Load balancing method and device, equipment and medium

Abstract

The embodiment of the invention provides a load balancing method and device, equipment and a medium. The method comprises the following steps: receiving a service request sent by a user, wherein the service request comprises a processing speed requirement on the service request and an arrival speed requirement of the service request; according to the service request, performing linear processing on a pre-obtained power cost model to obtain a target power cost model, the power cost model being a model for determining power cost obtained based on power information of each server and operation temperature information of the corresponding area; and performing minimization problem solving on the target power cost model to obtain server deployment numbers of different regions. According to the invention, the power cost of the server and the power cost of the cooling system in the distributed data center are uniformly distributed, so that the overall power cost of the distributed data center is reduced.

Description

Technical Field

[0001] The present application relates to the field of distributed computing technology, and in particular to a load balancing method, device, equipment and medium. Background Art

[0002] To improve data security and transmission speed and reduce operating costs, cloud service providers have established distributed data centers. With the increase in business demand and the surge in the number of server deployments, electricity costs have accounted for a large proportion of the overall operating costs of data centers. Therefore, it is particularly important to adopt a reasonable request load distribution management mechanism to reduce the electricity costs of data centers.

[0003] Existing distributed data center resource scheduling methods can be roughly divided into three categories. The first one aims to reduce the power consumption of servers running in a single data center, the second one aims to reduce the power consumption of the cooling system in a single data center, and the third one aims to save electricity costs between data centers.

[0004] However, a single server scheduling approach is less efficient in reducing overall operating costs. Summary of the invention

[0005] The present application provides a load balancing method, apparatus, device and medium for reducing the overall load operating cost of a distributed data center, so as to achieve the effect of optimizing resource allocation and reducing unnecessary energy consumption.

[0006] In a first aspect, the present application provides a load balancing method, which is applied to a distributed data center, where a plurality of servers are deployed, and the method comprises:

[0007] Receive service requests sent by users;

[0008] linearly processing the pre-acquired power cost model according to the service request to obtain a target power cost model, wherein the service request includes a processing speed requirement for the service request and an arrival speed requirement for the service request, wherein the power cost model is a model for determining power cost obtained based on power information of each server and operating temperature information of a corresponding area;

[0009] The target electricity cost model is minimised to obtain the number of servers deployed in different areas.

[0010] Optionally, before linearly processing the pre-acquired power cost model according to the service request to obtain the target power cost model, the method includes:

[0011] Obtaining server power cost information and cooling system power cost information based on power information of each server and operating temperature information of an area corresponding to each server;

[0012] The power cost model is determined according to the server power cost information and the cooling system power cost information.

[0013] Optionally, before acquiring the server power cost information and the cooling system power cost information according to the power information of each server and the operating temperature information of the area corresponding to each server, the method further includes:

[0014] Acquire power information of each server in the distributed data center, wherein the power information of each server includes power information of the server in an idle state, power information in a peak state, and CPU utilization information;

[0015] According to the temperature detection device arranged in the area corresponding to each server, the operating temperature information of the area corresponding to each server is obtained.

[0016] Optionally, obtaining server power cost information and cooling system power cost information according to power information of each server and operating temperature information of an area corresponding to each server includes:

[0017] Obtaining electricity price information of the distributed data center at various times;

[0018] Acquiring server power cost information according to the power price information and power information of each server;

[0019] The power cost information of the cooling system is obtained according to the power cost information of the servers and the operating temperature information of the areas corresponding to the servers.

[0020] Optionally, obtaining power cost information of a cooling system according to the power cost information of the server and the operating temperature information of the area corresponding to each server includes:

[0021] Acquire temperature information of a refrigeration unit in a cooling system of the distributed data center;

[0022] Determining a cooling performance coefficient of the cooling system according to temperature information of a refrigeration unit in the cooling system and operating temperature information of an area corresponding to each server;

[0023] The power cost information of the cooling system is calculated based on the server power cost information and the cooling performance coefficient.

