Cloud computing energy consumption optimization method and system based on task migration and medium

Abstract

The invention discloses a cloud computing energy consumption optimization method and system based on task migration and a medium. The method comprises the following steps: constructing a server total energy consumption model according to energy consumption sub-models of servers in different states in an established target cloud computing data center; by taking the total energy consumption output by the minimum server total energy consumption model as a target function, calculating a resource contention rate between a current task running on each virtual machine in each server and other tasks by using a multiple correlation coefficient method; determining a to-be-migrated server list according to the resource contention rate and the resource utilization rate of each server, and sorting the to-be-migrated servers in the to-be-migrated server list according to a preset priority; and according to the sorting result, the energy consumption optimization operation of task migration and shutdown of the virtual machine is executed on the server to be migrated in sequence, so that the energy consumption of the cloud computing data center is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of power cloud computing, and in particular, to a cloud computing energy consumption optimization method, system and medium based on task migration. Background Art

[0002] The business of the power industry is complex, including both tightly coupled critical business systems and numerous loosely coupled business systems. Moreover, the software and hardware platforms involved in these loosely coupled systems are also comprehensive, including databases, Web applications, etc. The hardware configuration and performance of these servers will affect the stable operation of the system. As a distributed computing paradigm, cloud computing (CC) not only provides massive computing and storage resources, but also deploys various flexible analysis and processing functions, providing an efficient solution for almost all types of large-scale computing. However, with the growing demand for cloud services and the rapid expansion of the scale of cloud data centers, the energy consumption of data centers has increased sharply.

[0003] In the process of dealing with the energy consumption of cloud computing data centers, the existing cooling system optimization technologies are often affected by the external environment, resulting in poor reduction effects on the energy consumption of cloud computing data centers and affecting the overall performance of servers.

[0004] Therefore, how to effectively reduce the energy consumption in the process of power cloud computing and improve the performance of servers has become a technical problem that needs to be urgently solved by those skilled in the art. Summary of the Invention

[0005] The present invention provides a cloud computing energy consumption optimization method, system and medium based on task migration to solve the problem of accurately determining the servers to be migrated and their energy consumption situations and ensuring the resource utilization rate.

[0006] To solve the above technical problems, embodiments of the present invention provide a cloud computing energy consumption optimization method, system and medium based on task migration, including: Constructing a total server energy consumption model according to the energy consumption sub-models of each server in the target cloud computing data center in different states; Taking the total energy consumption output by the total server energy consumption model as the objective function, and calculating the resource contention rate between the current tasks running on each virtual machine in each server and other tasks by using the multiple correlation coefficient method; Determining a list of servers to be migrated according to the resource contention rate and the resource utilization rate of each server, and sorting the servers to be migrated in the list of servers to be migrated according to a preset priority; Performing energy consumption optimization operations of virtual machine task migration and shutdown on the servers to be migrated in sequence according to the sorting result.

[0007] Further, constructing a total server energy consumption model based on the energy consumption sub-models of each server in different states includes: Establishing energy consumption sub-models for each server in the working, transitioning, idle, sleeping, and shutdown states, which reflect computing energy consumption, storage energy consumption, network energy consumption, idle energy consumption, sleeping energy consumption, and transitioning energy consumption; Integrating each of the energy consumption sub-models to establish the total server energy consumption model.

[0008] Further, calculating the resource contention rate between the current task running on each virtual machine in each server and other tasks includes: Obtaining the resource requirements of each task running on each virtual machine and representing them in a vectorized form; the resource requirements reflect the demand of the task for the CPU resources, storage resources, and network device resources of the server; Calculating the complex correlation coefficient between the first vector matrix corresponding to the resource requirements of the current task and the second vector matrix composed of the resource requirements of other tasks, and determining the resource contention rate between tasks in each virtual machine according to the complex correlation coefficient.

[0009] Further, determining a list of servers to be migrated according to the resource contention rate and the resource utilization rate of each server includes: Taking the virtual machine migration integration requirement as an adjustment parameter, and dynamically setting the high-load threshold and low-load threshold of the resource utilization rate by using the absolute median difference method; Including servers with a resource utilization rate lower than the low-load threshold and a resource contention rate meeting the preset value in the list of servers to be migrated.

