Electric power and thermal power cooperative scheduling control method based on server load rate

Abstract

The invention provides an electric power and thermal power cooperative scheduling control method based on a server load rate, and relates to the technical field of data center server cluster cooperative control, and the method comprises the steps: 1, measuring the unit calculation power energy consumption of a sample server under different calculation power conformity; 2, determining a corresponding optimal load rate range according to a unit calculation power energy consumption measurement result; 3, according to the optimal load rate range, the energy-saving refrigeration temperature of the sample server is calculated; according to the method, the optimal corresponding load rate range of the unit computing power energy consumption of the server is determined by obtaining the unit computing power energy consumption of the server under different computing power loads, so that the overall energy consumption of the data center is the lowest when the same computing power task is calculated, and the calculation efficiency of the data center is improved. And correspondingly, in the range, energy-saving refrigeration temperature calculation is performed, and the refrigeration temperature is regulated and controlled, so that the overall energy consumption of the data center machine room is reduced, and the economic benefit is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of coordinated control of data center server clusters, and in particular to an electric power and heat coordinated dispatching control method based on server load rate. Background Art

[0002] Data centers are facilities that are specifically designed to house computer systems and related communications and storage equipment. These facilities typically contain a large number of servers, network equipment, and storage systems that are used to process, store, and transmit large amounts of data. Data centers are key components of modern information technology infrastructure, providing support for a variety of online services, cloud computing, big data analysis, and other IT applications, while data center server clusters refer to a group of collaborative servers deployed in a data center that work together to provide computing resources and services. Server clusters can improve the service availability, scalability, and performance of data centers.

[0003] However, a data center has a large number of servers, and the servers generate a large amount of heat during operation, which not only increases the cooling cost of the data center, but may also cause the servers to overheat, affecting their performance and lifespan. Existing data centers usually focus on optimizing the cooling system to reduce the temperature of the data center in order to protect the performance of the servers, but do not take the cost of electricity use into consideration. When the load rate of the server changes, the heat dissipation and the server core temperature will change, which will cause the server to not operate at an appropriate temperature. At the same time, too low a cooling temperature will cause unnecessary additional energy consumption in the cooling system. Therefore, the present invention proposes an electric-thermal coordinated scheduling control method based on the server load rate to solve the problems existing in the prior art. Summary of the invention

[0004] In view of the above problems, the purpose of the present invention is to propose an electric power and thermal coordinated scheduling control method based on server load rate, which has the advantage of reducing the overall energy consumption of data center servers and can solve the problems existing in the prior art.

[0005] To achieve the purpose of the present invention, the present invention is implemented by the following technical solution: a power and heat coordinated dispatch control method based on server load rate, comprising the following steps:

[0006] Step 1: Select sample servers from the server cluster and measure the unit computing power energy consumption of the sample servers under different computing power loads;

[0007] Step 2: Determine the corresponding optimal load rate range based on the unit computing power energy consumption measurement results of the sample server obtained in step 1;

[0008] Step 3: According to the optimal load rate range in step 2, the energy-saving cooling temperature of the sample server at different load rates is measured and calculated to obtain the minimum cooling temperature range corresponding to the load rate range;

[0009] Step 4: Schedule and allocate the computing load among the server clusters according to the optimal load rate range, and set the cooling temperature according to the allocated computing load.

[0010] A further improvement is that in step 1, the specific steps of measurement are:

[0011] S1: Deploy energy consumption and performance indicator monitoring software on the sample server;

[0012] S2: Perform energy consumption tests on the sample server under different computing load levels;

[0013] S3: Calculate the energy consumption per unit computing power of the sample server based on the energy consumption test results.

[0014] A further improvement is that the range of the different computing power loads is 0% to 100%, and within this range, several groups of computing power load values ​​are set for testing.

[0015] A further improvement is that in step 2, based on the optimal load rate range, an optimal load rate within the range is determined, and the optimal load rate is the target load rate of the server.

[0016] A further improvement is that in step 3, the specific steps of measurement and calculation are:

[0017] SS1: Collect the heat output data of the server at different load rates, and then collect the efficiency data of the cooling equipment;

[0018] SS2: Based on the data collected by SS1, a digital model of heat output data and load rate, as well as a digital model of cooling efficiency, were established;

[0019] SS3: The two digital models in SS2 are coupled to obtain an energy consumption model. Based on the energy consumption model, the heat generated by the server and the energy consumption required by the cooling equipment under different load rates are calculated.

[0020] A further improvement is that in step three, a linear programming algorithm is used to minimize the energy consumption of the refrigeration system, and the refrigeration temperature is adjusted while ensuring that the server will not overheat, so as to find the refrigeration temperature with the lowest energy consumption.

