CN116595383A - Intelligent cooling method for storage unit of data center - Google Patents
Intelligent cooling method for storage unit of data center Download PDFInfo
- Publication number
- CN116595383A CN116595383A CN202310700589.5A CN202310700589A CN116595383A CN 116595383 A CN116595383 A CN 116595383A CN 202310700589 A CN202310700589 A CN 202310700589A CN 116595383 A CN116595383 A CN 116595383A
- Authority
- CN
- China
- Prior art keywords
- cooling
- heat
- data
- access
- node
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001816 cooling Methods 0.000 title claims abstract description 208
- 238000003860 storage Methods 0.000 title claims abstract description 126
- 238000012545 processing Methods 0.000 claims abstract description 70
- 238000004458 analytical method Methods 0.000 claims abstract description 17
- 238000005206 flow analysis Methods 0.000 claims abstract description 11
- 230000004927 fusion Effects 0.000 claims description 38
- 238000000034 method Methods 0.000 claims description 25
- 238000006243 chemical reaction Methods 0.000 claims description 23
- 238000013507 mapping Methods 0.000 claims description 20
- 238000009826 distribution Methods 0.000 claims description 19
- 239000000110 cooling liquid Substances 0.000 claims description 10
- 238000009499 grossing Methods 0.000 claims description 8
- 235000019633 pungent taste Nutrition 0.000 claims description 8
- 238000012549 training Methods 0.000 claims description 5
- 230000000694 effects Effects 0.000 abstract description 10
- 230000004044 response Effects 0.000 abstract description 6
- 238000013500 data storage Methods 0.000 abstract description 5
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 230000003044 adaptive effect Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 239000002699 waste material Substances 0.000 description 3
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000007417 hierarchical cluster analysis Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000009827 uniform distribution Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Air Conditioning Control Device (AREA)
Abstract
The application relates to the technical field of artificial intelligence, and provides an intelligent cooling method of a storage unit of a data center, which comprises the following steps: the raw connection storage unit is used for acquiring a data processing sample set; carrying out instruction control flow analysis and data access heat analysis by using access instructions corresponding to each sample data, outputting a plurality of storage flows and a plurality of access heats, and constructing a storage instruction tree; when a first task instruction is received, traversing a storage instruction tree, and outputting a first matching node set; the cooling control system is connected, the first cooling control parameters are output through integrated learning and fed back to the cooling control system, the technical problem that the cooling mode depends on manual intervention and cannot be subjected to cooling self-adaptive control is solved, the cold and hot data storage in the data center storage unit is modified, heat prediction is carried out, the cooling control parameters of the data center are converted according to the predicted heat temperature, the response speed and the efficiency of cooling control are improved, and the cooling self-adaptive control technical effect is achieved.
Description
Technical Field
The application relates to the technical field of artificial intelligence, in particular to an intelligent cooling method for a storage unit of a data center.
Background
The cooling control is commonly used for electronic equipment such as air conditioners, refrigerators and the like, and the traditional cooling mode comprises fixed temperature control, manual control and fixed time control, wherein the fixed temperature control does not consider the change of actual load and heat demand, which may cause the problems of excessive cooling or insufficient cooling, waste energy sources or incapability of effectively cooling; the manual control can be to manually adjust the temperature of the air conditioner or the speed of the fan, but the manual control is easy to cause human error and delay, and the cooling system cannot be monitored and adjusted in real time; the fixed time control is performed according to a fixed time schedule, but cannot be adjusted according to actual requirements, which may lead to waste of energy and poor cooling effect.
In summary, the prior art has the technical problem that the cooling mode depends on manual intervention and the cooling adaptive control cannot be performed.
Disclosure of Invention
The application provides an intelligent cooling method for a data center storage unit, and aims to solve the technical problem that a cooling mode in the prior art depends on manual intervention and cooling self-adaption control cannot be performed.
In view of the above, the present application provides an intelligent cooling method for a data center storage unit.
In a first aspect of the disclosure, an intelligent cooling method for a storage unit of a data center is provided, where the method includes: the method comprises the steps of connecting a storage unit of a first data center, and acquiring a data processing sample set of the storage unit, wherein the data processing sample set comprises access instructions corresponding to each sample data and access frequencies corresponding to each sample data; carrying out instruction control flow analysis by using the access instructions corresponding to the sample data, and outputting a plurality of storage flows; performing data access heat analysis according to the access frequency corresponding to each sample data, and outputting a plurality of access heats; dividing tree nodes according to the multiple storage flows, identifying the tree nodes according to the multiple access hotness, and building a storage instruction tree; when the storage unit receives a first task instruction, traversing the storage instruction tree according to the first task instruction, and outputting a first matched node set; and the cooling control system is connected with the storage unit, performs integrated learning according to the first matching node set and the cooling integrated prediction model, outputs a first cooling control parameter, and feeds back the first cooling control parameter to the cooling control system for control.