[0024] Optionally, linear processing is performed on the pre-acquired power cost model according to the service request to obtain a target power cost model, including:

[0025] Obtain the total number of servers, service request processing speed, service request arrival rate, and service request processing probability of the distributed data center;

[0026] Determining a delay requirement for processing the service request according to the processing speed requirement, the arrival speed requirement, the total number of servers, and the service request processing probability in the service request;

[0027] Determining constraint information of the power cost model according to the delay requirement;

[0028] According to the constraint information, the power cost model is linearly processed to obtain a target power cost model.

[0029] Optionally, the method further comprises:

[0030] Based on the number of server deployments in different regions, the server deployment in each region is scheduled.

[0031] In a second aspect, an embodiment of the present application provides a load balancing device, including:

[0032] A service request receiving module, used to receive service requests sent by users;

[0033] a target power cost model obtaining module, configured to perform linear processing on a pre-acquired power cost model according to the service request to obtain a target power cost model, wherein the service request information includes a processing speed requirement for the service request and an arrival speed requirement for the service request, wherein the power cost model is a model for determining power cost obtained based on power information of each server and operating temperature information of a corresponding area;

[0034] The server deployment quantity obtaining module is used to solve the minimization problem of the target power cost model to obtain the server deployment quantity in different areas.

[0035] In a third aspect, the present application provides a device, including: a memory, a processor;

[0036] The memory stores computer-executable instructions;

[0037] The processor executes the computer-executable instructions stored in the memory, so that the processor executes the above first aspect and / or various possible implementations of the first aspect.

[0038] In a fourth aspect, the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the above first aspect and / or various possible implementations of the first aspect.

[0039] The load balancing method, apparatus, device and medium provided in the embodiments of the present application receive a service request sent by a user; linearly process a pre-acquired power cost model according to the service request to obtain a target power cost model, wherein the service request includes a processing speed requirement for the service request and an arrival speed requirement for the service request, wherein the power cost model is a model for determining the power cost based on the power information of each server and the operating temperature information of the corresponding area; solve the minimization problem of the target power cost model to obtain a means of obtaining the number of server deployments in different areas, and evenly distribute the power costs of the servers in the distributed data center and the power costs of the cooling system, thereby reducing the overall power cost of the distributed data center. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0041] Figure 1 The overall architecture diagram of the load balancing method provided in this application;

[0042] Figure 2 The following is a flow chart of the first embodiment of the load balancing method provided in this application;

[0043] Figure 3 This is a flow chart of the second embodiment of the load balancing method provided by the present application;

[0044] Figure 4 This is a flowchart of the third embodiment of the load balancing method provided by the present application;

[0045] Figure 5 This is a flowchart of the fourth embodiment of the load balancing method provided by the present application;

[0046] Figure 6 This is a flowchart of the fifth embodiment of the load balancing method provided by the present application;

[0047] Figure 7 This is a flowchart of the sixth embodiment of the load balancing method provided by the present application;

[0048] Figure 8 A schematic diagram of the structure of the load device provided in this application;

[0049] Fig. 9 A schematic diagram of the structure of the electronic device provided in this application.

[0050] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0051] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0052] The existing distributed data center resource scheduling methods can be roughly divided into three categories: the first is to reduce the power consumption of servers running in a single data center, the second is to reduce the power consumption of the cooling system in a single data center, and the third is to save the power cost between data centers. It is often difficult to minimize the overall operating cost of the operator's distributed data center by optimizing a single influencing factor, and the mutual influence of various influencing factors should be considered at the same time.

[0053] In response to the above technical problems, the inventors designed a comprehensive electricity cost modeling method from the perspective of the dynamic changes in the operating environment of the data center and the electricity price. By obtaining the power information and electricity price information of each server at different times, the electricity cost of the server is accurately calculated. At the same time, combined with the operating temperature information of the area where each server is located, the parameter of the cooling performance coefficient is introduced. According to the output temperature of the refrigeration unit and the operating temperature of the server, the energy efficiency performance of the cooling system is dynamically evaluated, thereby more scientifically quantifying the electricity cost of the cooling system, optimizing the overall energy consumption structure, and providing accurate decision-making support for reducing the operation and maintenance costs of distributed data centers.

