CN114491943A

CN114491943A - Information processing method, temperature prediction model training method and device and electronic equipment

Info

Publication number: CN114491943A
Application number: CN202111594622.8A
Authority: CN
Inventors: 吉万鹏; 师晋辉; 王方舟; 王文韬
Original assignee: Beijing Dajia Internet Information Technology Co Ltd
Current assignee: Beijing Dajia Internet Information Technology Co Ltd
Priority date: 2021-12-23
Filing date: 2021-12-23
Publication date: 2022-05-13

Abstract

The disclosure relates to an information processing method, a temperature prediction model training method, a device and an electronic device, wherein the method comprises the following steps: acquiring real-time operation parameters of a target server in a target environment; inputting the real-time operation parameters into a temperature prediction model obtained by pre-training to obtain a target prediction temperature corresponding to the target server; and when the target predicted temperature corresponding to the target server is greater than the temperature threshold, generating temperature early warning information for the target server based on the target predicted temperature corresponding to the target server. Therefore, the target predicted temperature output by the temperature prediction model can be closer to the actual temperature of the target server operation by training the first submodel and the second submodel, so that the temperature prediction model can more accurately predict the temperature of the target server, measures can be timely taken in advance when the temperature of the target server is higher, and the safe operation of the target server is further ensured.

Description

Information processing method, temperature prediction model training method and device and electronic equipment

Technical Field

The present disclosure relates to the field of information processing, and in particular, to an information processing method, a temperature prediction model training method, an apparatus, and an electronic device.

Background

With the advent of the internet age, the size and number of data centers is becoming enormous. The server is a main construction facility of the data center, and is also a key point for ensuring the stable operation of the server, which is an operation and maintenance problem of the data center.

A large amount of heat is generated during the operation of the server, which increases the ambient temperature. Therefore, in the related art, an air conditioning and cooling device and an air circulation system are generally configured in a machine room of a data center to control the temperature of the machine room, so that the temperature of the machine room is always kept below an industry safety temperature red line (for example, 35 degrees celsius) of the operation of a server.

The server uses the rack to install in the computer lab as the unit, and the circulation refrigeration air circulates between the rack, and when the temperature exceeded trade temperature red line, computer lab management system can carry out temperature control fault alarm, and simultaneously, in order to discover server operation temperature fault conditions in advance, the temperature baseline has been established to most computer labs, when the cabinet temperature sensor discovers that server ambient temperature exceeded 31 degrees, can carry out the early warning. Since the server temperature often fluctuates on the baseline and the red line, it cannot be distinguished from the temperature broken line caused by the fault, and false alarm often results.

Disclosure of Invention

In order to overcome the problems in the related art, the present disclosure provides an information processing method, a temperature prediction model training method, an apparatus and an electronic device.

According to a first aspect of the embodiments of the present disclosure, there is provided an information processing method including:

acquiring real-time operation parameters of a target server in a target environment;

inputting the real-time operation parameters into a temperature prediction model obtained by pre-training to obtain a target prediction temperature corresponding to the target server; the temperature prediction model comprises a first submodel and a second submodel, wherein the first submodel is obtained by training by using historical operating parameters of each server in the target environment and the historical operating temperature of each server so as to obtain a first predicted temperature of each server; the second submodel is obtained by training by utilizing the first predicted temperature of each server and the historical operating temperature of each server to obtain a target predicted temperature corresponding to each server;

and when the target predicted temperature corresponding to the target server is greater than the temperature threshold, generating temperature early warning information for the target server based on the target predicted temperature corresponding to the target server.

Optionally, the method further includes:

determining the correlation between the servers according to the historical operating temperature of each server and the corresponding target predicted temperature;

and constructing a relevance graph network based on the server identifications of the servers and the relevance sizes among the servers, wherein the relevance graph network comprises a plurality of nodes, the node identification of each node corresponds to the server identification of one server, and the weight corresponding to the edge connecting any two nodes in the relevance graph network is the relevance size among the servers corresponding to the two nodes.

Optionally, the determining the correlation between the servers according to the historical operating temperatures of the servers and the corresponding target predicted temperatures includes:

determining events corresponding to the servers according to the historical operating temperatures of the servers and the corresponding target predicted temperatures, wherein the events are used for describing the change relationship between the historical operating temperatures of the servers and the corresponding target predicted temperatures;

for each server, calculating the number of the same events which occur with other services at the same time;

for each server, determining the correlation size between the server and other servers based on the number of the same events of the server and other servers, wherein the correlation size is in direct proportion to the number of the same events.

