CN117170851B - Task processing method in low power consumption state and data center - Google Patents
Task processing method in low power consumption state and data center Download PDFInfo
- Publication number
- CN117170851B CN117170851B CN202311422511.8A CN202311422511A CN117170851B CN 117170851 B CN117170851 B CN 117170851B CN 202311422511 A CN202311422511 A CN 202311422511A CN 117170851 B CN117170851 B CN 117170851B
- Authority
- CN
- China
- Prior art keywords
- task
- power consumption
- value
- data center
- heat dissipation
- 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.)
- Active
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 23
- 230000017525 heat dissipation Effects 0.000 claims description 108
- 230000006870 function Effects 0.000 claims description 34
- 238000012545 processing Methods 0.000 claims description 31
- 238000000034 method Methods 0.000 claims description 30
- 230000007613 environmental effect Effects 0.000 claims description 26
- 238000004458 analytical method Methods 0.000 claims description 25
- 238000004146 energy storage Methods 0.000 abstract description 29
- 238000005265 energy consumption Methods 0.000 abstract description 14
- 238000004364 calculation method Methods 0.000 description 5
- 239000000126 substance Substances 0.000 description 5
- 230000005540 biological transmission Effects 0.000 description 4
- 238000006243 chemical reaction Methods 0.000 description 3
- 230000003111 delayed effect Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- 238000012163 sequencing technique Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000033228 biological regulation Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000004134 energy conservation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000000053 physical method Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000008929 regeneration Effects 0.000 description 1
- 238000011069 regeneration method Methods 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
Classifications
-
- 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
- Power Sources (AREA)
Abstract
The invention provides a task processing method with low power consumption and a data center, comprising the following steps: acquiring tasks of a data center in real time; classifying and sorting the acquired tasks according to a preset task execution strategy; switching the data center into a low-power consumption state according to preset conditions; and when the data center is in a low-power consumption state, executing tasks according to the sorting result. By sorting the tasks and executing the tasks in a low-power consumption state, the task execution efficiency can be effectively improved, meanwhile, the energy consumption can be saved, the energy storage loss caused by long-time unused energy storage equipment can be avoided, and the problem that the energy consumption of the flywheel energy storage system used by the existing data center is large is solved.
Description
Technical Field
The invention relates to the technical field of energy storage and power supply of data centers, in particular to a task processing method with low power consumption and a data center.
Background
Data centers are globally coordinated, specific equipment networks used to communicate, accelerate, display, calculate, store data information over an internet network infrastructure. In order to ensure smooth and efficient operation of the data center, it is necessary to continuously provide stable electrical power to the data center. However, in the prior art, the data center has a great potential safety hazard in electricity consumption, because the data center generally adopts a chemical battery as a standby power supply at present, the chemical battery not only has the problem of battery leakage, but also can cause the service life attenuation due to the increase of the using time.
In order to avoid potential safety hazards caused by chemical batteries, research and development personnel propose using a flywheel energy storage system as a standby power supply, for example, a patent with application number 202210841196.1 and name of 'a regulation and control system of a green standby power supply for a data center', and provide a standby power supply system using flywheel energy storage. The flywheel energy storage means that the motor is used for driving the flywheel to rotate at high speed, and the flywheel is used for driving the generator to generate electricity when the flywheel is needed.
Although the flywheel energy storage system utilizes a physical method (electromechanical energy conversion) to realize energy storage, the limitation of a chemical battery is broken through, and the potential safety hazard caused by the chemical battery is avoided, the flywheel energy storage system has larger energy consumption in the electromechanical energy conversion process, and is not beneficial to energy conservation and environmental protection.
Disclosure of Invention
The invention aims to provide a task processing method with low power consumption and a data center, which at least solve the problem that the existing data center has larger energy consumption by using a flywheel energy storage system.
In order to solve the above technical problems, the present invention provides a task processing method in a low power consumption state, including:
acquiring tasks of a data center in real time;
Classifying and sorting the acquired tasks according to a preset task execution strategy;
switching the data center into a low-power consumption state according to preset conditions;
and when the data center is in a low-power consumption state, executing tasks according to the sorting result.
Optionally, in the low-power-consumption task processing method, the method for sorting the acquired tasks according to a preset task execution policy includes:
configuring a task type index table, a task analysis algorithm and a task sequence algorithm;
calculating a task priority value and a sequence priority value by using the task analysis algorithm and the task sequence algorithm;
storing the task with the task priority value lower than a preset priority reference value as a secondary task into a secondary task temporary storage area;
and adjusting the execution order of the secondary tasks stored in the secondary task temporary storage area by using the sequence priority value.
Optionally, in the low-power-consumption task processing method, the method for configuring a task type index table, a task analysis algorithm and a task sequence algorithm includes:
setting a task type index table, wherein task emergency values corresponding to different task types are stored in the task type index table;
The configuration task analysis algorithm is as follows:
wherein,a task priority value; />For a preset target time parameter, +.>The task type parameter is preset;a task target period corresponding to the task; />For the target completion time corresponding to the task, +.>Is the current moment; />A task urgency value for the task;
the configuration task sequence algorithm is as follows:
wherein,is a sequence priority value; />The task amount corresponding to the secondary task.
Optionally, in the task processing method in the low power consumption state, the method for switching the data center to the low power consumption state according to the preset condition includes:
configuring a switching reference threshold, a load prediction algorithm, a power consumption matching algorithm, an environment redundancy table and a power consumption supporting algorithm;
calculating a task load value by using the load prediction algorithm, calculating a power consumption matching value by using the power consumption matching algorithm, and calculating environmental heat dissipation efficiency by using the environmental redundancy table;
obtaining the heat dissipation efficiency of the equipment according to the heat dissipation equipment sub-efficiency of each heat dissipation equipment;
calculating a comprehensive switching value according to a power consumption supporting algorithm by using the task load value, the power consumption matching value, the environment heat dissipation efficiency and the equipment heat dissipation efficiency;
and if the comprehensive switching value is higher than the switching reference threshold value, switching the data center into a low-power consumption state.