[0054] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0055] Figure 1 The overall architecture diagram of the load balancing method provided in this application. Figure 1 As shown, the overall architecture 10 of the load balancing method includes a data collection module 101 , an electricity cost calculation module 102 , a load balancing decision module 103 and a control and execution module 104 .

[0056] Among them, the data acquisition module 101 is responsible for collecting basic operating data in the distributed data center to provide support for subsequent calculations, such as server power information, operating temperature information, refrigeration unit temperature information and electricity price information. The power cost calculation module 102 is used to calculate the power cost of the server and the cooling system based on the information provided by the data acquisition module. Among them, the server power cost calculation is to calculate the server power cost in different areas by combining the server power information with the electricity price information; the cooling system power cost calculation is based on the refrigeration unit temperature, operating temperature and server power cost, to determine the cooling system cooling performance coefficient, and calculate the power cost of the cooling system. The load balancing decision module 103 is used to optimize the server load distribution and achieve energy saving based on the power cost information and cooling system performance. The control and execution module 104 is used to apply the optimized scheduling scheme to the actual data center operation to achieve load balancing.

[0057] Figure 2 This is a flow chart of the first embodiment of the load balancing method provided by this application. Figure 2 As shown, the method includes:

[0058] S201: Receive a service request sent by a user.

[0059] Among them, service requests refer to specific needs from clients (such as individual users or corporate users), such as web browsing, file storage, data calculation and other operations. Service requests specifically include requirements for the processing speed of service requests and the arrival speed of service requests. Specifically, the processing speed requirement of service requests refers to the time that users expect to complete tasks, such as real-time requests with low latency requirements; the arrival speed requirement of service requests can refer to the frequency or number of requests per unit time, reflecting the load pressure of the server.

[0060] Exemplarily, the front-end server can receive user requests through the network interface, cache and queue the user requests in chronological order. Specifically, in the case of a large number of concurrent requests, the user requests can be initially distributed to the relevant servers using polling or hash distribution methods to prepare for the next round of processing.

[0061] S202: Perform linear processing on the pre-acquired power cost model according to the service request to obtain a target power cost model.

[0062] The power cost model is a mathematical model that combines the losses required for server operation, the power consumption of the cooling system, and the real-time electricity prices in the deployment area to calculate the actual power cost of running each server. Specifically, the power cost model is a model for determining power costs based on the power information of each server and the operating temperature information of the corresponding area. Linear processing is to convert complex power cost variables into simple linear functions through data models for optimization and calculation.

[0063] Specifically, based on the specific needs of user service requests (processing speed and request arrival speed), combined with the power consumption characteristics and cooling system performance of the area where each server is located, the existing electricity cost model is adjusted to simplify the complex problem into a linear form that is easy to optimize, thereby generating an optimized target electricity cost model for the next scheduling decision.

[0064] S203, solving the minimization problem of the target electricity cost model to obtain the number of servers deployed in different areas.

[0065] Among them, solving the minimization problem refers to using optimization algorithms to find the optimal solution in the model to minimize the total electricity cost. The number of servers deployed in different regions is the number of servers allocated to each data center to meet service requests, which is usually optimized based on the difference in electricity prices and cooling efficiency between regions.

[0066] In this step, mathematical optimization methods are applied to solve the target power cost model and calculate the number of servers that need to run in the data center to minimize the power cost of the entire distributed data center, thereby meeting user needs and achieving efficient use of resources.

[0067] The load balancing method provided in the embodiment of the present application combines user request characteristics, power cost model and server deployment strategy, comprehensively considers multi-dimensional factors such as server power, ambient temperature and electricity price, and dynamically optimizes resource allocation. By minimizing the power cost model, it not only improves the operating efficiency of the data center, but also effectively reduces energy consumption, providing technical support for the sustainable development of green data centers.

[0068] Figure 3 This is a flow chart of the second embodiment of the load balancing method provided by this application. Figure 3 As shown, in Figure 2 Based on the embodiment, before linearly processing the pre-acquired power cost model according to the service request to obtain the target power cost model, the method further includes:

[0069] S301. Obtain server power cost information and cooling system power cost information according to power information of each server and operating temperature information of an area corresponding to each server.

