Server resource scheduling method and server

Abstract

The invention relates to a server resource scheduling method and a server, and the server scheduling method comprises the steps: obtaining a resource demand quantity and the current temperature of each sub-server, inputting the resource demand quantity and the current temperature into a pre-trained prediction model, and outputting the predicted temperature and the predicted loadable capacity of each sub-server by the prediction model; and distributing the resource demand quantity to each sub-server by using a dynamic load balancing algorithm and the predicted load capacity of each sub-server. According to the method and the device, server crash / disconnection caused by overload of the sub-server is effectively avoided.

Description

Technical Field

[0001] The present invention relates to the field of server resource scheduling, and particularly to a server resource scheduling method and a server. Background Art

[0002] There are many factors affecting the server temperature, such as mechanical cooling systems, ambient temperature, etc. When the server temperature is too high, we will give priority to using the mechanical cooling system for heat dissipation. In the case of insufficient heat dissipation, the server load will be immediately reduced, and even the server will be controlled to stop serving to reduce the server temperature as soon as possible.

[0003] However, directly reducing the server load will cause the server to be unable to provide sufficient resources to the client within a short period of time, greatly affecting the utilization of server resources. Summary of the Invention

[0004] In view of the server resource scheduling problem, the present invention actively adjusts the load of each sub-server so that each sub-server works within a suitable temperature range.

[0005] In a first aspect, the present application provides a server resource scheduling method, which includes:

[0006] Obtain the resource demand and the current temperature of each sub-server, and input them into a pre-trained prediction model, and the prediction model outputs the predicted temperature and predicted load capacity of each sub-server;

[0007] Use the dynamic load balancing algorithm and the predicted load capacity of each sub-server to allocate the resource demand to each sub-server.

[0008] In some embodiments, the obtaining the resource demand and the current temperature of each sub-server, and inputting them into a pre-trained prediction model, and the prediction model outputs the predicted temperature and predicted load capacity of each sub-server includes:

[0009] Predict the predicted temperature of each sub-server when using the load balancing algorithm and the standard load capacity of each sub-server to allocate the resource demand to each sub-server;

[0010] Based on the predicted temperature of each sub-server, obtain the attenuation weight;

[0011] Multiply the attenuation weight by the standard load capacity of each sub-server to obtain the predicted load capacity.

[0012] In some embodiments, the obtaining the attenuation weight based on the predicted temperature of each sub-server includes:

[0013] When the predicted temperature of the sub-server is less than the first temperature threshold, configure the attenuation weight to be 1;

[0014] When the predicted temperature of the sub-server is greater than the first temperature threshold, calculate the attenuation weight according to the magnitude of the predicted temperature, where the predicted temperature is negatively correlated with the attenuation weight.

[0015] In some embodiments, the calculating the attenuation weight according to the magnitude of the predicted temperature includes:

[0016] Confirm the temperature range where the predicted temperature is located;

[0017] Confirm the attenuation weight based on the temperature range.

[0018] In some embodiments, the confirming the attenuation weight based on the temperature range includes:

[0019] Assign a standard weight range to each temperature range;

[0020] Perform mean interpolation using the predicted temperature and the larger end-point value of the temperature range where the predicted temperature is located;

[0021] Obtain the attenuation weight according to the result of the interpolation and the first formula:

[0022] First formula: y min +(1 - x i - x min )×(y max - y min );

[0023] x max - x min

[0024] where, y min is the smaller end-point value of the standard weight range, y max is the larger end-point value of the standard weight range, x min is the smaller end-point value of the temperature range, x max is the larger end-point value of the temperature range, x i is the result of the interpolation.

[0025] In some embodiments, the temperature range includes a first temperature range, a second temperature range, and a third temperature range; the first temperature range corresponds to a first attenuation weight, the second temperature range corresponds to a second attenuation weight, and the third temperature range corresponds to a third attenuation weight; where,

[0026] the temperature of the first temperature range is less than the temperature of the second temperature range, the temperature of the second temperature range is less than the temperature of the third temperature range; and the first attenuation weight is less than the second attenuation weight, and the second attenuation weight is less than the third attenuation weight.