Optionally, the method further includes:

when detecting that a first target node exists in the correlation graph network, acquiring a second target node of which the correlation size with the first target node is larger than a preset correlation size, wherein the target predicted temperature of a server corresponding to the first target node is larger than the temperature threshold value;

determining the number of nodes in the second target node, wherein the target predicted temperature of the corresponding server is greater than the temperature threshold;

and when the number of the nodes is larger than the preset number, generating temperature early warning information of the server corresponding to the first target node and the server corresponding to the second target node.

In a second aspect, an embodiment of the present disclosure provides a method for training a temperature prediction model, where the temperature prediction model includes a first sub-model and a second sub-model, and includes:

acquiring historical operating parameters of each server in a target environment and historical operating temperature of each server;

training the first submodel by using the historical operating parameters of each server and the historical operating temperature of each server to obtain a first predicted temperature corresponding to each server;

training the second submodel by using the first predicted temperature of each server and the historical operating temperature of each server to obtain a target predicted temperature corresponding to each server;

determining a loss function value of a temperature prediction model based on the target prediction temperature corresponding to each server and the historical operating temperature of each server;

and when the loss function value is smaller than a preset loss value, determining to obtain a trained temperature prediction model.

Optionally, the training the second submodel by using the first predicted temperature of each server and the historical operating temperature of each server includes:

for each server, forming a first key value pair by the first predicted temperature corresponding to the server and the historical operating temperature corresponding to the server;

determining the first key-value pair as a training sample;

inputting the training sample into the second submodel to train the second submodel.

Optionally, the inputting the training sample into the second submodel to train the second submodel includes:

sorting the plurality of first key value pairs according to the value of the first predicted temperature in the first key value pair to obtain a sorted first key value pair;

dividing the sorted first key value pairs into barrels, and acquiring the average value of the historical operating temperature of the first key value pairs in each barrel;

for each sorted first key value pair, forming a second key value pair by the first predicted temperature in the first key value pair and the mean value corresponding to the sub-barrel where the first key value pair is located;

inputting the second key value pair into a second submodel, and training the second submodel; and the first predicted temperature in the second key value pair is a characteristic value, and the mean value in the second key value pair is a label.

In a third aspect, an embodiment of the present disclosure provides an information processing apparatus, including:

the operation parameter acquisition module is configured to acquire real-time operation parameters of the target server in the target environment;

the temperature prediction module is configured to input the real-time operation parameters into a temperature prediction model obtained by pre-training to obtain a target prediction temperature corresponding to the target server; the temperature prediction model comprises a first submodel and a second submodel, wherein the first submodel is obtained by training by using historical operating parameters of each server in the target environment and the historical operating temperature of each server so as to obtain a first predicted temperature of each server; the second submodel is obtained by training by utilizing the first predicted temperature of each server and the historical operating temperature of each server to obtain a target predicted temperature corresponding to each server;

the first temperature early warning information generation module is configured to execute the generation of temperature early warning information for the target server based on the target predicted temperature corresponding to the target server when the target predicted temperature corresponding to the target server is greater than a temperature threshold value.

Optionally, the apparatus further comprises:

a correlation determination module configured to determine a correlation magnitude between the servers according to the historical operating temperatures of the servers and the corresponding target predicted temperatures;

and the graph network construction module is configured to execute construction of a correlation graph network based on the server identifications of the servers and the correlation sizes among the servers, wherein the correlation graph network comprises a plurality of nodes, the node identification of each node corresponds to the server identification of one server, and the corresponding weight of an edge connecting any two nodes in the correlation graph network is the correlation size among the servers corresponding to the two nodes.

Optionally, the correlation determination module is specifically configured to perform:

and for each server, determining the correlation size between the server and each other server based on the number of the same events of the server and each other server, wherein the correlation size is in direct proportion to the number of the same events.

Optionally, the apparatus further comprises:

the node acquisition module is configured to execute the steps of acquiring a second target node of which the correlation size with the first target node is larger than a preset correlation size when detecting that the correlation graph network has the first target node, wherein the target predicted temperature of a server corresponding to the first target node is larger than the temperature threshold value;

a node determination module configured to perform determining a number of nodes in the second target node for which the target predicted temperature of the corresponding server is greater than the temperature threshold;

and the second temperature early warning information generation module is configured to execute the generation of the temperature early warning information of the server corresponding to the first target node and the server corresponding to the second target node when the number of the nodes is greater than a preset number.

In a fourth aspect, an embodiment of the present disclosure provides a temperature prediction model training apparatus, where the temperature prediction model includes a first submodel and a second submodel, and includes:

the historical data acquisition module is configured to acquire historical operating parameters of each server in a target environment and historical operating temperatures of each server;

a first sub-model training module configured to perform training on the first sub-model by using the historical operating parameters of the servers and the historical operating temperatures of the servers to obtain first predicted temperatures corresponding to the servers;

the second sub-model training module is configured to train the second sub-model by using the first predicted temperature of each server and the historical operating temperature of each server to obtain a target predicted temperature corresponding to each server;

a loss function value determination module configured to perform a determination of a loss function value of a temperature prediction model based on the target predicted temperature corresponding to the respective server and the historical operating temperature of the respective server;

and the temperature prediction model determining module is configured to determine the trained temperature prediction model when the loss function value is smaller than a preset loss value.