[0070] Among them, power information includes the functional consumption of each server under different load conditions. Server power cost information is the direct energy consumption cost of the server calculated based on the server operating power and local electricity prices. Cooling system power cost information is determined by the heat generated when the server is running, which needs to be processed by the cooling system power consumption and is closely related to the regional temperature.

[0071] Furthermore, power information of each server in the distributed data center is obtained, wherein the power information of each server includes power information of the server in an idle state, power information in a peak state, and CPU utilization information.

[0072] Specifically, power information refers to the basic power consumption of the server when it is not processing tasks, power information in peak state refers to the maximum power consumption of the server when it is running at full load, and CPU utilization refers to the load ratio of the server CPU at the current moment, which affects the dynamic adjustment of the server's operating power.

[0073] Operating temperature information refers to data related to the ambient temperature of the area where the server is located, such as the temperature distribution of cold areas and hot areas, reflecting the changes in regional temperature under different loads.

[0074] Furthermore, operating temperature information of the area corresponding to each server is obtained according to the temperature detection device arranged in the area corresponding to each server.

[0075] The temperature detection device is installed in different areas of the data center to monitor the ambient temperature of the area in real time. The regional operating temperature information refers to the real-time temperature data in a specific area obtained by the temperature detection device, which is used to evaluate the heat load of the server operating environment in the area.

[0076] Specifically, in a distributed data center, each server is located in a specific physical area, and the temperature in different areas may be different due to the external environment or equipment heat release. By installing temperature detection devices in these areas, real-time operating temperature information of each area can be obtained. This temperature information provides a key reference for subsequent calculations of server power consumption (such as cooling requirements), and helps to evaluate the operating efficiency of the cooling system and the optimal configuration of distributed loads.

[0077] S302: Determine a power cost model according to the power cost information of the server and the power cost information of the cooling system.

[0078] Among them, the electricity cost model refers to a mathematical model that comprehensively considers multiple factors such as server power, regional temperature, electricity price, etc., and is used to characterize the total electricity consumption cost of distributed data centers.

[0079] The load balancing method provided in the embodiment of the present application can establish an electricity cost model by comprehensively analyzing the power cost information of the server and the power cost information of the cooling system. The model combines factors such as the operating power of the server, CPU utilization, regional electricity prices, and the operating temperature of the cooling system to describe the overall power cost of the data center under different load and environmental conditions. This model provides an important basis for subsequent load balancing and resource optimization, so that load distribution can not only reduce server power consumption, but also reduce the energy consumption of the cooling system, thereby optimizing the overall power cost of the distributed data center.

[0080] Figure 4 This is a flow chart of the third embodiment of the load balancing method provided in this application. Figure 4 As shown, in Figure 3 Based on the embodiment, according to the power information of each server and the operating temperature information of the area corresponding to each server, the server power cost information and the cooling system power cost information are obtained, including:

[0081] S401. Obtain electricity price information of the distributed data center at each time.

[0082] In this step, it is assumed that a distributed data center consists of N data centers, and the electricity price of data center i (1≤i≤N) at time t is P c (t) is the electricity cost information. The goal of this step is to obtain time-related electricity price information from the relevant market and provide basic data for subsequent electricity cost calculation.

[0083] S402: Obtain server power cost information based on power price information and power information of each server.

[0084] In this step, the average power of each server in idle state is set to P vavg , the average power of each server in peak state is P peak , the CPU utilization of the server is η i , then the power consumption of the running server in distributed data center i is for:

[0085]

[0086] In the cooling system of the distributed data center, the hot air generated by the running servers is sucked in by the liquid cooling system, and the cold air is sent into the computer room through the floor vents. Therefore, the distributed data center can be divided into cold zone, hot zone and warm zone according to the cold and hot air flow exchange mode. Therefore, the power cost of the running server of the distributed data center i is:

[0087]

[0088] Among them, P vavg It refers to the basic power consumption of the server when it is not processing tasks, and the power information P in the peak state peak It refers to the maximum power consumption of the server when it is running at full load, and the CPU utilization η i It refers to the load ratio of the server CPU at the current moment, which affects the dynamic adjustment of the server's operating power.