In another aspect of the present disclosure, an intelligent cooling system for a data center storage unit is provided, wherein the system comprises: the data processing sample set acquisition module is used for connecting a storage unit of the first data center, and acquiring a data processing sample set of the storage unit, wherein the data processing sample set comprises access instructions corresponding to each sample data and access frequencies corresponding to each sample data; the storage flow output module is used for carrying out instruction control flow analysis by using the access instructions corresponding to the sample data and outputting a plurality of storage flows; the access heat output module is used for carrying out data access heat analysis according to the access frequency corresponding to each sample data and outputting a plurality of access heats; the storage instruction tree building module is used for dividing tree nodes according to the storage flows, marking the tree nodes according to the access hotness and building a storage instruction tree; the first matching node set output module is used for traversing the storage instruction tree according to the first task instruction when the storage unit receives the first task instruction, and outputting a first matching node set; and the cooling control module is used for connecting the cooling control system of the storage unit, performing integrated learning according to the first matched node set and the cooling integrated prediction model, outputting a first cooling control parameter, and feeding back to the cooling control system for control.
One or more technical schemes provided by the application have at least the following technical effects or advantages:
because the raw connection storage unit is adopted, a data processing sample set is obtained; carrying out instruction control flow analysis and data access heat analysis by using access instructions corresponding to each sample data, outputting a plurality of storage flows and a plurality of access heats, and constructing a storage instruction tree; when the storage unit receives a first task instruction, traversing a storage instruction tree, and outputting a first matched node set; the cooling control system is connected with the storage unit, performs integrated learning according to the first matched node set and the cooling integrated prediction model, outputs first cooling control parameters, feeds back the first cooling control parameters to the cooling control system for control, realizes the transformation of cold and hot data storage in the storage unit of the data center, performs heat prediction, converts the cooling control parameters of the data center according to the predicted heat temperature, improves the response speed and efficiency of cooling control, and performs the technical effect of cooling adaptive control.
The foregoing description is only an overview of the present application, and is intended to be implemented in accordance with the teachings of the present application in order that the same may be more clearly understood and to make the same and other objects, features and advantages of the present application more readily apparent.
Drawings
FIG. 1 is a schematic diagram of a possible flow chart of an intelligent cooling method for a storage unit of a data center according to an embodiment of the present application;
FIG. 2 is a schematic diagram of a possible flow chart of outputting a first cooling control parameter in an intelligent cooling method of a data center storage unit according to an embodiment of the present application;
FIG. 3 is a schematic diagram of a possible flow chart of outputting dynamic cooling control parameters in an intelligent cooling method of a data center storage unit according to an embodiment of the present application;
fig. 4 is a schematic diagram of a possible configuration of an intelligent cooling system for a storage unit of a data center according to an embodiment of the present application.
Reference numerals illustrate: the system comprises a data processing sample set acquisition module 100, a storage flow output module 200, an access heat output module 300, a storage instruction tree building module 400, a first matching node set output module 500 and a cooling control module 600.
Detailed Description
The embodiment of the application provides an intelligent cooling method for a data center storage unit, which solves the technical problem that a cooling mode depends on manual intervention and cannot be subjected to cooling adaptive control, and realizes the technical effects of modifying cold and hot data storage in the data center storage unit, carrying out heat prediction, converting cooling control parameters of a data center according to the predicted heat temperature, improving the response speed and efficiency of cooling control and carrying out cooling adaptive control.
Having described the basic principles of the present application, various non-limiting embodiments of the present application will now be described in detail with reference to the accompanying drawings.
Example 1
As shown in fig. 1, an embodiment of the present application provides an intelligent cooling method for a storage unit of a data center, where the method includes:
s10: the method comprises the steps of connecting a storage unit of a first data center, and acquiring a data processing sample set of the storage unit, wherein the data processing sample set comprises access instructions corresponding to each sample data and access frequencies corresponding to each sample data;
s20: carrying out instruction control flow analysis by using the access instructions corresponding to the sample data, and outputting a plurality of storage flows;
s30: performing data access heat analysis according to the access frequency corresponding to each sample data, and outputting a plurality of access heats;
step S30 includes the steps of:
s31: performing differential analysis according to the access frequency of each sample data to obtain a differential index set;
s32: clustering is carried out according to the coordinate distribution of the difference index set, and a difference index clustering result is output;
s33: and marking the cold and hot degrees of each sample data according to the difference index clustering result, and outputting a plurality of access hot degrees.