[0021] A further improvement is that the condition that the server will not overheat is that the maximum operating temperature of the server is preset, and when the internal temperature of the server reaches more than 90% of the maximum operating temperature, the server is considered to be overheated.

[0022] A further improvement is that in step 4, the scheduling allocation rule is: tasks are allocated to servers with lower loads first, and the server load is brought close to the optimal load rate.

[0023] The beneficial effects of the present invention are as follows: the present invention provides an electric heat coordinated dispatching and control method based on server load rate, which measures the energy consumption of the server under different computing power loads, and obtains the unit computing power energy consumption of the server under different computing power loads, and then determines the optimal corresponding load rate range of the server unit computing power energy consumption, that is, the optimal load rate range, so that the overall energy consumption of the data center is minimized when calculating the same computing power tasks, and then according to the range, the computing power load is dispatched and allocated between the server clusters in the data center, and accordingly, within the range, the energy-saving refrigeration temperature is calculated, and the refrigeration temperature is adjusted by monitoring the changes in the server load rate, so that the overall energy consumption of the data center computer room is reduced and the economic benefits are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a schematic diagram of the steps of the present invention. DETAILED DESCRIPTION

[0025] In order to deepen the understanding of the present invention, the present invention will be further described in detail below in conjunction with examples. The examples are only used to explain the present invention and do not constitute a limitation on the protection scope of the present invention.

[0026] Embodiment 1

[0027] according to Figure 1 As shown, this embodiment proposes a power-heat coordinated dispatch control method based on server load rate, comprising the following steps:

[0028] Step 1: Select sample servers from the server cluster and measure the unit computing power energy consumption of the sample servers under different computing power loads. The specific steps of the measurement are as follows:

[0029] S1: Deploy energy consumption and performance indicator monitoring software on the sample server;

[0030] S2: Perform energy consumption tests on the sample server under different computing load levels;

[0031] S3: Calculate the unit computing power energy consumption of the sample server according to the energy consumption test results, where the range of different computing power loads is 0% to 100%. Within this range, set several groups of computing power load values ​​for testing;

[0032] Step 2: According to the unit computing power energy consumption measurement result of the sample server obtained in step 1, the corresponding optimal load rate range is determined, and according to the optimal load rate range, the optimal load rate within the range is determined, and the optimal load rate is the target load rate of the server;

[0033] Step 3: According to the optimal load rate range in step 2, measure and calculate the energy-saving cooling temperature of the sample server at different load rates to obtain the minimum cooling temperature range corresponding to the load rate range. The specific steps of measurement and calculation are as follows:

[0034] SS1: Collect the heat output data of the server at different load rates, and then collect the efficiency data of the cooling equipment;

[0035] SS2: Based on the data collected by SS1, a digital model of heat output data and load rate, as well as a digital model of cooling efficiency, were established;

[0036] SS3: The two digital models in SS2 are coupled to obtain an energy consumption model. Based on the energy consumption model, the heat generated by the server and the energy consumption required by the refrigeration equipment under different load rates are calculated. The linear programming algorithm is used to minimize the energy consumption of the refrigeration system. Under the premise of ensuring that the server will not overheat, the refrigeration temperature is adjusted to find the refrigeration temperature with the lowest energy consumption. The condition that the server will not overheat is: the maximum operating temperature of the server is preset. When the internal temperature of the server reaches more than 90% of the maximum operating temperature, the server is considered to be overheated;

[0037] Step 4: According to the optimal load rate range, the computing load between the server clusters is scheduled and allocated, and the cooling temperature is set according to the allocated computing load. In step 4, the scheduling and allocation rule is: tasks are preferentially allocated to servers with lower loads, and the server load is brought close to the optimal load rate.

[0038] Embodiment 2

[0039] according to Figure 1 As shown, this embodiment proposes a power-heat coordinated dispatch control method based on server load rate, comprising the following steps:

[0040] Step 1: Measure the unit computing power energy consumption of the server under different computing power loads. The main purpose is to measure the energy consumption of the server under different computing power loads, and obtain the unit computing power energy consumption of the server under different computing power loads, and then determine the optimal corresponding load rate of the server unit computing power energy consumption, that is, the optimal load rate. Therefore, under the optimal load rate, the energy consumption of the server is at a suitable position. Specifically, the specific steps of the measurement are:

[0041] S1: Since there are many servers in the data center, representative servers are selected as test objects in the data center, and other servers are debugged accordingly based on the test data of the representative servers. Then, energy consumption and performance indicator monitoring software is deployed on the servers. The software can monitor the real-time energy consumption data of the servers.