[0089] S403: Obtain power cost information of the cooling system according to the power cost information of the servers and the operating temperature information of the areas corresponding to the servers.

[0090] In this step, by obtaining the power cost information of the server, the regional operating temperature and the cooling system refrigeration performance coefficient, combined with the electricity price data, the power cost information of the cooling system can be calculated.

[0091] The load balancing method provided in the embodiment of the present application obtains the electricity price information of the distributed data center at each time and calculates the power cost of the server in combination with the power information of each server. Subsequently, based on the power cost information of the server and the operating temperature information of the corresponding area, the power cost of the cooling system is further obtained to achieve an accurate assessment of the overall power cost of the data center.

[0092] Figure 5 This is a flow chart of the fourth embodiment of the load balancing method provided by this application. Figure 5 As shown, in Figure 4 Based on the embodiment, the power cost information of the cooling system is obtained according to the power cost information of the server and the operating temperature information of the area corresponding to each server, including:

[0093] S501. Obtain temperature information of a refrigeration unit in a cooling system of a distributed data center.

[0094] The temperature information of the refrigeration unit is the key input for calculating the coefficient of performance (COP). It can be determined by the following formula:

[0095]

[0096]

[0097] in, is the output temperature of the refrigeration unit, is the maximum output temperature of the running server, T out is the outside temperature, T Max It is the highest safe temperature in the computer room. The percentage of the operating load of zone j in distributed data center i is After obtaining this temperature information, it can provide a basis for the next step of cooling performance coefficient calculation.

[0098] S502: Determine a cooling performance coefficient of the cooling system according to temperature information of a refrigeration unit in the cooling system and operating temperature information of an area corresponding to each server.

[0099] Among them, the cooling performance coefficient COP is related to the output temperature of the refrigeration unit There is a positive correlation between the COP and the refrigeration unit output temperature (the highest output of the three temperature zones) as a parameter, which can be expressed as:

[0100]

[0101] Through the above formula, the cooling efficiency of the cooling system under different cooling temperature conditions can be obtained, among which the cooling performance coefficient is a key parameter for optimizing the power consumption allocation of the cooling system.

[0102] S503: Calculate the power cost information of the cooling system according to the power cost information of the server and the cooling performance coefficient.

[0103] Combining the above server power cost calculation formula and cooling performance coefficient, the power cost calculation formula for the cooling system can be obtained:

[0104]

[0105] By combining the server power cost information and the cooling performance coefficient through the above calculation method, the power cost information ECCS of the cooling system can be obtained.

[0106] The load balancing method provided in the embodiment of the present application obtains the temperature information of the refrigeration unit, combines the operating temperature of the server area, calculates the cooling performance coefficient of the cooling system, and calculates the power cost information of the cooling system based on the server power cost information and the cooling performance coefficient. This process combines the server power consumption with the cooling efficiency to optimize the power cost evaluation model of the distributed data center.

[0107] Figure 6 This is a flow chart of the fifth embodiment of the load balancing method provided by this application. Figure 6 As shown, including:

[0108] S601. Obtain the total number of servers, service request processing speed, service request arrival rate, and service request processing probability of the distributed data center.

[0109] In this step, it is necessary to obtain various data in the distributed data center, including: the total number of servers N, which is used to calculate the total load and allocate workload; the service request processing speed, which is used to determine the performance requirements that the server needs to meet; the service request arrival rate, which is used to measure the load pressure of the data center; the service request processing probability, which is used to reflect whether the system can complete the request on time. These parameters provide basic input for the subsequent calculation of latency requirements and power consumption allocation.

[0110] S602: Determine the delay requirement for processing the service request according to the processing speed requirement, arrival speed requirement, total number of servers and service request processing probability in the service request.