[0010] Further, taking the virtual machine migration integration requirement as an adjustment parameter and dynamically setting the high-load threshold and low-load threshold of the resource utilization rate by using the absolute median difference method includes: The high-load threshold and the low-load threshold are represented by the following formulas: Where S is the requirement for virtual machine migration integration; is the high-load threshold, is the low-load threshold; is the absolute median difference.

[0011] Further, in the process of determining the list of servers to be migrated according to the resource contention rate and the resource utilization rate of each server, it also includes: Determining a group of candidate servers as candidate variables according to the complex correlation coefficient, and performing an information quantity score on each variable in the candidate variables; Traverse the scored candidate variables using a forward search strategy. When the cumulative information contribution rate reaches a preset condition, determine the target variable set, and include the corresponding servers in the target variable set in the list of servers to be migrated.

[0012] Further, according to the sorting result, sequentially perform virtual machine task migration and shutdown operations on the servers to be migrated, including: Include the servers with resource utilization rates exceeding the high-load threshold and the servers that have been shut down in the list of non-migratable servers; According to the sorting result, sequentially allocate target servers for the virtual machines in each server to be migrated, excluding the servers in the list of non-migratable servers; Shut down and migrate the servers for which virtual machine allocation has been performed to the list of non-migratable servers.

[0013] Another embodiment of the present invention provides a cloud computing energy consumption optimization system based on task migration, including: An energy consumption model construction module, configured to construct a total server energy consumption model according to the energy consumption sub-models of each server in the established target cloud computing data center in different states; A resource contention rate determination module, configured to use the multiple correlation coefficient method to calculate the resource contention rate between the current tasks running on each virtual machine in each server and other tasks, with the goal of minimizing the total energy consumption output by the total server energy consumption model; A list determination module of servers to be migrated, configured to determine the list of servers to be migrated according to the resource contention rate and the resource utilization rate of each server, and sort the servers to be migrated in the list of servers to be migrated according to a preset priority; An energy consumption optimization module, configured to sequentially perform energy consumption optimization operations of virtual machine task migration and shutdown on the servers to be migrated according to the sorting result.

[0014] Another embodiment of the present invention provides a computer device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the above-mentioned cloud computing energy consumption optimization method based on task migration is implemented.

[0015] Another embodiment of the present invention provides a computer-readable storage medium. The computer-readable storage medium stores a computer program. When the device where the computer-readable storage medium is located executes the computer program, the above-mentioned cloud computing energy consumption optimization method based on task migration is implemented.

[0016] Compared with the prior art, the beneficial effects of the embodiments of the present invention are at least one of the following: In the embodiments of the present invention, an energy consumption model is established to accurately evaluate the energy consumption of the server under different workloads and states; by setting an objective function with the goal of minimizing energy consumption and combining the multiple correlation coefficient to determine the task migration situation of the server, so that the system can effectively identify and optimize inefficient servers. Once the tasks on the server are migrated to the target server, the system can safely shut down these idle servers to ensure the maximization of resource utilization, thereby reducing the overall energy consumption. Brief Description of the Drawings

[0017] Figure 1 It is a schematic flowchart of a cloud computing energy consumption optimization method based on task migration in one embodiment of the present invention; Figure 2 It is a schematic structural diagram of a cloud computing energy consumption optimization system based on task migration in one embodiment of the present invention; Figure 3 It is a structural block diagram of a preferred embodiment of a computer device provided by the present invention. Detailed Embodiments

[0018] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

[0019] In the description of the present application, the terms "first", "second", "third", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first", "second", "third", etc. may explicitly or implicitly include one or more of such features. In the description of the present application, unless otherwise specified, the meaning of "a plurality" is two or more.

[0020] In the description of the present application, it should be noted that unless otherwise clearly defined and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two components. The terms "vertical", "horizontal", "left", "right", "up", "down" and similar expressions used herein are only for the purpose of illustration, rather than indicating or implying that the device or component referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation of the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific circumstances.