[0042] S2: Make the server perform energy consumption test under different computing load levels, the range of different computing loads is 0% to 100%. Within this range, several groups of computing load values ​​are set for testing. In this embodiment, five groups of computing load values ​​are set, and the values ​​are in percentage units, namely 20%, 40%, 60%, 80% and 100%, and then the test is performed;

[0043] S3: Calculate the energy consumption per unit computing power based on the energy consumption test results. The test results record the real-time energy consumption data of the server at each computing power load level, including power consumption and heat dissipation requirements.

[0044] Step 2: After the test, divide the total energy consumption at each load level by the corresponding computing power output to obtain the unit computing power energy consumption. Compare the unit computing power energy consumption at different load levels to determine the load range with the highest energy efficiency. Within this load range, the unit computing power energy consumption of the server reaches the lowest, that is, the energy consumed per unit computing power is the least. Therefore, according to the unit computing power energy consumption of the server in step 1, determine the corresponding optimal load rate range. Then, with the cooperation of steps 1 and 2, by determining the load rate range with the lowest energy consumption, the computing power resources of the data center can be more reasonably allocated, resource waste can be avoided, and energy utilization efficiency can be improved. By determining the optimal load rate range, the data center can reduce unnecessary energy consumption by adjusting the working status of the server, thereby reducing operating costs such as electricity bills.

[0045] Step 3: After obtaining the optimal load range, it means that in the daily work of the data center, controlling the server load within this optimal load range can avoid unnecessary energy consumption as much as possible. At the same time, in order to better control and reduce energy consumption, within the optimal load range, the energy-saving cooling temperature of the sample server at different load rates is measured and calculated. The measurement and calculation conditions are also within the optimal load range. Several groups of load values ​​are set, with at least three groups, to obtain the optimal cooling temperature range corresponding to the load rate range. The specific steps of measurement and calculation are:

[0046] SS1: Collect the heat output data of the server under different load rates. The heat output data is mainly used to analyze the heat output of the server under different load rates, and then collect the efficiency data of the refrigeration equipment (refrigeration system). The efficiency data is the performance data of the refrigeration equipment itself, including cooling capacity, power, etc.

[0047] SS2: Based on the data collected by SS1, digital models of heat output data and load rate are established respectively. The specific way to establish the digital model is as follows: a linear model is used to construct a linear relationship between heat output data and load rate. The formula is Q=a1+b1*L, where Q is heat output, L is load rate, a1 and b1 are constants, a1 is the basic heat output of the server, and b1 is the expected increase in heat output Q when the load rate L increases by one unit. It describes the sensitivity of the server heat output to the change of load rate. The digital model can reflect the relationship between it and the server heat output under different load rates. The refrigeration efficiency model can help evaluate the performance of the current refrigeration system of the data center, and includes the refrigeration coefficient and the energy efficiency ratio. A large refrigeration performance coefficient indicates that the energy utilization efficiency of the refrigeration system (refrigerator) is high, and the larger the energy efficiency ratio, the more electric energy is saved. In this embodiment, the energy efficiency ratio is used as the model data, and the model expression is obtained as follows: EER=a2+b2*P, where EER is the energy efficiency ratio, a2 and b2 are constants, a2 represents the energy efficiency ratio EER when the input power P of the refrigeration system is zero, and b2 is the ratio of the energy efficiency ratio to the input power;

[0048] SS3: The two digital models in SS2 are coupled to obtain an energy consumption model, that is, the heat output Q is used as the input of the cooling efficiency model, which is specifically expressed as: E=k*L+(a2+b2*Q), where E is the total energy consumption, which is the server energy consumption and the cooling system energy consumption, k is the energy consumption of the server unit load rate, and L is the load rate. Based on the energy consumption model, the heat generated by the server and the energy consumption required by the cooling equipment are calculated under different load rates. The linear programming algorithm is used to minimize the energy consumption of the cooling system. Under the premise of ensuring that the server will not overheat, the cooling temperature is adjusted to find the cooling temperature with the lowest energy consumption. The condition that the server will not overheat is: the preset maximum operating temperature of the server. When the internal temperature of the server reaches more than 90% of the maximum operating temperature, the server is considered to be overheated. Therefore, the overall energy consumption is reduced under the premise of ensuring the working efficiency of the server.

[0049] Step 4: According to the optimal load rate range, the computing load between server clusters is scheduled and allocated, and the cooling temperature is set according to the allocated computing load. The scheduling and allocation rules are: tasks are assigned to servers with lower loads first, and the server load is brought close to the optimal load rate.