Optionally, in the low-power-consumption task processing method, the method for configuring the switching reference threshold, the load prediction algorithm, the power consumption matching algorithm, the environmental redundancy table and the power consumption supporting algorithm includes:
the configuration load prediction algorithm is as follows:
wherein,is a task load value; />A historical load function corresponding to the i-th historical load data with the corresponding time characteristic; />As the i-th historical load functionSimilarity with the time characteristics of a future preset time period; n is the total number of historical load functions; />For the start time of the ith history load function, < +.>For the current moment +.>Presetting a time period for the future;
the configuration power consumption matching algorithm is as follows:
wherein,matching the power consumption with a value; />For a data center power consumption function within a previously preset period of time,/-, a data center power consumption function within a previously preset period of time>A data center power consumption function corresponding to the j-th historical power consumption data with the corresponding time characteristic; />For a previously preset period of time->Starting time of the power consumption function of the j-th data center; m is the total number of data center power consumption functions;
setting an environment redundancy table, wherein the environment redundancy table stores environment radiator values corresponding to different environment information;
the configuration power consumption support algorithm is:
Wherein,is the comprehensive switching value; />For a preset task load weight, +.>The weights are matched for a preset power consumption,is the preset heat dissipation efficiency weight->；/>Is a preset load reference value; />For the heat dissipation efficiency of the environment, the air conditioner is->For a preset ambient heat dissipation reference value, +.>For the heat dissipation efficiency of the device, < > for>For a preset device heat dissipation reference value, +.>And the heat dissipation safety reference value is a preset data center heat dissipation safety reference value.
Optionally, in the low-power-consumption-state task processing method, when the data center is in a low power consumption state, the method for executing the task according to the classification and sequencing result includes:
acquiring an idle value of a server and a task total value of a secondary task temporary storage area;
generating an equilibrium distribution instruction according to the sorting result so as to minimize the average difference between the idle value and the task total value;
and executing the task according to the balanced distribution instruction, and mirroring the task to a local temporary storage execution area.
In order to solve the above technical problems, the present invention further provides a data center, which uses the task processing method in a low power consumption state as described in any one of the above, the data center including:
the task management subsystem is used for acquiring tasks of the data center in real time, and sorting and executing the tasks according to a preset task execution strategy;
The low-consumption switching subsystem is used for calculating a task load value, a power consumption matching value, an environment heat dissipation efficiency and an equipment heat dissipation efficiency, and judging whether to switch the data center into a low-power consumption state according to the task load value, the power consumption matching value, the environment heat dissipation efficiency and the equipment heat dissipation efficiency;
and the low-consumption execution subsystem works when the data center is in a low-consumption state, and is used for starting the flywheel power supply to supply power to the data center and controlling the execution of the task according to the task balancing strategy.
Optionally, in the data center, the task management subsystem includes a task configuration module and a secondary task temporary storage area, which are disposed in a server; the task configuration module comprises a task analysis unit, a task management unit and a task execution unit; the task analysis unit is used for acquiring tasks of the data center in real time, calculating task priority values of the tasks, and storing the tasks with the task priority values lower than preset priority reference values as secondary tasks into the secondary task temporary storage area; the task management unit is used for calculating the sequence priority value of the secondary task in the secondary task temporary storage area in real time so as to adjust the execution sequence of the secondary task; and the task execution unit is used for sequentially calling and executing the tasks in the secondary task temporary storage area according to the execution sequence when the idle value of the server is larger than a preset idle reference value.
Optionally, in the data center, the low-consumption switching subsystem includes a load prediction module, a power consumption matching module, an environmental redundancy module, a heat dissipation support module, and a low-consumption switching module; the load prediction module is used for calculating a task load value in a preset time period in the future; the power consumption matching module is used for calculating a power consumption matching value in a preset time period before; the environment redundancy module is used for calculating the environment heat dissipation efficiency in a preset time period in the future; the heat dissipation support module is used for calculating the heat dissipation efficiency of the equipment in a future preset time period; the low-consumption switching module is used for calculating to obtain a comprehensive switching value according to the task load value, the power consumption matching value, the environment heat dissipation efficiency and the equipment heat dissipation efficiency, and judging whether to switch the data center into a low-power consumption state according to the comprehensive switching value.
Optionally, in the data center, the low-consumption execution subsystem includes a flywheel management module and a task balancing module; the flywheel management module is used for starting a flywheel power supply to supply power for the data center; the task balancing module is used for controlling the execution of the task according to the task balancing strategy.
The invention provides a task processing method with low power consumption and a data center, comprising the following steps: acquiring tasks of a data center in real time; classifying and sorting the acquired tasks according to a preset task execution strategy; switching the data center into a low-power consumption state according to preset conditions; and when the data center is in a low-power consumption state, executing tasks according to the sorting result. By sorting the tasks and executing the tasks in a low-power consumption state, the task execution efficiency can be effectively improved, meanwhile, the energy consumption can be saved, the energy storage loss caused by long-time unused energy storage equipment can be avoided, and the problem that the energy consumption of the flywheel energy storage system used by the existing data center is large is solved.
Drawings
FIG. 1 is a flow chart of a task processing method in a low power consumption state according to the present embodiment;
fig. 2 is a schematic structural diagram of a data center according to the present embodiment;
fig. 3 is a diagram of an example of a data center according to the present embodiment.
Detailed Description
The task processing method and the data center with low power consumption state provided by the invention are further described in detail below with reference to the accompanying drawings and the specific embodiments. It should be noted that the drawings are in a very simplified form and are all to a non-precise scale, merely for convenience and clarity in aiding in the description of embodiments of the invention. Furthermore, the structures shown in the drawings are often part of actual structures. In particular, the drawings are shown with different emphasis instead being placed upon illustrating the various embodiments.