[0021] In the description of the present application, it should be noted that unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as those commonly understood by those skilled in the technical field to which the present invention belongs. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific circumstances.

[0022] An embodiment of the present invention provides a cloud computing energy consumption optimization method based on task migration. Specifically, please refer to Figure 1 , Figure 1 which shows a schematic flow chart of the cloud computing energy consumption optimization method based on task migration in one of the embodiments of the present invention, including the following steps: S1. Construct a total server energy consumption model according to the energy consumption sub-models of each server in the established target cloud computing data center under different states.

[0023] It can be understood that in a cloud computing data center, the energy consumption of servers in the idle state during actual operation usually accounts for 70% of the energy consumption in the full-load state. Therefore, in this embodiment, β is used to represent the ratio of the idle power consumption to the full-load power consumption, and preferably β is 0.7. Then the relationship between the actual power of the server, the idle power, the full-load power, and the CPU utilization rate can be expressed as In the formula, represents the actual power, represents the power at idle, represents the power at full load, represents the utilization rate.

[0024] It can be seen from the above formula that Indicates that the runtime power of the server closely follows the CPU utilization rate and changes accordingly. Then, the energy consumption generated by the server within the target time is expressed as: It should be understood that to reduce the energy consumption of the server, it is necessary to reduce the idle time of the server. For a server that runs for a long time, only when it is used to process tasks does it mean that the energy consumption at this time is useful. During other times when the server is in an idle waiting state, it belongs to energy consumption waste.

[0025] For idle servers, if the shutdown operation is performed, it will have a direct effect on energy conservation. However, if the number of shutdowns is large and the number of working servers cannot meet the resource requirements of users, this will violate the SLA agreement, and it is necessary to start servers from the shutdown queue to increase the resources for processing user requests. Therefore, based on the above formula, as an example, in some embodiments of the present invention, the power consumption of the j-th server in the cloud data center is expressed as follows: In the formula, represents the power consumption in the idle state; represents the startup duration of the j-th server; represents the energy consumption of the k-th virtual machine running on this server, and w represents the number of virtual machines on this server.

[0026] Next, based on the above power consumption formula of the server, the total energy consumption of the server is modeled. It can be understood that the main energy-consuming objects in the energy consumption of the server are the CPU, storage, and network devices. The server uses virtualization technology to virtualize one server into multiple servers, and its purpose is to improve the CPU utilization rate and reduce the idle time. Based on this, the embodiments of the present invention reduce the energy consumption of the cloud computing data center by considering virtual machine migration. Assume that within a period of time [t1, t2], the virtual machines on the server need to be migrated to another server. Assume that the migration start time is , and the migration end time is , where the migration energy consumption of the virtual machine represents the sum of the energy consumption of these two servers during migration within , , as shown in the following formula: In the formula, , are the CPU utilization rates of the migrating virtual machine in the two servers respectively; and are the weight coefficients of the two servers respectively; P is a function.

[0027] The resource power consumption of the server varies in different states, including five states: working, transitioning, idle, sleeping, and shutdown.

[0028] In this embodiment, corresponding energy consumption sub-models are established according to the above five states, which are expressed as follows: 1. Sleep energy consumption in the sleep state: The sleep power consumption means that after saving the running state data to the hard disk, the whole machine stops power supply completely. In this embodiment, the sleep energy consumption sub-model is expressed as: In the formula, represents the sleep energy consumption of server j, refers to the sleep power consumption of server j, represents time.

[0029] 2. Idle energy consumption in the idle state: The idle energy consumption is mainly the consumption of the CPU, graphics card, memory, etc. in the idle state. In this embodiment, the idle energy consumption sub-model is expressed as: In the formula, represents the idle energy consumption of server j, refers to the idle power consumption of server j, refers to the idle time.

[0030] 3. Transition energy consumption in the transition state: The transition energy consumption refers to the energy consumption of the server during the process of transitioning from one state to another, which includes all components of the server, and the most important ones are the CPU, graphics card, etc. In this embodiment, the transition energy consumption sub-model is expressed as: In the formula, represents the transition energy consumption of server j, represents the transition power consumption of server j, represents the server transition time.