Specifically, the storage unit of the first data center is used for storing units or devices, such as a server, a hard disk and the like, of data in the data center, and is connected with the storage unit of the first data center and acquires a data processing sample set of the storage unit, wherein the data processing sample set comprises an access instruction and an access frequency corresponding to each sample data, each sample data corresponds to one access instruction and one access frequency, and the access instruction refers to an instruction for accessing or operating the data in the storage unit and can be read, write and the like; the access frequency refers to the frequency or the times of the access instruction corresponding to each sample data in a certain time, and is used for measuring the heat of the data or the frequency of access; the data in the data processing sample set is collected from actual operation, and the storage unit only supports reading and writing and does not support data rewriting;
analyzing and processing access instructions corresponding to each sample data to determine a control mode and sequence of a storage flow, performing instruction control flow analysis on the access instructions corresponding to each sample data, and outputting a plurality of storage flows through analyzing the relation among the instructions, wherein the plurality of storage flows describe the relation between the data processing flow and a storage unit, and the storage flows refer to an operation flow of the storage unit obtained through instruction control flow analysis and are used for guiding the cooling operation of the storage unit;
analyzing the access frequency corresponding to each sample data to determine the heat degree or the access frequency of the data, analyzing the data access heat degree according to the access frequency corresponding to each sample data, and outputting a plurality of access heat degrees, wherein the difference index set is obtained by comparing the access frequency of each sample data and reflects the difference degree between the access frequencies of different data;
according to the coordinate distribution of the difference index set, a clustering algorithm (such as K-means, hierarchical clustering and the like) can be used for clustering the data. Clustering to calculate similarity to divide the data clusters into different clusters, and performing bottom-up aggregation hierarchical clustering analysis on the material structure basic information and the connection fixed process information set to form a difference index clustering result;
if some clusters in the clustering result can be regarded as hot data, the data can be regarded as frequently accessed, and other clusters can be regarded as cold data, the data can be regarded as less accessed, according to the identifications, the cold and hot degrees of each sample data are identified according to the difference index clustering result, and a plurality of access heat degrees are output for further data access adjustment and cooling control operation;
in summary, by performing differential analysis and clustering on the access frequency of the sample data, the coldness and warmth of different data can be identified, so that a plurality of access warmth are output, and the optimization and energy-saving effects of the intelligent cooling system are supported.
S40: dividing tree nodes according to the multiple storage flows, identifying the tree nodes according to the multiple access hotness, and building a storage instruction tree;
s50: when the storage unit receives a first task instruction, traversing the storage instruction tree according to the first task instruction, and outputting a first matched node set;
s60: and the cooling control system is connected with the storage unit, performs integrated learning according to the first matching node set and the cooling integrated prediction model, outputs a first cooling control parameter, and feeds back the first cooling control parameter to the cooling control system for control.
Specifically, based on an FP-tree (Frequent Pattern Tree frequent pattern tree) as a model, dividing the multiple storage flows into multiple tree nodes according to the divided tree nodes and the identified access hotness, identifying the multiple tree nodes according to the multiple access hotness, and constructing a storage instruction tree, wherein the storage instruction tree can be used for representing the dependency relationship between the data processing flows and the nodes, reconstructing cold and hot data storage in a data center storage unit, and helping to effectively manage and organize storage instructions;
the first task instruction is an instruction for describing cooling operation on the storage unit, and may include specifying cooling time, a cooling method or other related parameters, when the storage unit receives the first task instruction, traversing the storage instruction tree according to the first task instruction, and determining which nodes need to perform cooling operation by selecting corresponding nodes according to requirements and dependency relations of the first task instruction in the traversing process, so as to output a first matching node set, where the first matching node set is used for characterizing a node set matched with the first task instruction found in the traversing process of the storage instruction tree, and elements in the first matching node set may be nodes needing cooling control;
the data center storage unit is connected with the cooling control system of the storage unit, receives the called instruction, performs integrated learning according to the first matching node set and the cooling integrated prediction model, and outputs a first cooling control parameter, wherein the first cooling control parameter is a cooling control parameter of the data center, and comprises related control parameters such as a cooling liquid flow control parameter and the like, and the cooling control parameter is obtained by heat temperature prediction of the cooling integrated prediction model; the first cooling control parameters are fed back to a cooling control system to carry out actual cooling control operation, so that the cooling strategy can be ensured to be adjusted according to the results of the matched nodes and the integrated learning model, and a more efficient and energy-saving cooling effect is provided; by combining data characteristics and cooling requirements, dynamic cooling adjustment is achieved by advanced techniques and automated control systems to improve energy utilization efficiency and reduce energy waste.