It is noted that "first", "second", etc. in the description and claims of the present invention and the accompanying drawings are used to distinguish similar objects so as to describe embodiments of the present invention, and not to describe a specific order or sequence, it should be understood that the structures so used may be interchanged under appropriate circumstances. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
The embodiment provides a task processing method in a low power consumption state, as shown in fig. 1, including:
s1, acquiring tasks of a data center in real time;
s2, sorting the acquired tasks according to a preset task execution strategy;
s3, switching the data center into a low-power consumption state according to preset conditions;
s4, executing tasks according to the sorting result when the data center is in a low-power consumption state.
According to the task processing method in the low power consumption state, the tasks are classified and ordered, and the tasks are executed in the low power consumption state, so that the task execution efficiency can be effectively improved, the energy consumption can be saved, the energy storage loss caused by long-time non-use of the energy storage equipment can be avoided, and the problem that the energy consumption of the flywheel energy storage system used by the existing data center is large is solved.
Specifically, in this embodiment, the tasks of the data center have different task load values, task types, task target periods, task emergency values and the like according to task content, and when the tasks of the data center are acquired in real time in step S1, the task load values, task types, task target periods, task emergency values and the like corresponding to the tasks may also be acquired according to actual situations, so that subsequent operation is facilitated. Methods for acquiring tasks of a data center in real time are well known to those skilled in the art and will not be described in detail herein.
Further, in this embodiment, step S2, the method for classifying and sorting the acquired tasks according to the preset task execution policy includes:
s21, configuring a task type index table, a task analysis algorithm and a task sequence algorithm.
Specifically, in this embodiment, the method for configuring the task type index table, the task analysis algorithm and the task sequence algorithm includes:
setting a task type index table, wherein task emergency values corresponding to different task types are stored in the task type index table;
the configuration task analysis algorithm is as follows:
wherein,a task priority value; />For a preset target time parameter, +. >The task type parameter is preset;a task target period corresponding to the task; />For the target completion time corresponding to the task, +.>Is the current moment; />A task urgency value for the task;
the configuration task sequence algorithm is as follows:
wherein,is a sequence priority value; />The task amount corresponding to the secondary task.
S22, calculating a task priority value and a sequence priority value by using the task analysis algorithm and the task sequence algorithm.
Specifically, in this embodiment, according to the obtained information such as the task type and the task target period of the task, the corresponding task emergency value is called from the task type index table by taking the task type as the index; and then obtaining a task priority value according to a task analysis algorithm, and obtaining a sequence priority value according to a task sequence algorithm. The numerical values of the parameters required in the task analysis algorithm and the task sequence algorithm configured in this embodiment can be obtained from task acquisition.
S23, storing the task with the task priority value lower than a preset priority reference value as a secondary task into a secondary task temporary storage area.
Specifically, in this embodiment, the pre-reference value may be set according to the actual situation of the task processing, or a reasonable value may be selected empirically. Of course, it is preferable to introduce a deep network learning model, by which the pre-reference value is optimized by using the historical data, and specific implementation is well known to those skilled in the art, and will not be described herein.
The secondary task registry may be located within the server. Secondary tasks refer to tasks that may delay processing, and in particular may be less important tasks, less urgent tasks, and so on. For tasks other than the secondary tasks, the data center can directly execute, and the data center is not in a low-power consumption state and does not use a standby power supply (flywheel power supply). The present application focuses on the processing execution of secondary tasks when the data center is in a low power state, so processing execution of non-secondary tasks is not described in great detail, and those skilled in the art can know from the prior art.
S24, adjusting the execution sequence of the secondary tasks stored in the secondary task temporary storage area by using the sequence priority value.
Specifically, in this embodiment, the task with the top sequence priority value will be preferentially executed. As can be seen from a task sequence algorithm, when the tasks are ordered, the task priority value and the task quantity are considered, namely the target period of the tasks, the emergency situation of the tasks and the task quantity are considered, and the tasks with long period, urgent tasks and large task quantity are processed preferentially. Of course, in other embodiments, the weights of the factors in the sequence priority values may be set according to actual needs, so that the execution priority of the task is more suitable for the actual needs.
According to the embodiment, the tasks of the data center are acquired and identified, the delayed processing tasks are marked and temporarily stored, when the data center is low in power consumption, the temporarily stored tasks are processed, the temporarily stored tasks are classified in the processing process, the files are ordered according to the importance degree and the storage size, the processing effect of the files is improved, and the task quantity can be effectively adjusted.
Further, in this embodiment, step S3, the method for switching the data center to the low power consumption state according to the preset condition includes:
s31, configuring a switching reference threshold, a load prediction algorithm, a power consumption matching algorithm, an environment redundancy table and a power consumption supporting algorithm.
Specifically, in this embodiment, the method for configuring the switching reference threshold, the load prediction algorithm, the power consumption matching algorithm, the environmental redundancy table, and the power consumption supporting algorithm includes:
setting a switching reference threshold, wherein the switching reference threshold can be set according to actual needs, and the overall task quantity of the data center, the distribution condition of the task emergency degree, the distribution condition of the task completion period and the like can be comprehensively considered in actual setting;
the configuration load prediction algorithm is as follows:
wherein,is a task load value; / >A historical load function corresponding to the i-th historical load data with the corresponding time characteristic; />Similarity of time characteristics of the ith historical load function and a future preset time period is obtained; n is the total number of historical load functions; />For the start time of the ith history load function, < +.>For the current moment +.>Presetting a time period for the future;
the configuration power consumption matching algorithm is as follows:
wherein,matching the power consumption with a value; />For a data center power consumption function within a previously preset period of time,/-, a data center power consumption function within a previously preset period of time>A data center power consumption function corresponding to the j-th historical power consumption data with the corresponding time characteristic; />For a previously preset period of time->Starting time of the power consumption function of the j-th data center; m is the total number of data center power consumption functions;
setting an environment redundancy table, wherein the environment redundancy table stores environment radiator values corresponding to different environment information;
the configuration power consumption support algorithm is:
wherein,is the comprehensive switching value; />For a preset task load weight, +.>The weights are matched for a preset power consumption,is the preset heat dissipation efficiency weight->；/>Is a preset load reference value; />For the heat dissipation efficiency of the environment, the air conditioner is->For a preset ambient heat dissipation reference value, +.>For the heat dissipation efficiency of the device, < > for >For a preset device heat dissipation reference value, +.>And the heat dissipation safety reference value is a preset data center heat dissipation safety reference value.