[0031] 4. Working energy consumption in the working state: The working state mainly includes computing energy consumption, storage energy consumption, and network energy consumption. In the working state, the execution of tasks mainly consumes the CPU and memory, and the corresponding energy consumption is called computing energy consumption. The computing energy consumption sub-model is expressed as: In the formula, represents the energy consumed by server j when executing task i, is the computing power consumption per unit time, represents the number of task instructions, Represents the processor speed.

[0032] The storage energy consumption represents the energy consumption generated when the disk reads and writes data. The storage energy consumption sub-model is expressed as: In the formula, is the storage energy consumption of server j, is the storage power consumption per unit time, is the read / write time of task i on server j, represents the amount of data read and written by the task, represents the speed of disk read and write.

[0033] The network energy consumption represents the energy consumption generated during the process of data transmission from the source server to the destination server. The transmission time is restricted by the network bandwidth, thus affecting the generated energy consumption. The network energy consumption sub-model is expressed as: In the formula, is the energy consumption during the transmission of task i to server j, represents the transmission power consumption, represents the time consumed by the transmission of task i, represents the number of bytes of task i, represents the network bandwidth.

[0034] Integrating each of the above energy consumption sub-models, the corresponding comprehensive server energy consumption model can be obtained, which is expressed by the following formula: Then the total energy consumption generated by the server during the migration process includes the energy consumption output by each of the above energy consumption sub-models and the virtual machine migration energy consumption. The server total energy consumption model is expressed as: By establishing an energy consumption model in the embodiments of the present invention, the energy consumption of the server under different workloads and states can be accurately evaluated, providing data support for subsequent migration optimization operations.

[0035] S2. Taking the total energy consumption output by the server total energy consumption model as the objective function, the resource contention rate between the current task running on each virtual machine in each server and other tasks is calculated by using the multiple correlation coefficient method.

[0036] It should be understood that the embodiments of the present invention aim to reduce the energy consumption of the cloud computing data center by optimizing the virtual machine migration method and shutting down the servers that present idle / low load after the virtual machine migration and allocation.

[0037] Based on this, the embodiments of the present invention will construct an objective function according to the established total server energy consumption model. Specifically, the present invention provides the following examples to describe the construction process of the objective function in detail: Suppose a cloud computing data center contains servers, and there are tasks waiting to be assigned. The triple corresponds to the CPU, storage, and network device resources of the th server respectively. Let be the demand for the three resources of the th task, as shown in the following formula: The allocation situation of tasks on the server can be represented by a two-dimensional matrix where represents the number of tasks, and represents the number of servers. Then being 0 means that task is not assigned to server , and being means that task is assigned to server

[0038] To allocate each task to a suitable server, first, the available resources of the server need to meet the requirements of the task. For each server, the following constraints need to be met: When the above constraints are met, the objective is to minimize the total energy consumption output by the total server energy consumption model, which can be expressed as: After constructing the objective function, the embodiments of the present invention take the optimal output result of the objective function as the goal and perform virtual machine task migration of the server. First, a list of servers to be migrated needs to be determined.

[0039] The embodiments of the present invention use the multiple correlation coefficient method to identify virtual machines with a high probability of resource contention, that is, virtual machines with highly correlated resource requirements between tasks. Resource requirements reflect the demand of tasks for the CPU resources, storage resources, and network device resources of the server. The following the present invention will provide a specific example to describe the process of determining the resource contention rate between tasks in detail: First, it is necessary to obtain the resource requirements of each task running on each virtual machine and represent them in a vectorized manner.