[0111] This step mainly combines the processing capacity and load pressure of the server to determine the speed at which the distributed data center processes data, that is, the latency requirement. Among them, the key influencing factors include processing speed requirements, arrival speed requirements, and load distribution impact. Specifically, the higher the load, the higher the server performance required, but the latency may increase; the faster the request arrival speed, the stricter the latency requirement; reasonable distribution of load and maximum processing capacity can ensure that the latency requirements are met.

[0112] For example, within a time interval, distributed data center I receives λ service requests, which needs to meet the following conditions:

[0113]

[0114] At the same time, when the number of running servers is n, the service request processing rate is v, the service request arrival rate is λ, and the probability of a service request in the processing queue is P c =1, the average delay is:

[0115]

[0116] In order to meet the latency requirements of distributed data center i, d i ≤D i .

[0117] S603: Determine constraint information of the power cost model according to the delay requirement.

[0118] Among them, the latency requirement imposes constraints on the power cost model. For example, temperature impact: the highest security issue of distributed data centers T Max Determines the operating temperature of the server Lower Corresponding to higher COP i , Power consumption relationship: Assumption of idle power consumption and peak power consumption P peak =2P vavg, which determines the basic structure of the power consumption model.,Processing margin: Ten percent of processing power is reserved as redundancy, providing a buffer for latency and energy efficiency.

[0119] S604: Perform linear processing on the power cost model according to the constraint information to obtain a target power cost model.

[0120] In this step, the complex nonlinear problem is simplified, that is, the objective function is converted into a convex function for easy solution. Specifically, the COP is represented by linear segmentation, and the one-dimensional linear regression method is used to convert the COP into a linear segmentation function of the load percentage. This method facilitates the processing of nonlinear models into segmented linear models. Then, the linearization formula is used to form an easy-to-calculate electricity cost model.

[0121] For example, based on the above, we can get the energy consumption estimate of the distributed data center:

[0122]

[0123] There may be N servers in a distributed data center, so when calculating the energy consumption of the entire distributed data center, it is necessary to calculate the sum of all servers; i It is a weight factor, which is related to the resource allocation or load ratio of each server.

[0124] in addition, is the correction factor, where 1 is the basic power consumption, b k is the temperature correlation coefficient, which is used to describe the impact of different temperature zones on power consumption. Represents the service of i in distributed data

[0125] The correction of energy efficiency by the load ratio of the device, where m i Indicates the current server load, M i Indicates the total capacity or maximum processing capability of the current server.

[0126] The embodiment of the present application establishes an optimization model of energy efficiency and power consumption through the real-time operating parameters of the distributed data center (such as load, power consumption and temperature), and uses the linear piecewise function representation of COP to convert the formula into a convex optimization temperature to simplify the solution process. In combination with the delay requirements and temperature limits, the power model and load correction factor are adjusted to ensure the balance between energy consumption and performance. Finally, through linearization and optimization technology, an electricity cost model that meets the delay requirements, temperature limits and the lowest energy consumption is obtained.

[0127] Figure 7 This is a flow chart of Example 6 of the load balancing method provided in this application. Figure 7 As shown, in Figures 2 to 6 Based on the embodiment, the method further includes:

[0128] S701. Schedule server deployment in each region based on the number of server deployments in different regions.

[0129] Aiming at the large-scale server deployment in distributed data centers, a method based on stepwise optimization is proposed to effectively solve the problem of minimizing the objective function. First, an efficient solver (Gurobi Optimizer, Gurobi for short) is used to calculate the number of servers deployed in a single data center, m. i An approximate solution is obtained to obtain a preliminary answer. Then, the cooling efficiency COP is calculated by a one-dimensional linear function. i Calculation is performed to further derive the maximum output temperature of the server.

[0130] Based on this, the number of servers deployed in different temperature zones is used as the key basis to dynamically schedule servers in each region to achieve reasonable resource allocation and energy efficiency optimization. This method not only improves the efficiency of solving large-scale problems, but also ensures the coordinated optimization effect of cooling and deployment in different temperature zones.

[0131] In the load balancing method provided in the embodiment of the present application, the computing capacity of each area is determined by the number of servers in different areas, and the overall energy consumption of the data center is reduced by optimizing the load distribution between areas while meeting the service request latency requirements.