S32, calculating a task load value by using the load prediction algorithm, calculating a power consumption matching value by using the power consumption matching algorithm, and calculating the environment heat dissipation efficiency by using the environment redundancy table.
Specifically, in the present embodiment, the historical load function is a function of the relationship between the load value of the data center and time, and different data centers may have different historical load functions. The historical load function and its similarity to the time characteristics of the future preset time period can be obtained through modeling, and specific implementation manners are well known to those skilled in the art, and are not described herein. The same power consumption function is a function of the power consumption value of the data center versus time, and different data centers may have different power consumption functions. The power consumption function may also be obtained through modeling, and the specific implementation is well known to those skilled in the art, and will not be described here again.
Because the environmental redundancy table stores a plurality of environmental heat dissipation sub-values, each environmental heat dissipation sub-value corresponds to environmental information, the embodiment obtains corresponding environmental heat dissipation efficiency by calling environmental prediction information in a preset time period in the future, calling the environmental heat dissipation sub-value at each moment for the environmental prediction information, and performing weighted calculation on the obtained environmental heat dissipation sub-value.
S33, obtaining the heat dissipation efficiency of the device according to the sub-efficiency of the heat dissipation device of each heat dissipation device.
Specifically, in this embodiment, the effective energy of the standby energy storage device (such as a flywheel power supply) in a preset time period in the future may be obtained, and the available heat dissipation devices are determined according to the effective energy, each available heat dissipation device corresponds to a heat dissipation device sub-efficiency, and the summation calculation is performed on all the heat dissipation device sub-efficiencies to obtain the heat dissipation efficiency of the device. When the standby energy storage equipment is a flywheel power supply, the effective energy comprises electric energy and mechanical energy, and the available heat dissipation equipment corresponding to the mechanical energy is a fan connected with the data center.
S34, calculating a comprehensive switching value according to a power consumption supporting algorithm by using the task load value, the power consumption matching value, the environment heat dissipation efficiency and the equipment heat dissipation efficiency.
Specifically, in this embodiment, the task load weight, the power consumption matching weight, and the heat dissipation efficiency weight in the power consumption support algorithm are set according to actual situations; the load reference value, the environment heat dissipation reference value, the equipment heat dissipation reference value and the heat dissipation safety reference value are reasonably set according to actual task related information, environment conditions and equipment conditions. This part is set up to a certain extent depending on the experience of the person.
And S35, if the comprehensive switching value is higher than the switching reference threshold value, switching the data center into a low-power consumption state.
Specifically, in this embodiment, the switching reference threshold needs to be set according to the actual application scenario, so that when the data center is switched to the low power consumption state, the secondary task can be processed and executed. When the flywheel is switched to a low-power consumption state, the data center is powered by a standby power supply (flywheel power supply), so that electric energy can be saved, long-time unused storage loss of the flywheel can be avoided, and pollution-free storage and regeneration of the electric energy are ensured through circulation.
Further, in this embodiment, step S4, when the data center is in a low power consumption state, the method for executing the task according to the result of classification and sequencing includes:
s41, acquiring an idle value of the server and a task total value of the secondary task temporary storage area.
Specifically, in the present embodiment, the server idle value may be obtained based on the duty metering, and the task total value may be obtained based on the task metering.
S42, generating an equilibrium distribution instruction according to the sorting result so as to minimize the average difference between the idle value and the task total value.
Specifically, in this embodiment, a task balancing policy may be preconfigured, and based on the sorting result of the tasks, and in combination with the task balancing policy, a balancing distribution instruction that minimizes the average difference between the idle value and the total task value is generated, that is, the balancing distribution instruction makes the amount of tasks received by each server and the time schedule approximately the same, so that the running power consumption of each server is substantially equal.
S43, executing the task according to the balanced distribution instruction, and mirroring the task to a local temporary storage execution area.
Specifically, in this embodiment, when the server forwards or processes the secondary task, the secondary task is mirrored to the local temporary storage execution area, until the secondary task is processed, and then the task is deleted from the temporary storage execution area.
According to the embodiment, through the task balancing strategy, a plurality of tasks can be balanced in a plurality of servers of the data center, so that the task execution efficiency is improved, and the power supply stability of each server can be ensured.
The embodiment also provides a data center, which uses the task processing method in the low power consumption state as described above, as shown in fig. 2, where the data center includes: the task management subsystem is used for acquiring tasks of the data center in real time, and sorting and executing the tasks according to a preset task execution strategy; the low-consumption switching subsystem is used for calculating a task load value, a power consumption matching value, an environment heat dissipation efficiency and an equipment heat dissipation efficiency, and judging whether to switch the data center into a low-power consumption state according to the task load value, the power consumption matching value, the environment heat dissipation efficiency and the equipment heat dissipation efficiency; and the low-consumption execution subsystem works when the data center is in a low-consumption state, and is used for starting the flywheel power supply to supply power to the data center and controlling the execution of the task according to the task balancing strategy.
According to the data center provided by the embodiment, the tasks of the data center are acquired and identified, the delayed processing tasks are marked and temporarily stored, the temporarily stored tasks are processed when the data center is low in power consumption, the temporarily stored tasks are classified in the processing process, the files are ordered according to the importance degree and the storage size, the processing effect of the files is improved, the task execution efficiency can be effectively improved, the energy consumption can be saved, the energy storage loss caused by long-time unused energy storage equipment can be avoided, and the problem that the flywheel energy storage system used by the existing data center is large in energy consumption is solved.