Optionally, the second sub-model training module includes:

the key value pair determining unit is configured to execute the step of combining a first predicted temperature corresponding to the server and a historical operating temperature corresponding to the server into a first key value pair for each server;

a training sample determination unit configured to perform determining the first key-value pair as a training sample;

a second submodel training unit configured to perform training of the second submodel by inputting the training samples into the second submodel.

Optionally, the second sub-model training unit is specifically configured to perform:

In a fifth aspect, an embodiment of the present disclosure provides an electronic device, including:

a processor;

a memory for storing processor-executable instructions;

wherein the processor is configured to execute the information processing method of the first aspect or the temperature prediction model training method of the second aspect.

In a sixth aspect, the present disclosure provides a computer-readable storage medium, where instructions of the storage medium, when executed by a processor of a mobile terminal, enable an electronic device to perform the information processing method according to the first aspect or the temperature prediction model training method according to the second aspect.

In a seventh aspect, the present disclosure provides a computer program product, which when run on a computer, causes the computer to execute the information processing method according to the first aspect or the temperature prediction model training method according to the second aspect.

The technical scheme provided by the embodiment of the disclosure can have the following beneficial effects:

according to the technical scheme provided by the embodiment of the disclosure, real-time operation parameters of a target server in a target environment are acquired; inputting the real-time operation parameters into a temperature prediction model obtained by pre-training to obtain a target prediction temperature corresponding to a target server; and when the target predicted temperature corresponding to the target server is greater than the temperature threshold, generating temperature early warning information for the target server based on the target predicted temperature corresponding to the target server. The temperature prediction model is trained through a large amount of historical operating data of each server in a target environment and comprises a first submodel and a second submodel, and the target prediction temperature output by the temperature prediction model can be closer to the real temperature of the target server through training the first submodel and the second submodel, so that the temperature of the target server can be predicted more accurately by the temperature prediction model, measures can be taken in advance in time when the temperature of the target server is higher, and the safe operation of the target server is ensured.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.

Drawings

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

FIG. 1 is a flow diagram illustrating an information processing method according to an exemplary embodiment;

FIG. 2 is a flow diagram illustrating another method of information processing according to an example embodiment;

FIG. 3 is a flowchart of one embodiment of step S140 of FIG. 1;

FIG. 4 is a schematic illustration of a variation of a historical operating temperature and a corresponding target predicted temperature;

FIG. 5 is a flow diagram illustrating another method of information processing according to an example embodiment;

FIG. 6 is a flow diagram illustrating a method of training a temperature prediction model in accordance with an exemplary embodiment;

FIG. 7 is a flowchart of one embodiment of step S630 in FIG. 6;

FIG. 8 is a flowchart of one embodiment of step S633 of FIG. 7;

FIG. 9 is a block diagram illustrating an information processing apparatus in accordance with an exemplary embodiment;

FIG. 10 is a block diagram illustrating a temperature prediction model training apparatus in accordance with an exemplary embodiment;

FIG. 11 is a block diagram illustrating an electronic device in accordance with an exemplary embodiment;

FIG. 12 is a block diagram illustrating an electronic device in accordance with an example embodiment.

Detailed Description

Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the exemplary embodiments below are not intended to represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure, as detailed in the appended claims.

In order to solve the problem of false alarm of the temperature of the server in the related art, the embodiment of the disclosure provides an information processing method, a temperature prediction model training device and electronic equipment.

In a first aspect, an information processing method provided by an embodiment of the present disclosure is first explained in detail.

Fig. 1 is a flowchart illustrating an information processing method according to an exemplary embodiment, which is used in a terminal, as shown in fig. 1, and may include the following steps:

in step S110, real-time operating parameters of the target server in the target environment are obtained.

The target environment in the embodiment of the present disclosure may be a space region including a plurality of servers, for example, a machine room, a cabinet, and the like, and in practical applications, the target environment may further include a refrigeration device. The target server may be any server included in the target environment.

The operation parameters of the target server may include a CPU utilization rate, a DISK read-write speed, a Memory read-write speed, a network transmission speed, an output power of the refrigeration device, and the like, which is not specifically limited in this embodiment of the disclosure.

In step S120, the real-time operating parameters are input into the pre-trained temperature prediction model to obtain a target predicted temperature corresponding to the target server.

The temperature prediction model comprises a first submodel and a second submodel, wherein the first submodel is obtained by training historical operating parameters of all servers in a target environment and historical operating temperatures of all servers so as to obtain first predicted temperatures of all the servers; the second submodel is obtained by training by utilizing the first predicted temperature of each server and the historical operating temperature of each server so as to obtain the target predicted temperature corresponding to each server.

In the embodiment of the disclosure, historical operating parameters of each server collected in a historical time period are obtained; and the historical operating temperature of the server collected by the server sensor is used as the characteristic of the first submodel, the collected historical operating temperature of each server is used as a label, the first submodel is trained, and the first submodel outputs a first predicted temperature. In an embodiment, the operation parameters of each server in the last month or several days may be determined as historical operation parameters, and of course, the operation parameters may be determined according to actual needs, which is not limited herein.

It can be understood that although the trained first sub-model can output the first predicted temperature, the first sub-model trained only through the historical operating parameters and the historical operating temperatures cannot accurately predict the temperature, and therefore, in this embodiment, the second sub-model is provided, and the second sub-model is trained through the first predicted temperature of each server and the historical operating temperature of each server output by the first sub-model. And when the second submodel is trained, the first predicted temperature is used as a characteristic, and the historical operating temperature is used as a label. And after the first submodel and the second submodel are trained, obtaining a temperature prediction model consisting of the first submodel and the second submodel.

Because the historical operating temperature is the real temperature of the server during operation, which is acquired by the server sensor, the predicted temperature output by the temperature prediction model can be closer to the real temperature of the server during operation by training the first submodel and the second submodel, so that the temperature of the target server can be predicted more accurately by the trained temperature prediction model.

In an embodiment, the first sub-model and the second sub-model may both use XGboost with higher precision and faster running time as models, for example, training the first XGboost model by using historical running parameters and historical running temperatures of each server in the last three months. And training the second XGboost model through the first predicted temperature and the historical operating temperature to obtain a temperature prediction model.

For completeness and clarity of description of the scheme, a specific training process of the temperature prediction model will be explained in detail in the following embodiments.

In step S130, when the target predicted temperature corresponding to the target server is greater than the temperature threshold, temperature warning information for the target server is generated based on the target predicted temperature corresponding to the target server.

After the target predicted temperature corresponding to the target server is obtained, if the target predicted temperature is greater than the temperature threshold, the target predicted temperature corresponding to the target server is higher, and temperature early warning information aiming at the target server is generated, so that relevant measures can be taken in time, and the safe operation of the target server is ensured.

The temperature threshold may be set according to an actual situation, which is not specifically limited in the embodiment of the present disclosure.

On the basis of the embodiment shown in fig. 1, as shown in fig. 2, the information processing method may further include the steps of:

and S140, determining the correlation among the servers according to the historical operating temperature of each server and the corresponding target predicted temperature.

Specifically, after the historical operating temperature and the target predicted temperature of each server are obtained, the correlation between any two servers can be determined according to the change trend of the historical operating temperature and the target predicted temperature and the magnitude relation between the historical operating temperature and the target predicted temperature. For example, the historical operating temperature and the target predicted temperature of the two servers have the same trend, and it can be determined that the correlation between the two servers is large. Conversely, the correlation between the two servers is small.

For completeness and clarity of description of the scheme, a detailed implementation of S140 will be set forth in the following examples.

S150, constructing a correlation graph network based on the server identification of each server and the correlation size among the servers.

The relevance graph network comprises a plurality of nodes, the node identification of each node corresponds to the server identification of one server, and the weight corresponding to the edge connecting any two nodes in the relevance graph network is the relevance between the servers corresponding to the two nodes.

Specifically, each server can be regarded as a node, and a network topology graph is established. From the air flow rate and the air flow path, the time required for the air of the refrigeration equipment to circulate for one week can be roughly estimated to be 24 seconds. The predicted operating temperature of each server can be sliced according to 24 seconds to obtain the key value pair (node identification, event) of each node, and accordingly a graph network is constructed. For example, 15 × 3600 × 3000 pieces of data can be obtained by data collection for 15 machine rooms for 15 days and temperature prediction of 3000 nodes.

Moreover, the weight of an edge connecting two nodes can be determined according to the correlation between any two nodes, and the greater the correlation, the greater the determination.

Therefore, through the embodiment, the relevance graph network is established, and the relevance of any two servers is further obtained, so that in the subsequent step, when the temperature of one server is abnormal, the early warning information that other servers with higher relevance are also abnormal is generated.

On the basis of the embodiment shown in fig. 2, in another embodiment provided by the present disclosure, as shown in fig. 3, S140 may include the following steps:

in S141, an event corresponding to each server is determined based on the historical operating temperature of each server and the corresponding target predicted temperature.

The event is used for describing the change relation between the historical operating temperature of each server and the corresponding target predicted temperature.

Specifically, the time corresponding to each server can be determined through two aspects, the first aspect is whether the temperature variation trend T is correct, and the second aspect is whether the temperature value V is correct. In general, there can be a total of 3 events:

1. t is the same, V is different: often appear as waveforms and the variance is close, but the true value, i.e. the historical operating temperature, tends to be higher/lower than the corresponding target predicted temperature, as in fig. 4 a.

2. T is different, V is the same: it is often shown that the mean value is almost the same, but the corresponding target predicted temperature fluctuates up and down around the real value, i.e. the historical operating temperature, or the real value, i.e. the historical operating temperature fluctuates up and down around the target predicted temperature, and the variance between the two is large, as shown in fig. 4 b.

3. T, V are all different: the long-term trends, which often show true values, i.e. the historical operating temperature and the target predicted temperature, are opposite, resulting in a large difference between the mean and the variance as shown in fig. 4 c.

In S142, the number of events that occur at the same time for each server is counted.

After the events occurring at each moment of each server are determined, the number of the events occurring at each moment of each server and the number of the events occurring at other servers can be counted.

In S143, for each server, the correlation size between the server and each of the other servers is determined based on the number of events in which the same event occurs between the server and each of the other servers.

Wherein the correlation magnitude is proportional to the number of occurrences of the same event.

It can be understood that the greater the number of events that two servers send the same event at the same time, the more similar the two servers are, i.e. the greater the correlation between the two servers. Conversely, the smaller the number of events sent by two servers at the same time, the lower the similarity between the two servers, i.e. the smaller the correlation between the two servers.

Therefore, according to the technical scheme provided by the embodiment, the correlation between any two servers can be accurately determined through the variation relation between the historical operating temperature of each server and the corresponding target predicted temperature.

On the basis of the embodiment shown in fig. 3, as shown in fig. 5, the method may further include the steps of:

in S510, when it is detected that a first target node exists in the plurality of nodes, a second target node having a correlation with the first target node greater than a preset correlation is acquired.

The target predicted temperature of the server corresponding to the first target node is greater than the temperature threshold.

Specifically, when the early warning system finds that the node temperature of a certain server is abnormal through the temperature prediction model and needs early warning, the node with high correlation with the current node is further calculated. The preset correlation may be determined according to an actual situation, and is not specifically limited herein.

In S520, the number of nodes in the second target node whose target predicted temperature of the corresponding server is greater than the temperature threshold is determined.

After the node temperature of one server is found to be abnormal, whether the server corresponding to the node with high node correlation is abnormal or not is judged, namely whether the target predicted temperature corresponding to the server with high correlation is larger than a temperature threshold value or not is judged. And counting the number of nodes of which the target predicted temperature of the corresponding server is greater than the temperature threshold in the second target nodes.

In S530, when the number of nodes is greater than the preset number, temperature warning information of the server corresponding to the first target node and the server corresponding to the second target node is generated.

Specifically, when the number of the nodes is larger than the preset number, the server with higher relevance sends temperature abnormity in a large proportion, and the server judges the temperature abnormity as a group temperature abnormity event.

According to the method and the device, the prediction accuracy is improved on the basis of the existing temperature prediction model, the historical operating temperature result is monitored and recorded, the server temperature correlation graph network is constructed in an off-line mode, the cluster with high correlation of each node on the temperature event is obtained through the graph network and the temperature correlation calculation, group alarm is achieved from the angles of various types of server groups and group cabinets, and error alarm can be reduced.

In a second aspect, a temperature prediction model training method provided in the embodiments of the present disclosure is described in detail.

The temperature prediction model provided by the embodiment of the disclosure includes a first sub-model and a second sub-model,

as shown in fig. 6, a method for training a temperature prediction model provided in an embodiment of the present disclosure may include the following steps:

s610, obtaining historical operating parameters of each server and historical operating temperature of each server in the target environment.

The operation parameters of each server may include a CPU utilization rate, a DISK read-write speed, a Memory read-write speed, a network transmission speed, an output power of a refrigeration device, and the like, which is not specifically limited in this embodiment of the disclosure.

Wherein the historical operating temperature of each server may be the temperature collected by the sensor of that server.

S620, training the first submodel by using the historical operating parameters of the servers and the historical operating temperatures of the servers to obtain first predicted temperatures corresponding to the servers.

When the first submodel is trained, the historical operating parameters of each server can be used as the characteristics of the first submodel, the collected historical operating temperature of each server is used as a label, and the output of the first submodel is a first predicted temperature. That is, the first submodel learns historical operating parameters versus temperature.

And S630, training the second submodel by using the first predicted temperature of each server and the historical operating temperature of each server to obtain the target predicted temperature corresponding to each server.

It can be understood that although the trained first sub-model can output the first predicted temperature, the first sub-model trained only through the historical operating parameters and the historical operating temperatures cannot accurately predict the temperature, and therefore, in this embodiment, the second sub-model is provided, and the second sub-model is trained through the first predicted temperature of each server and the historical operating temperature of each server output by the first sub-model. And when the second submodel is trained, the first predicted temperature is used as a characteristic, and the historical operating temperature is used as a label.

In an embodiment, the first sub-model and the second sub-model may both use XGboost with higher precision and faster running time as a model, and for example, the first XGboost model may be trained by historical operating parameters and historical operating temperatures of each server in the last three months. And training the second XGboost model through the first predicted temperature and the historical operating temperature to obtain a temperature prediction model.

And S640, determining a loss function value of the temperature prediction model based on the target prediction temperature corresponding to each server and the historical operating temperature of each server.

After the target predicted temperature corresponding to each server is obtained, the loss value of the temperature prediction model can be determined by calculating the sum variance, mean square error and the like between the target predicted temperature and the historical operating temperature of the server. The embodiment of the present disclosure is not particularly limited to this.

And S650, determining to obtain the trained temperature prediction model when the loss function value is smaller than a preset loss value.

Specifically, if the loss function value is smaller than the preset loss value, it indicates that the target predicted temperature is closer to the corresponding historical operating temperature, that is, the temperature prediction model can accurately predict the operating temperature of each server. Thus, the trained temperature prediction model is determined.

Conversely, if the loss function value is greater than the preset loss value, it indicates that the target predicted temperature is greater than the corresponding historical operating temperature, i.e., the accuracy of the temperature prediction model predicting the operating temperature of each server is lower. Therefore, the temperature prediction model needs to be continuously trained.

The temperature prediction model is trained through a large amount of historical operating data of each server in a target environment and comprises a first submodel and a second submodel, and the target prediction temperature output by the temperature prediction model can be closer to the real temperature of the target server through training the first submodel and the second submodel, so that the temperature of the target server can be predicted more accurately by the temperature prediction model, measures can be taken in advance in time when the temperature of the target server is higher, and the safe operation of the target server is further ensured.

Based on the embodiment shown in fig. 6, in order to elaborate how to train the second submodel, in a further embodiment provided by the present disclosure, as shown in fig. 7, the step S630 may include the following steps:

step S631, for each server, forming a first key-value pair from the first predicted temperature corresponding to the server and the historical operating temperature corresponding to the server.

In step S632, the first key-value pair is determined as a training sample.

Step S633, inputting the training sample into the second submodel to train the second submodel.

In this embodiment, by characterizing the first predicted temperature and the historical operating temperature as a label, the predicted temperature output from the temperature prediction model, which is the second submodel, can be made closer to the historical operating temperature, i.e., the temperature prediction model can more accurately predict the temperature of the target server.

Based on the embodiment shown in fig. 7, in an implementation manner, in step S633, inputting a training sample into the second submodel to train the second submodel, as shown in fig. 8, the following steps may be included:

in step S6331, the plurality of first key value pairs are sorted according to the value of the first predicted temperature in the first key value pair, so as to obtain a sorted first key value pair.

In this embodiment, the second sub-model is trained in a way of order-preserving regression, and specifically, after the N first predicted values and the historical operating temperatures are matched to form N first key-value pairs, the first key-value pairs may be sorted according to the first predicted values. The sorting may consist in arranging the true values to which the first predicted values are close as monotonically increasing or decreasing data as adjacent values to which the linear path is close.

In step S6332, the sorted first key value pairs are subjected to bucket sorting, and a mean value of the historical operating temperatures of the first key value pairs in each bucket is obtained.

In step S6333, for each sorted first key value pair, the first predicted temperature in the first key value pair and the average value corresponding to the sub-bucket where the first key value pair is located are combined into a second key value pair.

Specifically, the sorted first key value pairs are subjected to barrel division, and the average value of the historical operating temperature of the first key value pairs in each barrel division is obtained. By bucket averaging, a new set of key-value pairs, i.e. the second key-value pair, is obtained.

In step S6334, the second key value pair is input into the second submodel, and the second submodel is trained.

And the first predicted temperature in the second key value pair is a characteristic value, and the mean value in the second key value pair is a label.

Since the second key value pair is composed of the first predicted value and the mean value of the historical operating temperature, the first predicted value is closer to the mean value of the historical operating temperature, that is, the predicted values in the second key value pair are closer to the true temperature. Therefore, the temperature can be predicted more accurately by the obtained temperature prediction model by training the second submodel by taking the first predicted temperature in the second key value pair as the characteristic value and the mean value in the second key value pair as the label.

It can be seen that the temperature predicted by the second sub-model obtained by the embodiment is closer to the real temperature, so that the obtained temperature prediction model can predict the temperature more accurately.

In a third aspect, an embodiment of the present disclosure provides an information processing apparatus, as shown in fig. 9, including:

an operation parameter obtaining module 910 configured to perform obtaining real-time operation parameters of a target server in a target environment;

the temperature prediction module 920 is configured to input the real-time operation parameters into a temperature prediction model obtained through pre-training to obtain a target predicted temperature corresponding to the target server; the temperature prediction model comprises a first submodel and a second submodel, wherein the first submodel is obtained by training by using historical operating parameters of each server in the target environment and the historical operating temperature of each server so as to obtain a first predicted temperature of each server; the second submodel is obtained by training by utilizing the first predicted temperature of each server and the historical operating temperature of each server to obtain a target predicted temperature corresponding to each server;

a first temperature warning information generating module 930 configured to generate temperature warning information for the target server based on the target predicted temperature corresponding to the target server when the target predicted temperature corresponding to the target server is greater than the temperature threshold.

Optionally, the apparatus further comprises:

In a fourth aspect, an embodiment of the present disclosure provides a temperature prediction model training apparatus, where the temperature prediction model includes a first sub-model and a second sub-model, as shown in fig. 10, and includes:

a historical data acquisition module 1010 configured to perform acquisition of historical operating parameters of each server and historical operating temperatures of each server in a target environment;

a first sub-model training module 1020 configured to perform training on the first sub-model by using the historical operating parameters of the servers and the historical operating temperatures of the servers, so as to obtain first predicted temperatures corresponding to the servers;

a second submodel training module 1030 configured to perform training on the second submodel by using the first predicted temperature of each server and the historical operating temperature of each server, so as to obtain a target predicted temperature corresponding to each server;

a loss function value determination module 1040 configured to perform determining a loss function value of a temperature prediction model based on the target predicted temperature corresponding to the respective server and the historical operating temperature of the respective server;

a temperature prediction model determination module 1050 configured to determine the trained temperature prediction model when the loss function value is smaller than a preset loss value.

Optionally, the second sub-model training module includes:

a processor;

a memory for storing processor-executable instructions;

Fig. 11 is a block diagram illustrating an apparatus 1100 for information processing according to an example embodiment. For example, the apparatus 1100 is an electronic device, and may be embodied as a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, an exercise device, a personal digital assistant, and the like.

Referring to fig. 11, apparatus 1100 may include one or more of the following components: processing component 1102, memory 1104, power component 1106, multimedia component 11011, audio component 1110, input/output (I/O) interface 1112, sensor component 1114, and communications component 1116.

The processing component 1102 generally controls the overall operation of the device 1100, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations. The processing component 1102 may include one or more processors 1120 to execute instructions to perform all or a portion of the steps of the methods described above. Further, the processing component 1102 may include one or more modules that facilitate interaction between the processing component 1102 and other components. For example, the processing component 1102 may include a multimedia module to facilitate interaction between the multimedia component 11011 and the processing component 1102.

The memory 1104 is configured to store various types of data to support operation at the device 1100. Examples of such data include instructions for any application or method operating on device 1100, contact data, phonebook data, messages, pictures, videos, and so forth. The memory 1104 may be implemented by any type or combination of volatile or non-volatile memory devices 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 or optical disks.

A power component 1106 provides power to the various components of the device 1100. The power components 1106 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the apparatus 1100.

The multimedia component 1108 includes a screen that provides an output interface between the device 1100 and a user. In some embodiments, the screen may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive an input signal from a user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 1108 includes a front facing camera and/or a rear facing camera. The front-facing camera and/or the rear-facing camera may receive external multimedia data when the device 1100 is in an operating mode, such as a shooting mode or a video mode. Each front camera and rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.

The audio component 1110 is configured to output and/or input audio signals. For example, the audio component 1110 includes a Microphone (MIC) configured to receive external audio signals when the apparatus 1100 is in operating modes, such as a call mode, a recording mode, and a voice recognition mode. The received audio signals may further be stored in the memory 1104 or transmitted via the communication component 1116. In some embodiments, the audio assembly 1110 further includes a speaker for outputting audio signals.

The I/O interface 1112 provides an interface between the processing component 1102 and peripheral interface modules, which may be keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to: a home button, a volume button, a start button, and a lock button.