As shown in fig. 2, step S60 includes the steps of:
s61: acquiring a first matching node set, wherein the first matching node set comprises operation flow nodes obtained by matching after traversing the storage instruction tree based on the first task instruction;
s62: inputting the first set of matching nodes into the cooling integrated prediction model, the cooling integrated prediction model comprising a thermal predictor model and a thermal converter model;
s63: acquiring node heat corresponding to each matching node in the first matching node set, and outputting a matching node heat set;
s64: according to the heat predictor model, fusion prediction is carried out on the heat set of the matched node, and prediction heat information for identifying the first task instruction is output;
s65: and learning the predicted heat information as input data of the heat conversion submodel, and outputting a first cooling control parameter.
Specifically, performing integrated learning according to the first matching node set and the cooling integrated prediction model, and outputting a first cooling control parameter, wherein the first cooling control parameter comprises traversing a stored instruction tree according to a first task instruction, and finding out operation flow nodes matched with the instruction, wherein the obtained matching nodes form the first matching node set; inputting a first matched node set into a cooling integrated prediction model for temperature prediction, wherein the cooling integrated prediction model comprises two sub-models: a caloric predictor model and a caloric converter model; according to each matching node in the first matching node set, acquiring corresponding node heat, wherein the obtained corresponding node heat forms a matching node heat set, the node heat refers to access heat associated with each matching node in the first matching node set, and the output of the node heat set can help a system to perform heat-based mapping or decision in subsequent processing so as to better meet a cooling demand target;
according to the heat predictor model, fusion prediction is carried out on the heat set of the matched node, and fusion prediction output is carried out on prediction heat information of the first task instruction; and learning the predicted heat information as input data of the heat conversion sub-model, and outputting a first cooling control parameter, wherein the first cooling control parameter is a cooling control parameter of a data center by using the heat temperature predicted by the heat of the cooling integrated prediction model, and by using the cooling integrated prediction model, proper cooling control parameters can be predicted according to the heat condition of the matched node so as to realize intelligent cooling operation. Therefore, the targeted cooling control can be performed according to the heat conditions of different nodes, and the efficiency and the energy-saving effect of the system are improved.
Step S64 includes the steps of:
s641: judging whether the node heat stability of the matched node heat set is greater than or equal to the preset heat stability, and if the node heat stability of the matched node heat set is greater than or equal to the preset heat stability, presetting a first fusion weight layer;
S642-A: wherein the weight distribution difference of the first fused weight layer is smaller than theta 1 And 0.ltoreq.θ 1 ≤0.5；
S643-A: and carrying out fusion prediction on the matched node heat set according to the first fusion weight layer.
The embodiment of the application also comprises the following steps:
S642-B: if the node heat stability of the matched node heat set is smaller than the preset heat stability and larger than the preset heat stability, presetting a second fusion weight layer, wherein the weight distribution difference of the second fusion weight layer is larger than theta 1 ；
S643-B carries out fusion prediction on the matched node heat set according to the second fusion weight layer.
Specifically, according to the heat predictor model, fusion prediction is performed on the matched node heat set, which includes that firstly, node heat stability of the matched node heat set is calculated to determine whether validity of a preset heat stability threshold is met, and the node heat stability can be measured by calculating variance or standard deviation of node heat;
judging whether the node heat stability of the matched node heat set is greater than or equal to the preset heat stability: in the first case, if the node heat stability of the matching node heat set is greater than or equal to the preset heat stability, a first fusion weight layer is preset, where the first fusion weight layer refers to weighted average of heat values of different nodes to obtain weight distribution of fused heat values, and a form of the weight distribution may be selected according to specific situations, for example, uniform distribution or gaussian distribution is used, where a weight distribution difference of the first fusion weight layer is smaller than θ 1 And 0.ltoreq.θ 1 ≤0.5，θ 1 The weight distribution difference threshold value is used for representing the preset weight distribution difference threshold value and is custom set by a person skilled in the art; fusion prediction is carried out on the matched node heat set according to the first fusion weight layer: fusion prediction can be carried out on the heat sets of the matched nodes by using a first fusion weight layer, and the fusion prediction can be realized by multiplying heat values of different nodes by corresponding weights and then summing the heat values; the fused heat value can be used as an output result for further control parameter calculation or other purposes;
judging the said matchWhether the node heat stability of the node heat set is greater than or equal to the preset heat stability: in the second case, if the node heat stability of the matching node heat set is smaller than the preset heat stability and larger than the preset heat stability, presetting a second fused weight layer, and calculating the weight distribution difference of the second fused weight layer by using some difference measurement method, such as KL (Kullback-Leibler) divergence, euclidean distance, etc., where the weight distribution difference of the second fused weight layer is larger than θ 1 The method comprises the steps of carrying out a first treatment on the surface of the And carrying out fusion prediction on the matched node heat set according to the second fusion weight layer, and improving the satisfaction degree of the system on the cooling demand target by judging the stability of the node heat and selecting proper weight distribution.