Specifically, in this embodiment, the task management subsystem includes a task configuration module and a secondary task temporary storage area that are disposed in the server; the task configuration module comprises a task analysis unit, a task management unit and a task execution unit; the task analysis unit is used for acquiring tasks of the data center in real time, calculating task priority values of the tasks, and storing the tasks with the task priority values lower than preset priority reference values as secondary tasks into the secondary task temporary storage area; the task management unit is used for calculating the sequence priority value of the secondary task in the secondary task temporary storage area in real time so as to adjust the execution sequence of the secondary task; and the task execution unit is used for sequentially calling and executing the tasks in the secondary task temporary storage area according to the execution sequence when the idle value of the server is larger than a preset idle reference value.
In this embodiment, the low-consumption switching subsystem includes a load prediction module, a power consumption matching module, an environmental redundancy module, a heat dissipation support module, and a low-consumption switching module; the load prediction module is used for calculating a task load value in a preset time period in the future; the power consumption matching module is used for calculating a power consumption matching value in a preset time period before; the environment redundancy module is used for calculating the environment heat dissipation efficiency in a preset time period in the future; the heat dissipation support module is used for calculating the heat dissipation efficiency of the equipment in a future preset time period; the low-consumption switching module is used for calculating to obtain a comprehensive switching value according to the task load value, the power consumption matching value, the environment heat dissipation efficiency and the equipment heat dissipation efficiency, and judging whether to switch the data center into a low-power consumption state according to the comprehensive switching value.
In this embodiment, the low-consumption execution subsystem includes a flywheel management module and a task balancing module; the flywheel management module is used for starting a flywheel power supply to supply power for the data center; the task balancing module is used for controlling the execution of the task according to the task balancing strategy. The task balancing module comprises a duty metering unit and a task metering unit, wherein the duty metering unit is used for acquiring an idle value of each server, the task metering unit is used for a task total value of a secondary task execution area of each server, the task balancing module is configured with a task balancing strategy, the task balancing strategy acquires the idle value and the task total value of each server in real time, and generates a balancing distribution instruction so that the average difference value between the idle value and the task total value of the server is minimum, and when the balancing distribution instruction is received by the server, the corresponding secondary task is forwarded to other servers of the data center according to the balancing distribution instruction. Preferably, the task balancing module further includes a task mirroring sub-policy, when the server forwards or processes the secondary task, mirroring the secondary task to a local temporary storage execution area until the secondary task is processed, and deleting the corresponding secondary task from the temporary storage execution area.
In the following, referring to fig. 3, a specific embodiment is used to describe a task processing method and a data center in a low power consumption state provided in the present application:
as shown in fig. 3, the data center includes a task management subsystem, a low-consumption switching subsystem, and a low-consumption execution subsystem.
The task management subsystem comprises a task configuration module configured on the server, the task configuration module comprises a task analysis unit and a task management unit, and the task configuration module is pre-configured with a task execution strategy.
The server is configured with a secondary task temporary storage area, the task analysis unit acquires the task to be executed in real time, the task can be divided according to different importance degrees and emergency degrees by utilizing a pre-configured task analysis algorithm through the design of the secondary task temporary storage area, the task priority value of the task to be executed is calculated, and the task to be executed, of which the task priority value is lower than a preset priority reference value, is used as the secondary task to be configured in the secondary task execution area. The task analysis unit can acquire the target period of the task and the target completion time of the task in real time when acquiring the task, judge the emergency degree of the task, and then judge the importance degree of the task in a table look-up mode through the task type.
The task management unit is configured with a task sequence algorithm, so that a sequence priority value of a secondary task of the secondary task execution area is calculated in real time by using the task sequence algorithm to adjust the sequence of the secondary task. In this embodiment, taking into account the sorting value of the task amount, acquiring the task sorting value, where the number of acquired processing tasks is n, and n is a positive integer; setting the first task to be assigned value a 1 Assignment of the nth task to a … … n The first task has a memory size b 1 The storage size of the … … nth task is b n The task ranking value of the first task is pxz 1 Task order value for the nth task, … …, is pxz n The method comprises the steps of carrying out a first treatment on the surface of the The following equation may be used to find the ranking value of the tasks:
where i=1, 2, … …, n.
And arranging the acquired plurality of task ordering values according to descending order, and sequentially carrying out task processing according to the task ordering values in the task processing process.
In this embodiment, when judging the importance distance, by judging the type of each task file, for example, when judging the buffer file and the transmission task, the priority of the transmission task is higher than the priority of the buffer file, and the importance degree of the transmission task is higher than the importance degree of the buffer file; when caching the cache files with different storage sizes, preferentially storing the cache files with small storage size; when the file processing is performed, file transmission is ordered … … according to the file storage size and the assigned size, so that the task amount and the importance degree of the task are comprehensively calculated.
And when the idle value of the server is larger than a preset idle reference value, sequentially calling the tasks of the task execution area according to the task execution strategy and executing the tasks.
The low-consumption switching subsystem is configured with a load prediction module, a power consumption matching module, an environment redundancy module, a heat dissipation support module and a low-consumption switching module.
And the load prediction module calculates a task load value in a preset time period according to a load prediction algorithm. Specifically, the load prediction module obtains a workload prediction value of a data center, obtains a workload of the data center, obtains a processing speed of the workload in a time period, obtains an average value of the processing speed in the time period, obtains a processing time value according to the average value of the processing speed and the workload, records a workload processing time interval through a server, and judges the processing time according to the recording time: if the acquired processing time is within the workload processing time interval, the acquired processing time value is defined as a workload predicted value. The workload can be obtained through calculation of the workload predicted value, and judgment of the load condition is realized. The actual workload can be calculated by recording the workload processing interval of the server through the condition of the historical workload, and the time characteristics are the same time period, such as season, holiday, weekday and time period, if the conditions are the same, the corresponding time characteristics are considered, and the workload with the same time characteristics can be similar, so the corresponding workload can be determined through the analysis.