[0027] In some embodiments, the difference between the first attenuation weight and the second attenuation weight is less than the difference between the second attenuation weight and the third attenuation weight.

[0028] In some embodiments, after allocating the resource demand to each sub-server using the dynamic load balancing algorithm and the predicted load capacity of each sub-server, and before obtaining the next resource demand, the method further includes:

[0029] Periodically predicting the predicted temperature of each sub-server when allocating the resource demand to each sub-server using the dynamic load balancing algorithm and the predicted load capacity of each sub-server;

[0030] Updating the attenuation weight using the predicted temperature obtained periodically;

[0031] Updating the predicted load capacity using the updated attenuation weight.

[0032] In some embodiments, the pre-trained prediction model includes: a recurrent neural network, where

[0033] the training data of the recurrent neural network includes: a temperature change curve and a load capacity for a preset duration;

[0034] the label of the recurrent neural network includes: the measured temperature value.

[0035] In a second aspect, the present application provides a server, which includes:

[0036] a plurality of sub-servers;

[0037] a controller, connected to the temperature control systems of each sub-server to obtain the current temperature of each sub-server output by the temperature control system; the controller is configured to execute a server resource scheduling program to implement the above method.

[0038] The present application first adjusts the predicted load capacity using the predicted temperature information, and then performs task allocation / resource allocation based on the predicted load capacity dynamic load balancing algorithm, so as to dynamically adjust the load of each sub-server according to the temperature and workload, effectively avoiding server crashes / disconnections caused by sub-server overload. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a flowchart of an embodiment of the server resource scheduling method of the present application;

[0040] Figure 2 is a flowchart of another embodiment of the server resource scheduling method of the present application;

[0041] Figure 3It is a flowchart of another embodiment of the server resource scheduling method of this application. Detailed implementation manners

[0042] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying 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. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts fall within the protection scope of the present invention.

[0043] In the description of the present invention, the terms "first" and "second" are only used for descriptive purposes, and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present invention, "a plurality of" means two or more, unless otherwise specifically defined.

[0044] In the description of the present invention, the term "for example" is used to mean "serving as an example, illustration, or explanation". Any embodiment described as "for example" in the present invention is not necessarily construed as being more preferred or more advantageous than other embodiments. In order to enable any person skilled in the art to implement and use the present invention, the following description is given. In the following description, details are set forth for purposes of explanation. It should be understood that those skilled in the art can recognize that the present invention can be implemented without using these specific details. In other instances, well-known structures and processes are not elaborated in detail to avoid unnecessary details from obscuring the description of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed herein.

[0045] Example 1

[0046] Referring to Figure 1 , in a first aspect, this application provides a server resource scheduling method, and the method includes:

[0047] S100. Obtain the resource demand and the current temperature of each sub-server, and input them into a pre-trained prediction model, and the prediction model outputs the predicted temperature and predicted load capacity of each sub-server;

[0048] Exemplarily, a server may be composed of multiple sub-servers, and each sub-server has a pre-calibrated standard load capacity, which is the maximum load that can be provided when the sub-server works in the optimal temperature range.

[0049] In some embodiments, the prediction model can also be a non-neural network model, but a fitting curve, where the independent variable of the curve is the resource demand and the dependent variable is the predicted temperature of each sub-server.

[0050] In the embodiments of the present application, the pre-trained prediction module is preferably a neural network. In comparison, the neural network can more accurately regress and predict the predicted temperature and the predicted load capacity. Exemplarily, the neural network can be one or a combination of a convolutional neural network (CNN), a recurrent neural network (RNN), and a deep neural network (DNN), which is specifically selected according to actual needs and is not limited here. In this embodiment, a recurrent neural network is preferably used, which has continuity and is more suitable for processing continuous data and performing prediction functions.

[0051] In practical applications, to train the neural network, we can prepare training data and divide the training data into a training set, a test set, and a validation set. Label each training data with "predicted temperature". After iterative training, the prediction model obtains a set of converged parameters. Furthermore, based on the input resource demand and the current temperature of each sub-server, the predicted temperature and the predicted load capacity of each sub-server can be output.