[0132] Figure 8 A schematic diagram of the structure of the load balancing device provided in this application, such as Figure 8 As shown, the load balancing device 80 provided in this embodiment includes:

[0133] The service request receiving module 801 is used to receive a service request sent by a user;

[0134] The target power cost model obtaining module 802 is used to perform linear processing on the pre-acquired power cost model according to the service request to obtain the target power cost model, wherein the service request information includes a processing speed requirement for the service request and an arrival speed requirement for the service request, wherein the power cost model is a model for determining the power cost based on the power information of each server and the operating temperature information of the corresponding area;

[0135] The server deployment quantity obtaining module 803 is used to solve the minimization problem of the target power cost model to obtain the server deployment quantity in different areas.

[0136] Furthermore, the service request receiving module 801 is also specifically used for:

[0137] Obtaining server power cost information and cooling system power cost information based on power information of each server and operating temperature information of an area corresponding to each server;

[0138] An electricity cost model is determined based on the server electricity cost information and the cooling system electricity cost information.

[0139] Furthermore, the service request receiving module 801 is also specifically used for:

[0140] Obtaining power information of each server in the distributed data center, wherein the power information of each server includes power information of the server in an idle state, power information in a peak state, and CPU utilization information;

[0141] According to the temperature detection device arranged in the area corresponding to each server, the operating temperature information of the area corresponding to each server is obtained.

[0142] Furthermore, the service request receiving module 801 is also specifically used for:

[0143] Obtain electricity price information of distributed data centers at all times;

[0144] Obtain server power cost information based on power price information and power information of each server;

[0145] The power cost information of the cooling system is obtained according to the power cost information of the servers and the operating temperature information of the areas corresponding to the servers.

[0146] Furthermore, the service request receiving module 801 is also specifically used for:

[0147] Obtaining temperature information of a refrigeration unit in a cooling system of a distributed data center;

[0148] Determine the cooling performance coefficient of the cooling system according to the temperature information of the refrigeration unit in the cooling system and the operating temperature information of the area corresponding to each server;

[0149] The power cost information of the cooling system is calculated based on the server power cost information and the cooling performance coefficient.

[0150] Furthermore, the load balancing device 80 is also specifically used for:

[0151] Obtain the total number of servers, service request processing speed, service request arrival rate, and service request processing probability of the distributed data center;

[0152] Determine the delay requirement for processing the service request based on the processing speed requirement, arrival speed requirement, total number of servers and service request processing probability in the service request;

[0153] Determine the constraint information of the power cost model according to the delay requirements;

[0154] According to the constraint information, the power cost model is linearly processed to obtain the target power cost model.

[0155] Furthermore, the load balancing device 80 is also specifically used for:

[0156] Based on the number of server deployments in different regions, the server deployment in each region is scheduled.

[0157] The load balancing device provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and this embodiment will not be described in detail here.

[0158] Fig. 9 This is a schematic diagram of the structure of the electronic device provided in this application. Fig. 9 As shown, the electronic device 90 provided in this embodiment includes: at least one processor 901 and a memory 902. Optionally, the device 90 further includes a communication component 903. The processor 901, the memory 902 and the communication component 903 are connected via a bus 904.

[0159] In a specific implementation process, at least one processor 901 executes the computer execution instructions stored in the memory 902, so that at least one processor 901 executes the above method.

[0160] The specific implementation process of the processor 901 can be found in the above method embodiment, and its implementation principle and technical effect are similar, so this embodiment will not be repeated here.

[0161] In the above embodiments, it should be understood that the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the invention may be directly implemented as being executed by a hardware processor, or may be executed by a combination of hardware and software modules in the processor.

[0162] The memory may include a high-speed memory (Random Access Memory, RAM), and may also include a non-volatile memory (Non-volatile Memory, NVM), such as at least one disk memory.

[0163] The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, the bus in the drawings of the present application is not limited to only one bus or one type of bus.

[0164] The present application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the above method is implemented.

[0165] The above-mentioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium that can be accessed by a general or special-purpose computer.