[0040] Exemplarily, when there are n tasks, the resource requirements of each task can be vectorized and represented as a 1*3 matrix X, as shown in the following formula: The resource requirements of other tasks can be quantified as an (n-1)*3 matrix , as shown in the following formula: According to the above formula, calculate the complex correlation coefficient between the resource requirements of the current task corresponding to the first vector matrix X and the second vector matrix Y composed of the resource requirements of other tasks. Specifically, it includes the following steps: (1) Calculate the correlation coefficient of each row vector in X and Y. represents the covariance, represents the variance, then the correlation coefficient is expressed as: (2) Generate a new correlation coefficient matrix R from all the correlation coefficients: For each task, if each item in its correlation coefficient matrix is 1, it means that the probability of resource contention with other tasks is the highest. At this time, use to represent: In the embodiment of the present invention, by calculating and the reciprocal of the Euclidean distance between them is used as the complex correlation coefficient between X and Y. The greater the Euclidean distance, the farther the distance between the two vectors, indicating the lower the correlation. Therefore, the reciprocal of the Euclidean distance can be used to represent the complex correlation coefficient. and The Euclidean distance between them is as follows: (3) Then the complex correlation coefficient between X and Y is expressed as: According to the calculated complex correlation coefficient, determine the resource contention rate between tasks in each virtual machine. The greater the complex correlation coefficient, the higher the degree of linear correlation between tasks and the greater the probability of resource contention. In the embodiment of the present invention, by statistically quantifying the resource contention risk through correlation, high-conflict tasks can be preferentially migrated during subsequent migration, reducing performance interference after migration.

[0041] S3. Determine the list of servers to be migrated based on the resource contention rate and the resource utilization rate of each server, and sort the servers to be migrated in the list of servers to be migrated according to a preset priority.

[0042] It can be understood that if the resource contention rate of the virtual machine in this server is relatively high, it will lead to a decline in the performance of the virtual machine itself, cause delays in task execution, and high-contention virtual machines are usually tasks with intensive resource requirements. Spreading them across multiple servers can achieve load balancing.

[0043] Based on this, in the embodiment of the present invention, the servers with the resource contention rate meeting the preset value are included in the list of servers to be migrated, that is, the servers corresponding to the virtual machines with the largest complex correlation coefficient between tasks are selected and included in the list of servers to be migrated.

[0044] Moreover, the present invention fully considers the impact of the resource utilization rate (CPU utilization rate) on the server energy consumption, and also includes the servers with the resource utilization rate lower than the low-load threshold in the list of servers to be migrated.

[0045] Considering the characteristic that the task volume of the cloud computing data center changes with different time periods, in this embodiment, the high-load threshold and the low-load threshold of the resource utilization rate are dynamically set by using the absolute median difference method.

[0046] Specifically, in this embodiment, after reordering according to the change of the CPU utilization rate, the absolute median difference is calculated, and the virtual machine migration and integration requirement is used as an adjustment parameter. The high-load threshold and the low-load threshold of the resource utilization rate are dynamically set according to this MAD, which is expressed as: In the formula, s is the requirement for virtual machine migration and integration; is the high-load threshold, is the low-load threshold; MAD is the absolute median difference. It can be understood that s represents the intensity of migrating and integrating virtual machine tasks from low-load or high-resource contention rate servers to fewer servers. The smaller the s value, the less the need for migration. When the s value is large, timely migration and integration of virtual machines are required.

[0047] In the embodiment of the present invention, by setting dynamic thresholds, the number of servers to be migrated is controlled, and the "low-load" servers are added to the list of servers to be migrated.

[0048] It should be noted that in this embodiment, the joint entropy, conditional entropy, and mutual information in information theory are further used to further evaluate the correlation between relevant variables.

[0049] Specifically, a group of candidate servers are determined as candidate variables according to the obtained multiple correlation coefficient, and each variable (server) in the candidate variables is scored for information content, with the aim of selecting the server that contributes the most information to the target variable t (such as energy consumption). . Among them, the process of information content scoring is represented by the following formula: In the formula, V is the set of all candidate servers; S is the set of tasks that have been selected; is the number of data contained in the task set S; is the server and the server The total mutual information between them; is the mutual information between the target t and the candidate input variable ; is The information content score of.

[0050] It can be seen that the information correlation strength of the migration process is determined by the above multiple correlation coefficient. It should be understood that according to the above formula, the score The higher, the corresponding server The more should be preferentially added to the list of servers to be migrated.