The power consumption matching module is used for calculating a power consumption matching value in a preset time period. Specifically, the power consumption matching module acquires the working power of the data center, and sets a standard working power value through the server, and if the standard working power value is not larger than the standard working power value, the power consumption matching module judges that the standard working power value meets the expectations; similarly, by calculating the working power value, whether the difference value is in the predicted reference can be judged, if the difference value is larger, the situation that the data requirement of the current user side is unstable is indicated, the situation that the data risk or the load risk possibly occurs at the moment is indicated, the data center is not suitable to be switched to the low-power consumption state, otherwise, if the difference value is relatively close to the previous situation, the situation that the data center is suitable to be switched to the low-power consumption state is indicated.
The environment redundancy module is used for calculating the environment heat dissipation efficiency in a preset time period in the future. Specifically, the environmental redundancy module is configured with an environmental redundancy table, and the environmental redundancy table stores a plurality of environmental heat dissipation sub-values, and each environmental heat dissipation sub-value corresponds to environmental information. The environment redundancy module invokes environment prediction information in a future preset time period through an external database, invokes an environment heat dissipation sub-value at each moment according to the environment prediction information, and performs weighted calculation on the obtained environment heat dissipation sub-value to obtain corresponding environment heat dissipation efficiency. For example, the temperature and the humidity of the data center are obtained through a temperature and humidity sensor, and whether the operating environment conditions are met is judged according to the temperature and the humidity; setting a safe temperature interval and a safe humidity interval according to the running temperature and the humidity value of each component in the data center, wherein the environmental radiator values corresponding to different safe intervals are different; by the arrangement, the corresponding environment heat dissipation efficiency can be judged according to the actual temperature and humidity in a data table calling mode. The corresponding ambient heat dissipation table can be generated according to the prior data and manual configuration.
The heat dissipation support module is used for calculating the heat dissipation efficiency of the equipment in a preset time period in the future. Specifically, the heat dissipation support module obtains effective energy of the flywheel power supply in a preset time period in the future, determines available heat dissipation devices according to the effective energy, each available heat dissipation device corresponds to a heat dissipation device sub-efficiency, and performs summation calculation on the heat dissipation device sub-efficiency to obtain the heat dissipation efficiency of the device. The effective energy comprises electric energy and mechanical energy, and the available heat dissipation equipment corresponding to the mechanical energy is a fan connected with the data center. When the available heat dissipation power of the heat dissipation device is obtained, different working powers can be set according to different types of heat dissipation devices. Because the flywheel is energy-storage realized by electromechanical energy conversion, the embodiment judges equipment capable of being started under the energy supply of the flywheel and corresponding efficiency by counting mechanical energy and electric energy, thereby calculating heat dissipation efficiency.
The low-consumption switching module is used for calculating to obtain a comprehensive switching value according to the task load value, the power consumption matching value, the environment heat dissipation efficiency and the equipment heat dissipation efficiency, and judging whether to switch the data center into a low-power consumption state according to the comprehensive switching value. Specifically, the low-consumption switching module is configured with a power consumption supporting algorithm and a switching reference threshold, and switches the data center to a low-consumption state when the comprehensive switching value is higher than the switching reference threshold. Judging the power consumption of the data center, and if the power consumption of the data center is high, not processing; and if the power consumption of the data center is judged to be low, processing the task for collecting temporary storage. Thus, whether to switch the data center to the low power consumption state can be judged. Under the conditions of lower future processing load, higher load stability and higher heat dissipation efficiency of the data center, the data center is switched to a low-power consumption mode, and flywheel power supply is utilized to achieve the purpose of reducing power consumption.
The low-consumption execution subsystem works in a low-consumption state and comprises a flywheel management module and a task balancing module.
The flywheel management module is used for providing electric energy for the data center through a flywheel power supply.
The task balancing module is configured with a task balancing strategy and comprises a duty metering unit and a task metering unit. The duty metering unit is used for acquiring the idle value of each server; the task metering unit is used for the total task value of the secondary task execution area of each server. The task balancing strategy acquires the idle value and the task total value of each server in real time, generates a balancing distribution instruction so as to minimize the average difference value between the idle value and the task total value of the server, and forwards the corresponding secondary task to other servers of the data center according to the balancing distribution instruction when the server receives the balancing distribution instruction.
The task balancing module further comprises a task mirroring sub-strategy, when the server forwards or processes the secondary task, the secondary task is mirrored to a local temporary storage execution area until the secondary task is processed, and the corresponding secondary task is deleted from the temporary storage execution area.
In addition to using flywheel to supply power, the low-consumption execution subsystem of the embodiment also enables the plurality of servers to jointly process a plurality of tasks through a task balancing algorithm: if the first server analyzes that the task ordering value of the predicted subsequent moment is 3 and the server receives a task with the task ordering value of 5, the task ordering value of the subsequent task is acquired, if the task ordering value is greater than 5, the task with the task ordering value greater than 5 is delayed, and if the task ordering value is less than 5, the task with the task ordering value of 5 is distributed to other servers for processing. Classifying the task quantity, judging the importance degree of each task according to the classification of the task quantity, and assigning a value to each task according to the importance degree, wherein the larger the importance degree is, the larger the value is; the idle value of the server can be obtained by directly invoking the working parameters of the server, the total task value is the total occupation amount of processing resources of all secondary tasks, and the total task value can be obtained by the task type and the task size.
Therefore, through the task processing method and the data center in the low power consumption state, the task execution efficiency can be effectively improved, meanwhile, the energy consumption can be saved, the energy storage loss caused by long-time non-use of the energy storage equipment can be avoided, and the problem that the energy consumption of the flywheel energy storage system used by the existing data center is large is solved.