[0052] S200. Use the dynamic load balancing algorithm and the predicted load capacity of each sub-server to allocate the resource demand to each sub-server.

[0053] Among them, the dynamic load balancing algorithm is an algorithm that determines task allocation according to the real-time load status of the server. Different from the static load balancing algorithm, the dynamic load balancing algorithm can be adjusted according to the current performance and status of the server, so as to more effectively utilize resources and improve the overall performance of the system.

[0054] Based on the dynamic balancing algorithm, the present application updates the predicted load capacity using the predicted temperature information, and then uses the dynamic balancing algorithm to allocate tasks based on the predicted load capacity.

[0055] The present application first adjusts the predicted load capacity using the predicted temperature information, and then performs task allocation / resource allocation based on the predicted load capacity and the dynamic load balancing algorithm. Furthermore, the load of each sub-server can be dynamically adjusted according to the temperature and workload, effectively avoiding server crashes / disconnections caused by sub-server overload.

[0056] Refer to Figure 2 , in some embodiments, S100. Obtain the resource demand and the current temperature of each sub-server, and input them into the pre-trained prediction model, and the prediction model outputs the predicted temperature and the predicted load capacity of each sub-server, including:

[0057] S101. Predict the predicted temperature of each sub-server when allocating the resource demand to each sub-server using the load balancing algorithm and the standard load capacity of each sub-server;

[0058] Among them, the standard load capacity can be confirmed by calibration, that is, the load capacity of the server when the preset server works in the best environment. Step S101 can be implemented by a neural network model.

[0059] S102. Obtain the attenuation weight based on the predicted temperature of each sub-server;

[0060] In this application, the higher the predicted temperature, the greater the attenuation weight. And in this application, the predicted temperature and the attenuation weight are not in a linear relationship, but rather as the predicted temperature increases, the attenuation weight changes at a greater rate of change. For example, an exponential relationship.

[0061] S103. Multiply the attenuation weight by the standard load capacity of each sub-server to obtain the predicted load capacity. That is, as the predicted temperature increases, the predicted load capacity decreases, and decreases at a faster rate.

[0062] In some embodiments, S102. Obtain the attenuation weight based on the predicted temperature of each sub-server, including:

[0063] When the predicted temperature of the sub-server is less than the first temperature threshold, configure the attenuation weight to be 1;

[0064] When the predicted temperature of the sub-server is greater than the first temperature threshold, calculate the attenuation weight according to the magnitude of the predicted temperature, where the predicted temperature is negatively correlated with the attenuation weight.

[0065] In this application, when it is lower than the first temperature threshold, it means that the server is working in a better temperature range and no attenuation is required. When it is higher than the first temperature threshold, the attenuation weight needs to be calculated according to the magnitude of the predicted temperature.

[0066] Exemplarily, this application gives several methods for confirming the attenuation weight. Refer to the following embodiments:

[0067] Refer to Figure 3 , in some embodiments, the calculating the attenuation weight according to the magnitude of the predicted temperature includes:

[0068] S1021. Confirm the temperature range where the predicted temperature is located;

[0069] Step to confirm the attenuation weight based on the temperature range.

[0070] That is, we first divide multiple temperature ranges. Exemplarily, the division of temperature ranges is not equidistant. Instead, the higher the temperature, the smaller the range of the temperature range, and the lower the temperature, the larger the range of the temperature range. This is more in line with the fact that the server is more affected by temperature at high temperatures and less affected by temperature at low temperatures.

[0071] Specifically, the step of confirming the attenuation weight based on the temperature range includes:

[0072] S10221. Assign a standard weight range to each temperature range;

[0073] S10222. Perform mean interpolation using the predicted temperature and the larger endpoint value of the temperature range where the predicted temperature is located;

[0074] S10223. Obtain the attenuation weight according to the result of the interpolation and the first formula:

[0075] First formula:

[0076] where y min is the smaller endpoint value of the standard weight range, y max is the larger endpoint value of the standard weight range, x min is the smaller endpoint value of the temperature range, x max is the larger endpoint value of the temperature range, x i is the result of the interpolation.