[0051] Based on this, in some embodiments of the present invention, a forward search strategy is used to traverse the scored candidate variables, and the target variable set is determined when the cumulative information contribution rate reaches a preset condition. The specific process includes the following steps: 1. Calculate the maximum correlation of all candidate servers, and select the server that makes the largest as the first input variable, which is expressed as: 2. Traverse step by step through the forward search strategy , ,..., and each time select the server that maximizes until the cumulative information contribution rate reaches the threshold.

[0052] Specifically, the selection (sorting) of the remaining variables is obtained by comparing the sizes of the information content scores.

[0053] Exemplarily, in step m ( ), the remaining variables are ( ), where is the ordered set ( ) of the selected (or sorted) variables in ( ), so the variable selected in m is: For each selected input variable, the algorithm dynamically searches one step forward. Thus, this process incrementally evaluates all candidate variables until step , all servers V are evaluated and sorted. The information content score of each variable after sorting can be calculated according to the following formula: 3. Cumulative information contribution rate to determine the target variable set.

[0054] Define the cumulative information contribution rate as , when reaches the maximum value, the corresponding servers in the target ordered variable set will be included in the list of servers to be migrated.

[0055] After determining the list of servers to be migrated, sort the servers to be migrated in the list according to a preset priority. In the embodiment of the present invention, the servers are sorted in ascending order of resource utilization rate.

[0056] S4. According to the sorting result, perform energy consumption optimization operations for virtual machine task migration and shutdown operations on the servers to be migrated in sequence.

[0057] According to the sorting result, allocate target servers to the virtual machines in each server to be migrated in sequence. It should be noted that during the migration allocation process, in order to avoid resource conflicts, ensure service quality (SLA), and ensure the effectiveness of the migration operation during the task migration process, unavailable servers need to be excluded.

[0058] In this embodiment, servers with a resource utilization rate exceeding the high load threshold and servers that have been shut down are included in the list of non-migratable servers as unavailable servers. When allocating target servers to virtual machines, the servers in the list of non-migratable servers will be excluded, and at the same time, the servers that have completed virtual machine allocation will be shut down and migrated to the list of non-migratable servers after the allocation is completed. It should be noted that to ensure the performance of the server during the migration process, the selection of the target server needs to ensure that its CPU computing power does not exceed 90%.

[0059] When selecting the target server, servers with lower loads are given priority. In actual operation, it is necessary to monitor the load conditions of the cloud data center servers in real time to timely grasp the status of the servers.

[0060] Next, the present invention will provide an example to refine the determination of the list of servers to be migrated and the process of virtual machine task migration and shutdown: (1) Set the requirement s for virtual machine migration and consolidation; (2) Obtain the list of unavailable servers, including overloaded servers and shut-down servers; (3) Obtain the resource utilization rate of each server. If the utilization rate is less than the low-load threshold, add the server to the list of servers to be migrated. (4) If the list of servers to be migrated is not empty, execute step (5); otherwise, execute step (14). (5) Sort the servers in the list of servers to be migrated in ascending order according to the resource usage rate. (6) Obtain the list of virtual machines on the first server in the sorted list of servers to be migrated. (7) Allocate a target server for each virtual machine in the list of servers to be migrated. (8) If no target server is found, judge whether each host in the list of servers to be migrated can receive the current virtual machine from the back to the front. (9) If no target server is still found, cancel the migration, exit the current loop, and execute step (4); otherwise, execute step (10). (10) Migrate the tasks of the current virtual machine to the target server. (11) Judge the next virtual machine in the list of servers to be migrated. If the loop ends, execute step (12); otherwise, execute step (7). (12) Delete the current server from the list of servers to be migrated and add it to the list of servers to be shut down. (13) Judge the next low-load server in the list and execute step (4). (14) End.

[0061] It can be understood that if no target server is found during the allocation process, other servers in the list of servers to be migrated will be searched in reverse. In this process, unavailable servers still need to be excluded.