In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the difference from other embodiments, so that the same similar parts of each embodiment are referred to each other.
The low-power-consumption task processing method and the data center provided by the embodiment comprise the following steps: acquiring tasks of a data center in real time; classifying and sorting the acquired tasks according to a preset task execution strategy; switching the data center into a low-power consumption state according to preset conditions; and when the data center is in a low-power consumption state, executing tasks according to the sorting result. By sorting the tasks and executing the tasks in a low-power consumption state, the task execution efficiency can be effectively improved, meanwhile, the energy consumption can be saved, the energy storage loss caused by long-time unused energy storage equipment can be avoided, and the problem that the energy consumption of the flywheel energy storage system used by the existing data center is large is solved.
The above description is only illustrative of the preferred embodiments of the present invention and is not intended to limit the scope of the present invention, and any alterations and modifications made by those skilled in the art based on the above disclosure shall fall within the scope of the appended claims.
Claims (7)
1. The task processing method in the low power consumption state is characterized by comprising the following steps of:
acquiring tasks of a data center in real time;
classifying and sorting the acquired tasks according to a preset task execution strategy;
switching the data center to a low power consumption state according to preset conditions, wherein the method comprises the following steps:
configuring a switching reference threshold, a load prediction algorithm, a power consumption matching algorithm, an environment redundancy table and a power consumption supporting algorithm, wherein the switching reference threshold, the load prediction algorithm, the power consumption matching algorithm, the environment redundancy table and the power consumption supporting algorithm comprise:
the configuration load prediction algorithm is as follows:
wherein,is a task load value; />A historical load function corresponding to the i-th historical load data with the corresponding time characteristic; />Similarity of time characteristics of the ith historical load function and a future preset time period is obtained; n is the total number of historical load functions; />For the start time of the ith history load function, < +.>For the current moment +.>Presetting a time period for the future;
the configuration power consumption matching algorithm is as follows:
wherein,matching the power consumption with a value; / >For a data center power consumption function within a previously preset period of time,/-, a data center power consumption function within a previously preset period of time>A data center power consumption function corresponding to the j-th historical power consumption data with the corresponding time characteristic; />For a previously preset period of time,is the jth numberStarting time of the power consumption function according to the center; m is the total number of data center power consumption functions;
setting an environment redundancy table, wherein the environment redundancy table stores environment radiator values corresponding to different environment information;
the configuration power consumption support algorithm is:
wherein,is the comprehensive switching value; />For a preset task load weight, +.>Weights are matched for preset power consumption, +.>Is the preset heat dissipation efficiency weight->；/>Is a preset load reference value; />For the heat dissipation efficiency of the environment, the air conditioner is->For a preset ambient heat dissipation reference value, +.>For the heat dissipation efficiency of the device, < > for>For a preset device heat dissipation reference value, +.>A preset data center heat dissipation safety reference value is set;
calculating a task load value by using the load prediction algorithm, calculating a power consumption matching value by using the power consumption matching algorithm, and calculating environmental heat dissipation efficiency by using the environmental redundancy table;
obtaining the heat dissipation efficiency of the equipment according to the heat dissipation equipment sub-efficiency of each heat dissipation equipment;
calculating a comprehensive switching value according to a power consumption supporting algorithm by using the task load value, the power consumption matching value, the environment heat dissipation efficiency and the equipment heat dissipation efficiency;
If the comprehensive switching value is higher than the switching reference threshold value, switching the data center into a low-power consumption state;
and when the data center is in a low-power consumption state, executing tasks according to the sorting result.
2. The method for processing tasks in a low power consumption state according to claim 1, wherein the method for sorting the acquired tasks according to a preset task execution policy comprises:
the method comprises the steps of configuring a task type index table, a task analysis algorithm and a task sequence algorithm, wherein the task type index table, the task analysis algorithm and the task sequence algorithm comprise the following steps:
setting a task type index table, wherein task emergency values corresponding to different task types are stored in the task type index table;
the configuration task analysis algorithm is as follows:
wherein,a task priority value; />For a preset target time parameter, +.>The task type parameter is preset; />A task target period corresponding to the task; />For the target completion time corresponding to the task, +.>Is the current moment; />A task urgency value for the task;
the configuration task sequence algorithm is as follows:
wherein,is a sequence priority value; />The task quantity corresponding to the secondary task;
calculating a task priority value and a sequence priority value by using the task analysis algorithm and the task sequence algorithm;
Storing the task with the task priority value lower than a preset priority reference value as a secondary task into a secondary task temporary storage area;
and adjusting the execution order of the secondary tasks stored in the secondary task temporary storage area by using the sequence priority value.
3. The method for processing tasks in a low power consumption state according to claim 1, wherein when the data center is in the low power consumption state, the method for executing tasks according to the result of classification ordering comprises:
acquiring an idle value of a server and a task total value of a secondary task temporary storage area;
generating an equilibrium distribution instruction according to the sorting result so as to minimize the average difference between the idle value and the task total value;
and executing the task according to the balanced distribution instruction, and mirroring the task to a local temporary storage execution area.
4. A data center using the task processing method in a low power consumption state according to any one of claims 1 to 3, wherein the data center comprises:
the task management subsystem is used for acquiring tasks of the data center in real time, and sorting and executing the tasks according to a preset task execution strategy;
the low-consumption switching subsystem is used for calculating a task load value, a power consumption matching value, an environment heat dissipation efficiency and an equipment heat dissipation efficiency, and judging whether to switch the data center into a low-power consumption state according to the task load value, the power consumption matching value, the environment heat dissipation efficiency and the equipment heat dissipation efficiency;
And the low-consumption execution subsystem works when the data center is in a low-consumption state, and is used for starting the flywheel power supply to supply power to the data center and controlling the execution of the task according to the task balancing strategy.