The sensor assembly 1114 includes one or more sensors for providing various aspects of state assessment for the device 1100. For example, the sensor assembly 1114 may detect an open/closed state of the device 1100, the relative positioning of components, such as a display and keypad of the apparatus 1100, the sensor assembly 1114 may also detect a change in position of the apparatus 1100 or a component of the apparatus 1100, the presence or absence of user contact with the apparatus 1100, an orientation or acceleration/deceleration of the apparatus 1100, and a change in temperature of the apparatus 1100. Sensor assembly 1114 may include a proximity sensor configured to detect the presence of nearby objects in the absence of any physical contact. The sensor assembly 1114 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 1114 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

The communication component 1116 is configured to facilitate wired or wireless communication between the apparatus 1100 and other devices. The apparatus 1100 may access a wireless network based on a communication standard, such as WiFi, an operator network (such as 2G, 3G, 4G, or 5G), or a combination thereof. In an exemplary embodiment, the communication component 1116 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 1116 also includes a Near Field Communication (NFC) module to facilitate short-range communications. For example, the NFC module may be implemented based on Radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, Ultra Wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

In an exemplary embodiment, the apparatus 1100 may be implemented by one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), controllers, micro-controllers, microprocessors or other electronic components for performing the above-described information processing methods.

In an exemplary embodiment, a computer-readable storage medium comprising instructions, such as the memory 1104 comprising instructions, executable by the processor 1120 of the apparatus 1100 to perform the method described above is also provided. For example, the computer readable storage medium may be a ROM, a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.

FIG. 12 is a block diagram illustrating an apparatus 1200 for information processing or temperature prediction model training in accordance with an exemplary embodiment. For example, the apparatus 1200 may be provided as a server. Referring to fig. 12, the apparatus 1200 includes a processing component 1222 that further includes one or more processors, and memory resources, represented by memory 1232, for storing instructions, such as application programs, that are executable by the processing component 1222. The application programs stored in memory 1232 may include one or more modules that each correspond to a set of instructions. Further, the processing component 1222 is configured to execute instructions to perform the information processing method or the temperature prediction model training method described above.

The apparatus 1200 may also include a power supply component 1226 configured to perform power management of the apparatus 1200, a wired or wireless network interface 1250 configured to connect the apparatus 1200 to a network, and an input output (I/O) interface 1258. The apparatus 1200 may operate based on an operating system stored in the memory 1232, such as Windows Server, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, or the like.

Because the temperature prediction model is trained through a large amount of historical operating data of each server in a target environment and comprises a first submodel and a second submodel, the target prediction temperature output by the temperature prediction model can be closer to the actual temperature of the target server through training the first submodel and the second submodel, so that the temperature of the target server can be more accurately predicted by the temperature prediction model, measures can be taken in advance in time when the temperature of the target server is higher, and the safe operation of the target server is ensured.

In a seventh aspect, the present disclosure provides a computer program product, which when run on a computer, causes the computer to execute the information processing method according to the first aspect or the temperature prediction model training method according to the second aspect. The temperature prediction model is trained through a large amount of historical operating data of each server in a target environment and comprises a first submodel and a second submodel, and the target prediction temperature output by the temperature prediction model can be closer to the real temperature of the target server through training the first submodel and the second submodel, so that the temperature of the target server can be predicted more accurately by the temperature prediction model, measures can be taken in advance in time when the temperature of the target server is higher, and the safe operation of the target server is further ensured.

In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. The procedures or functions described in accordance with the embodiments of the disclosure are, in whole or in part, generated when the computer program instructions are loaded and executed on a computer. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored on a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website, computer, server, or data center to another website, computer, server, or data center via wire (e.g., coaxial cable, fiber, DSL (Digital Subscriber Line)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that incorporates one or more of the available media. The usable medium may be a magnetic medium (e.g., a floppy Disk, a hard Disk, a magnetic tape), an optical medium (e.g., a DVD (Digital Versatile Disk)), or a semiconductor medium (e.g., an SSD (Solid State Disk)), etc.

It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in the process, method, article, or apparatus that comprises the element.

Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

It will be understood that the present disclosure is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. An information processing method characterized by comprising:

2. The method of claim 1, further comprising:

3. The method of claim 2, wherein determining the magnitude of the correlation between the servers based on the historical operating temperatures of the servers and the corresponding target predicted temperatures comprises:

4. The method of claim 3, further comprising:

5. A temperature prediction model training method is characterized in that the temperature prediction model comprises a first submodel and a second submodel, and comprises the following steps:

6. An information processing apparatus characterized by comprising:

7. A temperature prediction model training apparatus, wherein the temperature prediction model comprises a first sub-model and a second sub-model, comprising:

8. An electronic device, comprising:

a processor;

a memory for storing processor-executable instructions;

wherein the processor is configured to perform the information processing method of any one of claims 1 to 4, or the temperature prediction model training method of claim 5.

9. A computer-readable storage medium, wherein instructions in the storage medium, when executed by a processor of a mobile terminal, enable an electronic device to perform the information processing method of any one of claims 1 to 4, or the temperature prediction model training method of claim 5.

10. A computer program product, which, when run on a computer, causes the computer to perform the information processing method of any one of claims 1 to 4, or the temperature prediction model training method of claim 5.