[0062] In summary, in the embodiment of the present invention, by establishing and integrating the energy consumption models of each server in different states, with the goal of minimizing the total energy consumption output by the server, server virtual machine migration and shutdown are performed. In this process, the multiple correlation coefficient between the current task running on each virtual machine in each server and other tasks is calculated as the resource contention rate, and combined with the method of dynamically adjusting the resource utilization threshold, the servers to be migrated are accurately identified and determined, so as to realize the migration of virtual machines from low-load servers and shut down these servers after migration, ensuring the maximization of resource utilization and significantly reducing the energy consumption of the cloud computing data center.

[0063] An embodiment of the present invention provides a cloud computing energy consumption optimization system based on task migration. Specifically, please refer to Figure 2 , Figure 2Shown is a schematic structural diagram of a cloud computing energy consumption optimization system based on task migration in one embodiment of the present invention, including: An energy consumption model construction module M1, configured to construct a total server energy consumption model according to the energy consumption sub-models of each server in the established target cloud computing data center under different states; A resource contention rate determination module M2, configured to use the multiple correlation coefficient method to calculate the resource contention rate between the current task running on each virtual machine in each server and other tasks, with the goal of minimizing the total energy consumption output by the total server energy consumption model; A server to be migrated list determination module M3, configured to determine a list of servers to be migrated according to the resource contention rate and the resource utilization rate of each server, and sort the servers to be migrated in the list of servers to be migrated according to a preset priority; An energy consumption optimization module M4, configured to sequentially perform energy consumption optimization operations of virtual machine task migration and shutdown on the servers to be migrated according to the sorting result.

[0064] As Figure 3 shown, an embodiment of the present invention also provides a computer device, Figure 3 which is a structural block diagram of a preferred embodiment of a computer device provided by the present invention. The computer device includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the above-mentioned cloud computing energy consumption optimization method based on task migration.

[0065] Preferably, the computer program can be divided into one or more modules / units (such as computer program 1, computer program 2,...). The one or more modules / units are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units can be a series of computer program instruction segments capable of performing specific functions, and these instruction segments are used to describe the execution process of the computer program in the computer device.

[0066] The processor may be a Central Processing Unit (CPU), or may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor, or the processor may also be any conventional processor. The processor is the control center of the terminal device, and connects various parts of the terminal device through various interfaces and circuits.

[0067] The memory mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function, etc., and the data storage area can store relevant data, etc. In addition, the memory may be a high-speed random access memory, or may also be a non-volatile memory, such as a plug-in hard disk, a SmartMedia Card (SMC), a Secure Digital (SD) card, a Flash Card, etc., or the memory may also be other volatile solid-state storage devices.

[0068] It should be noted that the above terminal device may include, but is not limited to, a processor and a memory. Those skilled in the art can understand that Figure 3 The structural block diagram is only an example of the terminal device, and does not constitute a limitation on the terminal device. It may include more or fewer components than shown in the figure, or combine certain components, or different components. Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it may include the processes of the embodiments of the above methods. Among them, the storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), or a Random Access Memory (RAM), etc.

[0069] Correspondingly, an embodiment of the present invention provides a computer-readable storage medium, which includes a stored computer program. When the computer program runs, it controls the device where the computer-readable storage medium is located to execute the steps in the method of the above embodiment, for example Figure 1 the steps S1 to S4 described above.

[0070] The embodiments described above merely represent several implementation manners of the present invention. The description is relatively specific and detailed, but it should not be construed as a limitation to the scope of the patent for the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all fall within the protection scope of the present invention. Therefore, the protection scope of the patent for the present invention shall be subject to the appended claims.

Claims

1. A cloud computing energy consumption optimization method based on task migration, characterized in that: include: Construct a server total energy consumption model based on the energy consumption sub-models of each server in different states in the target cloud computing data center; Taking minimizing the total energy consumption output by the server total energy consumption model as the objective function, the resource contention rate between the current task and other tasks running on each virtual machine in each server is calculated using the complex correlation coefficient method; Determining a list of servers to be migrated according to the resource contention rate and the resource utilization rate of each server, and sorting the servers to be migrated in the list of servers to be migrated according to a preset priority; According to the sorting results, energy consumption optimization operations of virtual machine task migration and shutdown are performed in sequence on the servers to be migrated.

2. The cloud computing energy consumption optimization method based on task migration according to claim 1, characterized in that: The server total energy consumption model is constructed according to the established energy consumption sub-models of each server in different states, including: Establish energy consumption sub-models that reflect computing energy consumption, storage energy consumption, network energy consumption, idle energy consumption, sleep energy consumption and conversion energy consumption of each server in working, conversion, idle, sleep and shutdown states; The energy consumption sub-models are integrated to establish the total energy consumption model of the server.

3. The cloud computing energy consumption optimization method based on task migration according to claim 1, characterized in that: The calculating of the resource contention rate between the current task and other tasks running on each virtual machine in each server includes: Obtain resource requirements for each task running on each virtual machine and express them quantitatively; the resource requirements reflect the amount of CPU resources, storage resources, and network device resources that the task requires from the server; The complex correlation coefficient between the first vector matrix corresponding to the resource demand of the current task and the second vector matrix composed of the resource demands of other tasks is calculated, and the resource contention rate between the tasks in each virtual machine is determined according to the complex correlation coefficient.

4. The cloud computing energy consumption optimization method based on task migration according to claim 1, characterized in that: Determining the list of servers to be migrated according to the resource contention rate and the resource utilization rate of each server includes: The virtual machine migration and integration requirements are used as adjustment parameters, and the absolute median difference method is used to dynamically set the high-load threshold and low-load threshold of resource utilization; The servers whose resource utilization rate is lower than the low load threshold and whose resource contention rate meets the preset value are included in the list of servers to be migrated.

5. The cloud computing energy consumption optimization method based on task migration according to claim 4, characterized in that: The virtual machine migration integration requirement is used as the adjustment parameter, and the absolute median difference method is used to dynamically set the high load threshold and the low load threshold of the resource utilization rate, including: The high load threshold and the low load threshold are expressed by the following formula: Where S is the requirement for virtual machine migration and integration; is the high load threshold, is the low load threshold; is the absolute median difference.

6. The cloud computing energy consumption optimization method based on task migration according to claim 3, characterized in that: The process of determining the server list to be migrated according to the resource contention rate and the resource utilization rate of each server also includes: Determine a group of candidate servers as candidate variables according to the multiple correlation coefficient, and perform information scoring on each of the candidate variables; The forward search strategy is used to traverse the scored candidate variables. When the cumulative information contribution rate reaches the preset condition, the target variable set is determined, and the corresponding servers in the target variable set are included in the list of servers to be migrated.

7. The cloud computing energy consumption optimization method based on task migration according to claim 4, characterized in that: The step of sequentially performing virtual machine task migration and shutdown operations on the server to be migrated according to the sorting results includes: Adding servers whose resource utilization exceeds the high load threshold and servers that have been shut down into a list of non-migratable servers; According to the sorting result, the target servers excluding the servers in the non-migratable server list are sequentially allocated to the virtual machines in the servers to be migrated; The server that has executed the virtual machine allocation is shut down and migrated to the non-migratable server list.

8. A cloud computing energy consumption optimization system based on task migration, characterized in that: include: An energy consumption model building module is used to build a server total energy consumption model based on the energy consumption sub-models of each server in different states in the established target cloud computing data center; A resource contention rate determination module is used to calculate the resource contention rate between the current task and other tasks running on each virtual machine in each server by using a complex correlation coefficient method, taking the total energy consumption output by the server total energy consumption model as the objective function; A server list determination module to be migrated is used to determine the server list to be migrated according to the resource contention rate and the resource utilization rate of each server, and to sort the servers to be migrated in the server list to be migrated according to a preset priority; The energy consumption optimization module is used to perform energy consumption optimization operations of virtual machine task migration and shutdown in sequence on the server to be migrated according to the sorting results.

9. A computer device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the method for optimizing cloud computing energy consumption based on task migration as described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein when the device where the computer-readable storage medium is located executes the computer program, the cloud computing energy consumption optimization method based on task migration as described in any one of claims 1 to 7 is implemented.

Images