[0077] Exemplarily, the temperature range is (60 - 75], the standard weight range is (0.8 - 0.9], and the predicted temperature is 68 degrees. At this time, mean interpolation can be performed between 68 and 75, and the obtained result is 71.5; then normalization is performed in the temperature range (60 - 75], that is, the data is mapped to the range of 0 to 1 for processing, and the normalization result is That is, 0.77. At this time, it is 0.8 + (1 - 0.77) × (0.9 - 0.8) = 0.823.

[0078] In the above calculation formula, as the predicted temperature is closer to the larger endpoint value, the attenuation weight is greater, and thus the temperature can be better adjusted and the predicted load capacity can be confirmed.

[0079] In other embodiments, a standard weight is assigned to each temperature range. For example, the temperature range is (60 - 75], and the standard weight is 0.85.

[0080] Exemplarily, the temperature range includes a first temperature range, a second temperature range, and a third temperature range; the first temperature range corresponds to a first attenuation weight, the second temperature range corresponds to a second attenuation weight, and the third temperature range corresponds to a third attenuation weight; where

[0081] The temperature of the first temperature range is less than the temperature of the second temperature range, and the temperature of the second temperature range is less than the temperature of the third temperature range; and the first attenuation weight is less than the second attenuation weight, and the second attenuation weight is less than the third attenuation weight.

[0082] In some embodiments, the difference between the first attenuation weight and the second attenuation weight is less than the difference between the second attenuation weight and the third attenuation weight.

[0083] In this embodiment, the attenuation weights in the same temperature range are the same, which can reduce the calculation amount;

[0084] Between different temperature ranges, as the temperature increases, the attenuation weight increases non-linearly, and the slope becomes larger, that is, as the temperature increases, the attenuation weight increases in an approximately exponential form.

[0085] In some embodiments, after allocating the resource demand to each sub-server by using the dynamic load balancing algorithm and the predicted load capacity of each sub-server, and before obtaining the next resource demand, the method further includes:

[0086] Periodically predicting the predicted temperature of each sub-server when using the dynamic load balancing algorithm and the predicted load capacity of each sub-server to allocate the resource demand to each sub-server;

[0087] Updating the attenuation weight by using the periodically obtained predicted temperature;

[0088] Updating the predicted load capacity by using the updated attenuation weight.

[0089] In this embodiment, the frequency of obtaining resource demands may be relatively low. For example, during model training, tasks are allocated only once an hour or even half a day. To better adjust the temperature of each sub-server, the attenuation weight update strategy added in this application, that is, periodically (for example, every 10 minutes) predicting the predicted temperature of each sub-server, and then using this predicted temperature to update the attenuation weight, and further updating the load capacity.

[0090] In some embodiments, the pre-trained prediction model includes: a recurrent neural network, where,

[0091] The training data of the recurrent neural network includes: a temperature change curve and a load capacity with a preset duration;

[0092] The label of the recurrent neural network includes: the measured temperature value.

[0093] It should be noted that the recurrent neural network can be an RNN, LSTM, GRU, or other more complex recurrent neural networks, which is not limited in this embodiment. Preferably, it is an LSTM. Compared with the RNN, it avoids the problem of excessive memory, and compared with the GRU, its computational complexity is smaller.

[0094] In practical applications, the training data and labels can be matched first, and then divided into a training set, a test set, and a validation set, and then training, testing, and validation are performed.

[0095] Example 2

[0096] This application provides a server, which includes:

[0097] Multiple sub-servers;

[0098] A controller, connected to the temperature control systems of each sub-server to obtain the current temperature of each sub-server output by the temperature control system; the controller is configured to execute a server resource scheduling program to implement the following method:

[0099] S100. Obtain the resource demand and the current temperature of each sub-server, and input them into a pre-trained prediction model, and the prediction model outputs the predicted temperature and predicted load capacity of each sub-server;

[0100] S200. Use the dynamic load balancing algorithm and the predicted load capacity of each sub-server to allocate the resource demand to each sub-server.

[0101] Refer to Figure 2 , in some embodiments, S100. Obtain the resource demand and the current temperature of each sub-server, and input them into a pre-trained prediction model, and the prediction model outputs the predicted temperature and predicted load capacity of each sub-server, including:

[0102] S101. Predict the predicted temperature of each sub-server when using the load balancing algorithm and the standard load capacity of each sub-server to allocate the resource demand to each sub-server;

[0103] S102. Obtain the attenuation weight based on the predicted temperature of each sub-server;

[0104] S103. Multiply the attenuation weight by the standard load capacity of each sub-server to obtain the predicted load capacity.

[0105] Refer to Figure 3 , in some embodiments, S102. Obtain the attenuation weight based on the predicted temperature of each sub-server, including:

[0106] When the predicted temperature of the sub-server is less than the first temperature threshold, configure the attenuation weight to be 1;

[0107] When the predicted temperature of the sub-server is greater than the first temperature threshold, calculate the attenuation weight according to the magnitude of the predicted temperature, where the predicted temperature is negatively correlated with the attenuation weight.

[0108] In some embodiments, calculating the attenuation weight according to the magnitude of the predicted temperature includes:

[0109] S1021. Confirm the temperature range where the predicted temperature is located;

[0110] S1022. Confirm the attenuation weight based on the temperature range.

[0111] In some embodiments, S1022. Confirm the attenuation weight based on the temperature range, includes:

[0112] S10221. Assign a standard weight range to each temperature range;

[0113] S10222. Perform mean interpolation using the predicted temperature and the larger endpoint value of the temperature range where the predicted temperature is located;

[0114] S10223. Obtain the attenuation weight according to the result of the interpolation and the first formula:

[0115] First formula:

[0116] where y min is the smaller endpoint value of the standard weight range, y max is the larger endpoint value of the standard weight range, x min is the smaller endpoint value of the temperature range, x max is the larger endpoint value of the temperature range, x i is the result of the interpolation.

[0117] In some embodiments, the temperature range includes a first temperature range, a second temperature range, and a third temperature range; the first temperature range corresponds to a first attenuation weight, the second temperature range corresponds to a second attenuation weight, and the third temperature range corresponds to a third attenuation weight; where

[0118] the temperature of the first temperature range is less than the temperature of the second temperature range, and the temperature of the second temperature range is less than the temperature of the third temperature range; and the first attenuation weight is less than the second attenuation weight, and the second attenuation weight is less than the third attenuation weight.

[0119] In some embodiments, the difference between the first attenuation weight and the second attenuation weight is less than the difference between the second attenuation weight and the third attenuation weight.

[0120] In some embodiments, after allocating the resource demand to each sub-server by using the dynamic load balancing algorithm and the predicted load capacity of each sub-server, and before obtaining the next resource demand, the method further includes:

[0121] Periodically predicting the predicted temperature of each sub-server when allocating the resource demand to each sub-server by using the dynamic load balancing algorithm and the predicted load capacity of each sub-server;

[0122] Updating the decay weight by using the periodically obtained predicted temperature;

[0123] Updating the predicted load capacity by using the updated decay weight.

[0124] In some embodiments, the pre-trained prediction model includes: a recurrent neural network, wherein,

[0125] The training data of the recurrent neural network includes: a temperature change curve and a load capacity for a preset duration;

[0126] The label of the recurrent neural network includes: the measured temperature value.

[0127] It should be noted that in the above embodiments, the descriptions of the respective embodiments have their own focuses. For parts not described in detail in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0128] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0129] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, and the combination of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0130] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means that implement the function specified in one or more processes and / or blocks Figure 1 in the flowchart Figure 1 a flowchart or flowcharts and / or blocks

[0131] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable apparatus provide steps for implementing the function specified in one or more processes and / or blocks Figure 1 in the flowchart Figure 1 a flowchart or flowcharts and / or blocks

[0132] Although the preferred embodiments of the present invention have been described, additional changes and modifications can be made by those skilled in the art once they learn of the basic inventive concept. Therefore, the appended claims are intended to be construed to cover the preferred embodiments as well as all changes and modifications falling within the scope of the present invention

[0133] It is obvious that those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations

Claims

1. A server resource scheduling method, characterized in that: include: Obtain resource demand and current temperature of each sub-server, and input them into a pre-trained prediction model, wherein the prediction model outputs predicted temperature and predicted load capacity of each sub-server; The resource demand is distributed to each sub-server using a dynamic load balancing algorithm and the predicted load capacity of each sub-server.

2. The server resource scheduling method according to claim 1, characterized in that: The resource demand and the current temperature of each sub-server are obtained and input into a pre-trained prediction model, and the prediction model outputs the predicted temperature and the predicted load capacity of each sub-server, including: Predicting the predicted temperature of each sub-server when the resource demand is allocated to each sub-server using a load balancing algorithm and a standard load capacity of each sub-server; Based on the predicted temperature of each sub-server, the attenuation weight is obtained; The decay weight is multiplied by the standard load capacity of each sub-server to obtain the predicted load capacity.

3. The server resource scheduling method according to claim 2, characterized in that: The step of obtaining the attenuation weight based on the predicted temperature of each sub-server includes: When the predicted temperature of the sub-server is less than the first temperature threshold, the attenuation weight is configured to be 1; When the predicted temperature of the sub-server is greater than the first temperature threshold, the attenuation weight is calculated according to the predicted temperature, wherein the predicted temperature is negatively correlated with the attenuation weight.

4. The server resource scheduling method according to claim 3, characterized in that: The calculating the attenuation weight according to the predicted temperature includes: confirming the temperature interval within which the predicted temperature lies; The attenuation weight is determined based on the temperature range.

5. The server resource scheduling method according to claim 4, characterized in that: The determining the attenuation weight based on the temperature interval includes: Assign a standard weight interval to each temperature interval; Perform mean interpolation using the predicted temperature and the larger endpoint value of the temperature interval in which the predicted temperature is located; The attenuation weight is obtained according to the interpolation result and the first formula: First formula: Among them, y min is the smaller endpoint value of the standard weight interval, y max is the larger endpoint value of the standard weight interval, x min is the smaller endpoint value of the temperature range, x max is the larger endpoint value of the temperature range, x i is the result of interpolation.

6. The server resource scheduling method according to claim 4, characterized in that: The temperature interval includes a first temperature interval, a second temperature interval, and a third temperature interval; the first temperature interval corresponds to a first attenuation weight, the second temperature interval corresponds to a second attenuation weight, and the third temperature interval corresponds to a third attenuation weight; wherein, The temperature of the first temperature interval is lower than the temperature of the second temperature interval, and the temperature of the second temperature interval is lower than the temperature of the third temperature interval; and the first attenuation weight is lower than the second attenuation weight, and the second attenuation weight is lower than the third attenuation weight.

7. The server resource scheduling method according to claim 6, characterized in that: A difference between the first attenuation weight and the second attenuation weight is smaller than a difference between the second attenuation weight and the third attenuation weight.

8. The server resource scheduling method according to any one of claims 1 to 7, characterized in that: After allocating the resource demand to each sub-server using the dynamic load balancing algorithm and the predicted load capacity of each sub-server, and before obtaining the next resource demand, the method further includes: Periodically predicting the predicted temperature of each sub-server when allocating the resource demand to each sub-server using a dynamic load balancing algorithm and the predicted available load of each sub-server; Using the predicted temperature obtained periodically, updating the attenuation weight; The predicted load capacity is updated using the updated decay weight.

9. The server resource scheduling method according to claim 1, characterized in that: The pre-trained prediction models include: recurrent neural network, where: The training data of the recurrent neural network includes: a temperature change curve and a load amount of a preset time length; The label of the recurrent neural network includes: a measured temperature value.

10. A server, characterized in that: include: Multiple sub-servers; A controller is connected to the temperature control system of each sub-server to obtain the current temperature of each sub-server output by the temperature control system; the controller is configured to execute a server resource scheduling program to implement the method according to any one of claims 1-9.

Images