The embodiment of the application also comprises the following steps:
s651: the cooling control system is connected with the storage unit and used for acquiring cooling liquid information, cooling flow rate and heat loss data in the cooling control system;
s652: training by taking the cooling liquid information as a fixed variable and taking the cooling flow rate and the heat loss data as variables, and outputting a heat conversion sub-model for identifying a rate-heat conversion relation;
s653: and inputting the predicted heat information into the heat conversion submodel, and outputting the first cooling control parameter.
Specifically, data collection is carried out, and a cooling control system connected with the storage unit is used for acquiring cooling liquid information, cooling flow rate and heat loss data in the cooling control system; training with the cooling liquid information as a fixed variable and the cooling flow rate and the heat loss data as variables, and training a heat conversion submodel for identifying a rate-heat conversion relationship to identify a conversion relationship between rate and heat, wherein preferably, the heat conversion submodel base for identifying the rate-heat conversion relationship can be a regression model, a neural network or other suitable algorithm;
and inputting the heat information to be predicted into a trained heat conversion sub-model, and outputting corresponding first cooling control parameters by the heat conversion sub-model according to the input heat information and the learned conversion relation, wherein the first cooling control parameters can be parameters for controlling the flow rate of cooling liquid or other related cooling control systems, so that the optimization of the cooling control system is realized.
As shown in fig. 3, the embodiment of the present application further includes:
s71: judging whether the processing time length of the first task instruction is longer than a preset processing time length or not;
s72: if the processing time of the first task instruction is longer than the preset processing time, acquiring node processing time sequence information corresponding to the first matched node set under the first task instruction;
s73: taking the first cooling control as a parameter as a cooling demand target, mapping the first cooling control according to the node processing time sequence information corresponding to the first matching node set and the matching node heat set corresponding to the first matching node set, and outputting cooling control parameters based on time sequence mapping;
s74: and smoothing the cooling control parameters, and outputting dynamic cooling control parameters based on time sequence mapping.
Specifically, the control performed by feeding back the first cooling control parameter to the cooling control system is instantaneous cooling control, if long-time cooling control is required, dynamic control is required to be correspondingly performed according to the change of the flow node, and based on the control, whether the processing duration of the first task instruction is longer than the preset processing duration is judged; if the processing time of the first task instruction is longer than the preset processing time, the cooling control system cannot finish the cooling task within a specified time, corresponding adjustment is needed, node processing time sequence information corresponding to the first matching node set under the first task instruction is obtained, and the node processing time sequence information can provide information about the time and sequence of node processing;
taking the first cooling control as a parameter as a cooling demand target, dynamically adjusting the first cooling control parameter according to node processing time sequence information corresponding to the first matched node set so as to adapt to the time change of node processing, and based on the first cooling control parameter, dynamically controlling the cooling control system according to the node processing time sequence information so as to realize long-time cooling control;
according to the node processing time sequence information corresponding to the first matched node set and the matched node heat set, mapping is carried out, cooling control parameters based on time sequence mapping are output, smoothing processing is carried out on the cooling control parameters based on the time sequence mapping, so that dynamic cooling control parameters based on the time sequence mapping are obtained, and the cooling control system can be more efficient and adapt to different working conditions through mapping and smoothing processing; the cooling control parameters are dynamically adjusted according to the node processing time sequence information, the problem of long-time cooling control is solved, and meanwhile, the dynamic cooling control parameters based on time sequence mapping can improve the response capability of the system.
In summary, the intelligent cooling method for the data center storage unit provided by the embodiment of the application has the following technical effects:
1. because the raw connection storage unit is adopted, a data processing sample set is obtained; carrying out instruction control flow analysis and data access heat analysis by using access instructions corresponding to each sample data, outputting a plurality of storage flows and a plurality of access heats, and constructing a storage instruction tree; when the storage unit receives a first task instruction, traversing a storage instruction tree, and outputting a first matched node set; the intelligent cooling method for the data center storage unit is used for realizing transformation of cold and hot data storage in the data center storage unit, carrying out heat prediction, converting the cooling control parameters of the data center according to the predicted heat temperature, improving the response speed and efficiency of cooling control and carrying out the technical effect of cooling adaptive control.
2. Judging whether the processing time is longer than the preset processing time or not is adopted; if the processing time length is longer than the preset processing time length, acquiring node processing time sequence information corresponding to a first matched node set under a first task instruction; the first cooling control parameter is used as a cooling demand target, mapping is carried out with a matched node heat set according to node processing time sequence information, cooling control parameters are output and smoothing processing is carried out, dynamic cooling control parameters based on time sequence mapping are output, and the cooling control system can be more efficient and adapt to different working conditions through mapping and smoothing processing; the cooling control parameters are dynamically adjusted according to the node processing time sequence information, the problem of long-time cooling control is solved, and meanwhile, the dynamic cooling control parameters based on time sequence mapping can improve the response capability of the system.
Example two
Based on the same inventive concept as the intelligent cooling method of a data center storage unit in the foregoing embodiment, as shown in fig. 4, an embodiment of the present application provides an intelligent cooling system of a data center storage unit, where the system includes:
the data processing sample set obtaining module 100 is configured to connect to a storage unit of a first data center, and obtain a data processing sample set of the storage unit, where the data processing sample set includes an access instruction corresponding to each sample data and an access frequency corresponding to each sample data;
the storage flow output module 200 is configured to perform instruction control flow analysis according to the access instruction corresponding to each sample data, and output a plurality of storage flows;
the access heat output module 300 is configured to perform data access heat analysis according to the access frequency corresponding to each sample data, and output a plurality of access heats;
the storage instruction tree building module 400 is configured to divide tree nodes according to the multiple storage flows, identify the tree nodes according to the multiple access hotness, and build a storage instruction tree;
the first matching node set output module 500 is configured to output a first matching node set when the storage unit receives a first task instruction, and traverse the storage instruction tree according to the first task instruction;
and the cooling control module 600 is used for connecting the cooling control system of the storage unit, performing integrated learning according to the first matching node set and the cooling integrated prediction model, outputting a first cooling control parameter, and feeding back to the cooling control system for control.
Further, the system includes:
the first matching node set acquisition module is used for acquiring a first matching node set, wherein the first matching node set comprises operation flow nodes obtained by matching after traversing the storage instruction tree under the first task instruction;
the first matched node set input module is used for inputting the first matched node set into the cooling integrated prediction model, and the cooling integrated prediction model comprises a heat prediction sub-model and a heat conversion sub-model;
the matching node heat set output module is used for acquiring the node heat corresponding to each matching node in the first matching node set and outputting a matching node heat set;
the predicted heat information output module is used for carrying out fusion prediction on the matched node heat set according to the heat predictor model and outputting predicted heat information for identifying the first task instruction;
and the first cooling control parameter output module is used for learning the predicted heat information as input data of the heat conversion submodel and outputting a first cooling control parameter.
Further, the system includes:
the cooling related index acquisition module is used for connecting a cooling control system of the storage unit and acquiring cooling liquid information, cooling flow rate and heat loss data in the cooling control system;
the heat conversion sub-model output module is used for training by taking the cooling liquid information as a fixed variable and taking the cooling flow rate and the heat loss data as variables, and outputting a heat conversion sub-model for identifying a rate-heat conversion relation;
and the first cooling control parameter output module is used for inputting the predicted heat information into the heat conversion submodel and outputting the first cooling control parameter.
Further, the system includes:
the difference analysis module is used for carrying out difference analysis according to the access frequency of each sample data to obtain a difference index set;
the difference index clustering result output module is used for clustering according to the coordinate distribution of the difference index set and outputting a difference index clustering result;
and the access heat output module is used for identifying the heat of each sample data according to the difference index clustering result and outputting a plurality of access heats.
Further, the system includes:
the first judging module is used for judging whether the node heat stability of the matched node heat set is greater than or equal to the preset heat stability, and presetting a first fusion weight layer if the node heat stability of the matched node heat set is greater than or equal to the preset heat stability;
a first case analysis module, configured to, in the first fused weight layer, have a weight distribution difference less than θ 1 And 0.ltoreq.θ 1 ≤0.5；
And the first fusion prediction module is used for carrying out fusion prediction on the matched node heat set according to the first fusion weight layer.
Further, the system includes:
a second condition analysis module, configured to preset a second fused weight layer if the node heat stability of the matched node heat set is less than the preset heat stability and greater than the preset heat stability, where the weight distribution difference of the second fused weight layer is greater than θ 1 ；
And the second fusion prediction module is used for carrying out fusion prediction on the matched node heat set according to the second fusion weight layer.
Further, the system includes:
the second judging module is used for judging whether the processing time length of the first task instruction is longer than a preset processing time length or not;
the node processing time sequence information acquisition module is used for acquiring node processing time sequence information corresponding to the first matched node set under the first task instruction if the processing time length of the first task instruction is longer than the preset processing time length;
the cooling control parameter output module is used for taking the first cooling control as a parameter as a cooling demand target, processing time sequence information according to the nodes corresponding to the first matching node set, mapping the heat set of the matching nodes corresponding to the first matching node set, and outputting cooling control parameters based on time sequence mapping;
and the smoothing processing module is used for carrying out smoothing processing on the cooling control parameters and outputting dynamic cooling control parameters based on time sequence mapping.
Any of the steps of the methods described above may be stored as computer instructions or programs in a non-limiting computer memory and may be called by a non-limiting computer processor to identify any method for implementing an embodiment of the present application, without unnecessary limitations.
Further, the first or second element may not only represent a sequential relationship, but may also represent a particular concept, and/or may be selected individually or in whole among a plurality of elements. It will be apparent to those skilled in the art that various modifications and variations can be made to the present application without departing from the scope of the application. Thus, the present application is intended to include such modifications and alterations insofar as they come within the scope of the application or the equivalents thereof.
Claims (8)
1. A method for intelligent cooling of a data center storage unit, the method comprising:
the method comprises the steps of connecting a storage unit of a first data center, and acquiring a data processing sample set of the storage unit, wherein the data processing sample set comprises access instructions corresponding to each sample data and access frequencies corresponding to each sample data;
carrying out instruction control flow analysis by using the access instructions corresponding to the sample data, and outputting a plurality of storage flows;
performing data access heat analysis according to the access frequency corresponding to each sample data, and outputting a plurality of access heats;
dividing tree nodes according to the multiple storage flows, identifying the tree nodes according to the multiple access hotness, and building a storage instruction tree;
when the storage unit receives a first task instruction, traversing the storage instruction tree according to the first task instruction, and outputting a first matched node set;
and the cooling control system is connected with the storage unit, performs integrated learning according to the first matching node set and the cooling integrated prediction model, outputs a first cooling control parameter, and feeds back the first cooling control parameter to the cooling control system for control.
2. The method of claim 1, wherein the first cooling control parameter is output based on the first set of matched nodes and a cooling integrated prediction model for integrated learning, the method comprising:
acquiring a first matching node set, wherein the first matching node set comprises operation flow nodes obtained by matching after traversing the storage instruction tree based on the first task instruction;
inputting the first set of matching nodes into the cooling integrated prediction model, the cooling integrated prediction model comprising a thermal predictor model and a thermal converter model;
acquiring node heat corresponding to each matching node in the first matching node set, and outputting a matching node heat set;
according to the heat predictor model, fusion prediction is carried out on the heat set of the matched node, and prediction heat information for identifying the first task instruction is output;
and learning the predicted heat information as input data of the heat conversion submodel, and outputting a first cooling control parameter.
3. The method of claim 2, wherein the method further comprises:
the cooling control system is connected with the storage unit and used for acquiring cooling liquid information, cooling flow rate and heat loss data in the cooling control system;
training by taking the cooling liquid information as a fixed variable and taking the cooling flow rate and the heat loss data as variables, and outputting a heat conversion sub-model for identifying a rate-heat conversion relation;
and inputting the predicted heat information into the heat conversion submodel, and outputting the first cooling control parameter.
4. The method of claim 1, wherein the data access heat analysis is performed at an access frequency corresponding to the respective sample data, the method comprising:
performing differential analysis according to the access frequency of each sample data to obtain a differential index set;
clustering is carried out according to the coordinate distribution of the difference index set, and a difference index clustering result is output;
and marking the cold and hot degrees of each sample data according to the difference index clustering result, and outputting a plurality of access hot degrees.
5. The method of claim 2, wherein fusion predicting the set of matching node hotspots according to the thermal predictor model comprises:
judging whether the node heat stability of the matched node heat set is greater than or equal to the preset heat stability, and if the node heat stability of the matched node heat set is greater than or equal to the preset heat stability, presetting a first fusion weight layer;
wherein the weight distribution difference of the first fused weight layer is smaller than theta 1 And 0.ltoreq.θ 1 ≤0.5；
And carrying out fusion prediction on the matched node heat set according to the first fusion weight layer.
6. The method of claim 5, wherein the method further comprises:
if the node heat stability of the matched node heat set is smaller than the preset heat stability and larger than the preset heat stability, presetting a second fusion weight layer, wherein the weight distribution difference of the second fusion weight layer is larger than theta 1 ；
And carrying out fusion prediction on the matched node heat set according to the second fusion weight layer.
7. The method of claim 1, wherein the method further comprises:
judging whether the processing time length of the first task instruction is longer than a preset processing time length or not;
if the processing time of the first task instruction is longer than the preset processing time, acquiring node processing time sequence information corresponding to the first matched node set under the first task instruction;
taking the first cooling control as a parameter as a cooling demand target, mapping the first cooling control according to the node processing time sequence information corresponding to the first matching node set and the matching node heat set corresponding to the first matching node set, and outputting cooling control parameters based on time sequence mapping;
and smoothing the cooling control parameters, and outputting dynamic cooling control parameters based on time sequence mapping.