5. The data center of claim 4, wherein the task management subsystem comprises a task configuration module and a secondary task registry provided on a server; the task configuration module comprises a task analysis unit, a task management unit and a task execution unit; the task analysis unit is used for acquiring tasks of the data center in real time, calculating task priority values of the tasks, and storing the tasks with the task priority values lower than preset priority reference values as secondary tasks into the secondary task temporary storage area; the task management unit is used for calculating the sequence priority value of the secondary task in the secondary task temporary storage area in real time so as to adjust the execution sequence of the secondary task; and the task execution unit is used for sequentially calling and executing the tasks in the secondary task temporary storage area according to the execution sequence when the idle value of the server is larger than a preset idle reference value.
6. The data center of claim 4, wherein the low-consumption switching subsystem comprises a load prediction module, a power consumption matching module, an environmental redundancy module, a heat dissipation support module, and a low-consumption switching module; the load prediction module is used for calculating a task load value in a preset time period in the future; the power consumption matching module is used for calculating a power consumption matching value in a preset time period before; the environment redundancy module is used for calculating the environment heat dissipation efficiency in a preset time period in the future; the heat dissipation support module is used for calculating the heat dissipation efficiency of the equipment in a future preset time period; the low-consumption switching module is used for calculating to obtain a comprehensive switching value according to the task load value, the power consumption matching value, the environment heat dissipation efficiency and the equipment heat dissipation efficiency, and judging whether to switch the data center into a low-power consumption state according to the comprehensive switching value.
7. The data center of claim 4, wherein the low-consumption execution subsystem comprises a flywheel management module and a task balancing module; the flywheel management module is used for starting a flywheel power supply to supply power for the data center; the task balancing module is used for controlling the execution of the task according to the task balancing strategy.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311422511.8A CN117170851B (en) | 2023-10-31 | 2023-10-31 | Task processing method in low power consumption state and data center |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311422511.8A CN117170851B (en) | 2023-10-31 | 2023-10-31 | Task processing method in low power consumption state and data center |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117170851A CN117170851A (en) | 2023-12-05 |
| CN117170851B true CN117170851B (en) | 2024-03-19 |
Family
ID=88947070
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311422511.8A Active CN117170851B (en) | 2023-10-31 | 2023-10-31 | Task processing method in low power consumption state and data center |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117170851B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115309603A (en) * | 2021-11-30 | 2022-11-08 | 宁波亮控信息科技有限公司 | Data center energy consumption prediction optimization method, system, medium and computing device |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20230205591A1 (en) * | 2016-07-19 | 2023-06-29 | Pure Storage, Inc. | System Having Dynamic Power Management |
| US10303241B2 (en) * | 2017-06-19 | 2019-05-28 | Alibaba Group Holding Limited | System and method for fine-grained power control management in a high capacity computer cluster |
| US11239678B2 (en) * | 2018-07-23 | 2022-02-01 | Ajay Khoche | Low-cost task specific device scheduling system |
| WO2022038658A1 (en) * | 2020-08-17 | 2022-02-24 | 日本電信電話株式会社 | Task management device, task management method, and task management program |
-
2023
- 2023-10-31 CN CN202311422511.8A patent/CN117170851B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115309603A (en) * | 2021-11-30 | 2022-11-08 | 宁波亮控信息科技有限公司 | Data center energy consumption prediction optimization method, system, medium and computing device |
Non-Patent Citations (1)
| Title |
|---|
| 常带宽服务器混合任务低功耗调度算法;张忆文，等;计算机研究与发展;第52卷(第9期);2094-2104 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117170851A (en) | 2023-12-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102891496B (en) | Electricity adjusting device and method, electricity adjustment programme and electric power supply system | |
| CN106505560B (en) | A kind of network optimization operation method of more policy co-ordinations based on response priority | |
| Latif et al. | Plug in hybrid vehicle-wind-diesel autonomous hybrid power system: Frequency control using FA and CSA optimized controller | |
| WO2018174027A1 (en) | Aggregation system, method for controlling same, and complex power conversion device | |
| CN110165713B (en) | Multi-energy complementary park demand response method based on power grid peak regulation demand | |
| CN113496098A (en) | Energy consumption control method, computing equipment and medium for carbon emission reduction | |
| CN110830560A (en) | Multi-user mobile edge calculation migration method based on reinforcement learning | |
| CN117170851B (en) | Task processing method in low power consumption state and data center | |
| WO2019227273A1 (en) | Hierarchical concept based neural network model for data center power usage effectiveness prediction | |
| CN116826814B (en) | Electric energy management method based on battery cluster, energy manager and related medium | |
| CN116826815B (en) | Charging management method based on battery module, energy manager and related medium | |
| CN113765099A (en) | Intelligent scheduling system based on Internet of things and load scheduling method thereof | |
| US9946286B2 (en) | Information processing apparatus, power-consuming body, information processing method, and program | |
| CN116857798A (en) | Air conditioner electricity consumption prediction method, power grid load prediction method and device thereof | |
| CN111106415A (en) | Battery management method and device, battery and cloud server | |
| CN116454983A (en) | Wind-solar-energy-storage combined optimal control management method, system and equipment | |
| CN105227674A (en) | Internet data center's power cost control method and system | |
| CN116667443B (en) | Photovoltaic equipment and photovoltaic equipment control system | |
| CN118199025B (en) | Air conditioner load cluster aggregation and regulation method and device considering switching frequency constraint | |
| CN115265036B (en) | Power distribution method and device of intelligent refrigerator, storage medium and electronic equipment | |
| US11722002B2 (en) | Management of power supply in enclosed space | |
| CN118611075A (en) | Control method, device, equipment and medium for user side demand response resource cluster | |
| CN118645993A (en) | Power grid distribution personalized management method and device | |
| WO2024047276A1 (en) | Management of a distributed energy storage, des, arrangement | |
| CN115076915A (en) | Control method and device of air conditioner |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |