum of the energy efficiency weight and the time efficiency weight is 1.

3. The computing power scheduling method of an ARM cluster server according to claim 2, characterized in that: Calculating an energy efficiency weight correction coefficient according to the plurality of occupancy rate information includes: Calculating the average of the plurality of occupancy rate information to obtain average occupancy rate information; The ratio of the average occupancy rate information to the occupancy rate threshold value is calculated to obtain an energy efficiency weight correction coefficient.

4. The computing power scheduling method of an ARM cluster server according to claim 1, characterized in that: Obtaining a historical computing power scheduling scheme for processing edge data of the multiple edge nodes by the multiple ARM servers at a previous time point, including: A historical computing power scheduling scheme is obtained by allocating multiple historical edge data collected by the multiple edge nodes at a previous moment to the multiple ARM servers for processing.

5. The computing power scheduling method of an ARM cluster server according to claim 1, characterized in that: Calculating the scheduling deviations between multiple optimized computing power scheduling schemes in the optimized computing power scheduling scheme library and the historical computing power scheduling scheme, and outputting the optimized computing power scheduling scheme with the smallest scheduling deviation as the optimal computing power scheduling scheme, including: Extracting a first optimized computing power scheduling scheme from the optimized computing power scheduling scheme library, obtaining a first processing mapping relationship between a plurality of edge data collected by the plurality of edge nodes in the first optimized computing power scheduling scheme and the plurality of ARM servers, wherein the first processing mapping relationship includes a plurality of first mapping tuples of edge nodes and ARM servers; Acquire a historical processing mapping relationship between the plurality of edge nodes and the plurality of ARM servers in the historical computing power scheduling scheme, wherein the historical processing mapping relationship includes a plurality of historical mapping tuples of edge nodes and ARM servers; Calculating the proportions of different mapping tuples in the plurality of first mapping tuples and the plurality of historical mapping tuples to obtain a first scheduling deviation; Continue to calculate the scheduling deviations of other optimized computing power scheduling plans and the historical computing power scheduling plan, and output the optimized computing power scheduling plan with the smallest scheduling deviation as the optimal computing power scheduling plan.

6. A computing power scheduling system for an ARM cluster server, characterized in that: The system for implementing the computing power scheduling method of an ARM cluster server according to any one of claims 1 to 5 comprises an ARM cluster server, wherein the ARM cluster server comprises a plurality of ARM servers in communication connection with a plurality of edge nodes, and further comprises: A scheduling weight optimization module is used to monitor the multiple edge data currently collected by the multiple edge nodes, monitor the multiple current occupancy information of the multiple ARM servers, and perform computing power scheduling optimization weight configuration to obtain energy efficiency weight and time efficiency weight; A scheduling scheme optimization module is used to construct a computing power scheduling processing space based on the multiple edge data, and optimize the computing power scheduling scheme for scheduling the multiple edge data to the multiple ARM servers for processing according to the energy efficiency weight and the timeliness weight, so as to obtain an optimized computing power scheduling scheme library; A historical scheduling scheme acquisition module, used to acquire a historical computing power scheduling scheme for processing edge data of the multiple edge nodes by the multiple ARM servers at a previous time point; The optimal scheduling scheme acquisition module is used to calculate the scheduling deviations between multiple optimized computing power scheduling schemes in the optimized computing power scheduling scheme library and the historical computing power scheduling scheme, and output the optimized computing power scheduling scheme with the smallest scheduling deviation as the optimal computing power scheduling scheme. According to the optimal computing power scheduling scheme, the multiple edge nodes are controlled to transmit the multiple edge data to the corresponding multiple ARM servers for processing.

7. An electronic device, characterized in that: include: Memory for storing computer software programs; A processor is used to read and execute the computer software program, thereby implementing the steps of a computing power scheduling method for an ARM cluster server as described in any one of claims 1 to 5.

8. A non-transitory computer-readable storage medium, characterized in that: The storage medium stores a computer software program, which, when executed by a processor, implements the steps of a computing power scheduling method for an ARM cluster server as described in any one of claims 1 to 5.

Images