[0166] An exemplary readable storage medium is coupled to a processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (Application Specific Integrated Circuits, referred to as: ASIC). Of course, the processor and the readable storage medium can also exist in the device as discrete components.

[0167] The division of units is only a logical function division, and there may be other divisions in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interface, device or unit, which can be electrical, mechanical or other forms.

[0168] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0169] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0170] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0171] Those skilled in the art can understand that all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, disk or optical disk and other media that can store program codes.

[0172] Finally, it should be noted that those skilled in the art will readily conceive of other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses or adaptations of the present invention, which follow the general principles of the present invention and include common knowledge or customary technical means in the art not disclosed by the present invention, are not limited to the precise structure described above and shown in the drawings, and may be modified and changed in various ways without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims

1. A load balancing method, characterized in that: Applied to a distributed data center, where a plurality of servers are deployed, the method comprises: Receiving a service request sent by a user, wherein the service request includes a processing speed requirement for the service request and an arrival speed requirement for the service request; According to the service request, a pre-acquired power cost model is linearly processed to obtain a target power cost model, wherein the power cost model is a model for determining power cost based on power information of each server and operating temperature information of a corresponding area; The target electricity cost model is minimised to obtain the number of servers deployed in different areas.

2. The method according to claim 1, characterized in that Before linearly processing the pre-acquired power cost model according to the service request to obtain the target power cost model, the method includes: Obtaining server power cost information and cooling system power cost information based on power information of each server and operating temperature information of an area corresponding to each server; The power cost model is determined according to the server power cost information and the cooling system power cost information.

3. The method according to claim 2, characterized in that Before acquiring the server power cost information and the cooling system power cost information according to the power information of each server and the operating temperature information of the area corresponding to each server, the method further includes: Acquire power information of each server in the distributed data center, wherein the power information of each server includes power information of the server in an idle state, power information in a peak state, and CPU utilization information; According to the temperature detection device arranged in the area corresponding to each server, the operating temperature information of the area corresponding to each server is obtained.

4. The method according to claim 3, characterized in that The obtaining of server power cost information and cooling system power cost information according to the power information of each server and the operating temperature information of the area corresponding to each server includes: Obtaining electricity price information of the distributed data center at various times; Acquiring server power cost information according to the power price information and power information of each server; The power cost information of the cooling system is obtained according to the power cost information of the servers and the operating temperature information of the areas corresponding to the servers.

5. The method according to claim 4, characterized in that The obtaining the power cost information of the cooling system according to the power cost information of the server and the operating temperature information of the area corresponding to each server includes: Acquiring temperature information of a refrigeration unit in a cooling system of the distributed data center; Determining a cooling performance coefficient of the cooling system according to temperature information of a refrigeration unit in the cooling system and operating temperature information of an area corresponding to each server; The power cost information of the cooling system is calculated based on the server power cost information and the cooling performance coefficient.

6. The method according to any one of claims 1 to 5, characterized in that: The linear processing of the pre-acquired power cost model according to the service request to obtain the target power cost model includes: Obtain the total number of servers, service request processing speed, service request arrival rate, and service request processing probability of the distributed data center; Determining a delay requirement for processing the service request according to the processing speed requirement, the arrival speed requirement, the total number of servers, and the service request processing probability in the service request; Determining constraint information of the power cost model according to the delay requirement; According to the constraint information, the power cost model is linearly processed to obtain a target power cost model.

7. The method according to any one of claims 1 to 5, characterized in that: The method further comprises: Based on the number of server deployments in different regions, the server deployment in each region is scheduled.

8. A load balancing device, characterized in that: include: A service request receiving module, used to receive service requests sent by users; a target power cost model obtaining module, configured to perform linear processing on a pre-acquired power cost model according to the service request to obtain a target power cost model, wherein the service request information includes a processing speed requirement for the service request and an arrival speed requirement for the service request, wherein the power cost model is a model for determining power cost obtained based on power information of each server and operating temperature information of a corresponding area; The server deployment quantity obtaining module is used to solve the minimization problem of the target power cost model to obtain the server deployment quantity in different areas.

9. An electronic device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 7 when executed by a processor.

Images