[0050] Embodiment 3

[0051] This embodiment is the same as the second embodiment, but the difference between this embodiment and the second embodiment is that the energy-saving refrigeration temperature is measured and calculated. The actual test is carried out. After the actual test is completed, it is found that according to the measurement and calculation results, the server can meet the usage requirements, which means that there is no problem with the measurement and calculation results. If it is applied to other servers, if it cannot meet the usage requirements, the performance of the server fluctuates, and then corresponding adjustments need to be made.

[0052] Embodiment 4

[0053] This embodiment is the same as the third embodiment, but the difference between this embodiment and the third embodiment is that since a representative server is selected for debugging and the debugging data is applied to other servers, the performance of each server is different. Although some of them are similar, there are still certain differences. Then, in subsequent use, when it is found that the energy consumption and load rate of some servers fluctuate greatly, separate testing and debugging are performed on the server.

[0054] Therefore, the present invention involves two methods, one of which is an optimized scheduling method for computing loads between data center server clusters, and the other is a cooling temperature control method that changes with the load rate of data center servers, and then the two are combined to measure the energy consumption of servers under different computing loads, and obtain the unit computing power energy consumption of servers under different computing loads, and then determine the optimal corresponding load rate of the unit computing power energy consumption of the server, that is, the optimal load rate. By scheduling and allocating computing loads between data center server clusters, the overall energy consumption of the data center is minimized when calculating the same computing power tasks, and at the same time, the most energy-saving cooling temperature of the server under different load rates is calculated and tested, and the cooling temperature is adjusted by monitoring the changes in the server load rate, so that the overall energy consumption of the data center computer room is minimized.

[0055] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the framework and scope of application of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for coordinated dispatching and controlling electric power and heat power based on server load rate, characterized in that: The following steps are involved: Step 1: Select sample servers from the server cluster and measure the unit computing power energy consumption of the sample servers under different computing power loads; Step 2: Determine the corresponding optimal load rate range based on the unit computing power energy consumption measurement results of the sample server obtained in step 1; Step 3: According to the optimal load rate range in step 2, the energy-saving cooling temperature of the sample server at different load rates is measured and calculated to obtain the minimum cooling temperature range corresponding to the load rate range; Step 4: Schedule and allocate the computing load among the server clusters according to the optimal load rate range, and set the cooling temperature according to the allocated computing load.

2. The method for controlling the coordinated dispatch of electric power and heat power based on server load rate according to claim 1 is characterized in that: In step 1, the specific steps of measurement are: S1: Deploy energy consumption and performance indicator monitoring software on the sample server; S2: Perform energy consumption tests on the sample server under different computing load levels; S3: Calculate the energy consumption per unit computing power of the sample server based on the energy consumption test results.

3. The method for controlling the coordinated dispatch of electric power and heat power based on server load rate according to claim 2 is characterized in that: The range of the different computing power loads is 0% to 100%. Within this range, several groups of computing power load values ​​are set for testing.

4. The method for controlling the coordinated dispatch of electric power and heat power based on server load rate according to claim 1 is characterized in that: In the step 2, based on the optimal load rate range, the optimal load rate within the range is determined, and the optimal load rate is the target load rate of the server.

5. The method for controlling the coordinated dispatch of electric power and heat power based on server load rate according to claim 1 is characterized in that: In step 3, the specific steps of measurement and calculation are: SS1: Collect the heat output data of the server at different load rates, and then collect the efficiency data of the cooling equipment; SS2: Based on the data collected by SS1, a digital model of heat output data and load rate, as well as a digital model of cooling efficiency, were established; SS3: The two digital models in SS2 are coupled to obtain an energy consumption model. Based on the energy consumption model, the heat generated by the server and the energy consumption required by the cooling equipment under different load rates are calculated.

6. The method for controlling the coordinated dispatch of electric power and heat power based on server load rate according to claim 5 is characterized in that: In the step three, a linear programming algorithm is used to minimize the energy consumption of the refrigeration system, and the refrigeration temperature is adjusted under the premise of ensuring that the server will not overheat, so as to find the refrigeration temperature with the lowest energy consumption.

7. The method for controlling the coordinated dispatch of electric power and heat power based on server load rate according to claim 6 is characterized by: The condition that the server will not overheat is: the maximum operating temperature of the server is preset, and when the internal temperature of the server reaches more than 90% of the maximum operating temperature, the server is considered to be overheated.

8. The method for controlling electric power and heat coordinated dispatch based on server load rate according to claim 1 is characterized in that: In the step 4, the scheduling allocation rule is: preferentially allocate tasks to servers with lower loads, and make the server loads approach the optimal load rate.

Images