8. An intelligent cooling system for a data center storage unit, for implementing the intelligent cooling method for a data center storage unit of any one of claims 1-7, comprising:
the data processing sample set acquisition module is used for connecting a storage unit of the first data center, and acquiring a data processing sample set of the storage unit, wherein the data processing sample set comprises access instructions corresponding to each sample data and access frequencies corresponding to each sample data;
the storage flow output module is used for carrying out instruction control flow analysis by using the access instructions corresponding to the sample data and outputting a plurality of storage flows;
the access heat output module is used for carrying out data access heat analysis according to the access frequency corresponding to each sample data and outputting a plurality of access heats;
the storage instruction tree building module is used for dividing tree nodes according to the storage flows, marking the tree nodes according to the access hotness and building a storage instruction tree;
the first matching node set output module is used for traversing the storage instruction tree according to the first task instruction when the storage unit receives the first task instruction, and outputting a first matching node set;
and the cooling control module is used for connecting the cooling control system of the storage unit, performing integrated learning according to the first matched node set and the cooling integrated prediction model, outputting a first cooling control parameter, and feeding back to the cooling control system for control.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310700589.5A CN116595383A (en) | 2023-06-14 | 2023-06-14 | Intelligent cooling method for storage unit of data center |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310700589.5A CN116595383A (en) | 2023-06-14 | 2023-06-14 | Intelligent cooling method for storage unit of data center |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116595383A true CN116595383A (en) | 2023-08-15 |
Family
ID=87589984
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310700589.5A Pending CN116595383A (en) | 2023-06-14 | 2023-06-14 | Intelligent cooling method for storage unit of data center |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116595383A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117066496A (en) * | 2023-10-17 | 2023-11-17 | 南通盟鼎新材料有限公司 | Casting cooling control method and system |
-
2023
- 2023-06-14 CN CN202310700589.5A patent/CN116595383A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117066496A (en) * | 2023-10-17 | 2023-11-17 | 南通盟鼎新材料有限公司 | Casting cooling control method and system |
| CN117066496B (en) * | 2023-10-17 | 2024-01-23 | 南通盟鼎新材料有限公司 | Casting cooling control method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107392368B (en) | Meteorological forecast-based office building dynamic heat load combined prediction method | |
| JP2020040119A (en) | Parameter proposal method of soldering process | |
| CN107025334B (en) | Method and device for establishing overall dynamic model of heat users of central heating system | |
| CN116595383A (en) | Intelligent cooling method for storage unit of data center | |
| CN110380444B (en) | Capacity planning method for distributed wind power orderly access to power grid under multiple scenes based on variable structure Copula | |
| CN116882708B (en) | Steel process flow control method and device based on digital twin and related equipment | |
| CN115685941B (en) | Machine room operation regulation and control method and device based on machine cabinet hot spot temperature prediction | |
| CN103345660A (en) | Cold load prediction method based on support vector machine parameters optimized with cat swarm algorithm | |
| CN112884012A (en) | Building energy consumption prediction method based on support vector machine principle | |
| CN117234301A (en) | Server thermal management method based on artificial intelligence | |
| CN114970358A (en) | Data center energy efficiency optimization method and system based on reinforcement learning | |
| CN111652422A (en) | Heat supply system load prediction method, device and system based on building classification | |
| CN117066496B (en) | Casting cooling control method and system | |
| CN116820214B (en) | Intelligent regulation and control method and system for ventilation and heat dissipation of server | |
| CN116955963B (en) | Heating ventilation energy-saving ladder optimizing control method based on historical data analysis | |
| CN116358107B (en) | Temperature-sensing air conditioner weight relation calculation method based on pre-regulation and control | |
| CN111083201B (en) | Energy-saving resource allocation method for data-driven manufacturing service in industrial Internet of things | |
| CN114992772B (en) | Method, device and storage medium for evaluating adjustable potential of air conditioner temperature control load cluster | |
| CN117332288A (en) | Building energy saving technology evaluation system and method | |
| JP6607264B2 (en) | Optimal scheduling and real-time control method for x management system | |
| Chanpiwat et al. | Using cluster analysis and dynamic programming for demand response applied to electricity load in residential homes | |
| Schachinger et al. | Adaptive learning-based time series prediction framework for building energy management | |
| Wen et al. | Development and validation of online parameter estimation for HVAC systems | |
| CN111933226A (en) | Artificial intelligence material screening method based on classification model | |
| Kan et al. | Multi-zone building control system for energy and comfort management |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |