WO2022199608A1 - 一种应用于机房的节能方法、装置及系统 - Google Patents

一种应用于机房的节能方法、装置及系统 Download PDF

Info

Publication number
WO2022199608A1
WO2022199608A1 PCT/CN2022/082480 CN2022082480W WO2022199608A1 WO 2022199608 A1 WO2022199608 A1 WO 2022199608A1 CN 2022082480 W CN2022082480 W CN 2022082480W WO 2022199608 A1 WO2022199608 A1 WO 2022199608A1
Authority
WO
WIPO (PCT)
Prior art keywords
computing resource
server
cabinet
candidate
temperature
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.)
Ceased
Application number
PCT/CN2022/082480
Other languages
English (en)
French (fr)
Inventor
雷高伟
陈娟
杨丹军
师磊
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huawei Technologies Co Ltd
Original Assignee
Huawei Technologies Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Huawei Technologies Co Ltd filed Critical Huawei Technologies Co Ltd
Priority to EP22774264.0A priority Critical patent/EP4300262A4/en
Publication of WO2022199608A1 publication Critical patent/WO2022199608A1/zh
Priority to US18/472,789 priority patent/US20240023296A1/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/16—Constructional details or arrangements
    • G06F1/20—Cooling means
    • H—ELECTRICITY
    • H05—ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
    • H05K—PRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
    • H05K7/00—Constructional details common to different types of electric apparatus
    • H05K7/20—Modifications to facilitate cooling, ventilating, or heating
    • H05K7/2089—Modifications to facilitate cooling, ventilating, or heating for power electronics, e.g. for inverters for controlling motor
    • H05K7/20945—Thermal management, e.g. inverter temperature control
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26—Power supply means, e.g. regulation thereof
    • G06F1/32—Means for saving power
    • G06F1/3203—Power management, i.e. event-based initiation of a power-saving mode
    • G06F1/3234—Power saving characterised by the action undertaken
    • G06F1/329—Power saving characterised by the action undertaken by task scheduling
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/16—Constructional details or arrangements
    • G06F1/20—Cooling means
    • G06F1/206—Cooling means comprising thermal management
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26—Power supply means, e.g. regulation thereof
    • G06F1/32—Means for saving power
    • G06F1/3203—Power management, i.e. event-based initiation of a power-saving mode
    • G06F1/3234—Power saving characterised by the action undertaken
    • G06F1/3293—Power saving characterised by the action undertaken by switching to a less power-consuming processor, e.g. sub-CPU
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5094—Allocation of resources, e.g. of the central processing unit [CPU] where the allocation takes into account power or heat criteria
    • H—ELECTRICITY
    • H05—ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
    • H05K—PRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
    • H05K7/00—Constructional details common to different types of electric apparatus
    • H05K7/20—Modifications to facilitate cooling, ventilating, or heating
    • H05K7/20709—Modifications to facilitate cooling, ventilating, or heating for server racks or cabinets; for data centers, e.g. 19-inch computer racks
    • H05K7/20836—Thermal management, e.g. server temperature control
    • 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

Definitions

  • the present application relates to the field of computer technology, and in particular, to an energy-saving method, device, and system applied to a computer room.
  • Servers and air conditioners are usually stored in the computer room.
  • the server provides information technology (Internet technology, IT) services for users.
  • the air conditioner is used to control the air temperature in the computer room, so that the server and other supporting devices in the computer room can work in a suitable temperature/humidity range.
  • the energy consumption of the computer room is mainly composed of the IT energy consumption generated by the server and the cooling energy consumption generated by the air conditioner. Therefore, how to save the energy consumption of the computer room is mainly considered from these two aspects. Since the server system will receive tasks from the outside (such as virtual machine creation tasks) in real time, the tasks will be assigned to one of the servers for execution through the task scheduling algorithm, and the execution of tasks will occupy the central processing unit (CPU) resources of the server. , thereby increasing the IT load, thereby increasing IT energy consumption and cooling energy consumption. Therefore, how to formulate an appropriate task scheduling algorithm is very important to save the total energy consumption of the computer room. In addition, the air conditioning system is responsible for processing the heat generated by the IT system in the computer room.
  • the server system will receive tasks from the outside (such as virtual machine creation tasks) in real time, the tasks will be assigned to one of the servers for execution through the task scheduling algorithm, and the execution of tasks will occupy the central processing unit (CPU) resources of the server. , thereby increasing the IT load, thereby increasing
  • the heat generated by the IT system is approximately equal to the IT energy consumption. Therefore, on the premise that the air conditioning system meets the cooling capacity requirement, how to dynamically adjust the set temperature and air volume according to the IT energy consumption is very important to save the cooling energy consumption of the equipment room.
  • task scheduling instructions and temperature setting parameters are generally determined by building a predictive model. Specifically, by collecting data such as server resource utilization, computer room air conditioning system parameters, total energy consumption of the computer room, and external environmental parameters for a long period of time, model training is performed to obtain a total energy consumption prediction model and an air conditioning parameter prediction model. Based on the prediction model, the total energy consumption and air-conditioning parameters after each task is scheduled to any server can be predicted in turn, so that the server with the smallest total energy consumption can be selected as the target server for executing each task, and the corresponding air-conditioning parameters are the final air-conditioning parameters. Setting parameters.
  • the present application provides an energy-saving method and a related device applied to a computer room, which can minimize the total power of the computer room while meeting business requirements.
  • the present application provides an energy-saving method applied to a computer room
  • the computer room includes a plurality of cabinets and at least one temperature adjustment device, wherein each cabinet includes at least one server, and each server is used for providing computing resources, each A temperature adjustment device is used to adjust the temperature of the computer room, the computer room is further configured with a facility management device, and the facility management device is used to manage the computing resources and temperature adjustment equipment of the computer room, and the method includes: obtaining the The cooling conduction relationship information of the computer room and the computing resource status information of the computer room; according to the cooling conduction relationship information and the computing resource status information, trigger the facility management device to execute the computing resource management instruction set, so as to meet the business needs
  • the total power of the equipment room is minimized while meeting the demand; wherein, the cooling conduction relationship information is used to determine the cooling capacity of each cabinet in the equipment room to each temperature adjustment device in the equipment room,
  • the computing resource status information is used to determine the operating position and operating status of the computing resources contained in the computer room. In this way, the total power
  • the computing resource management instruction set is a computing resource adjustment instruction set
  • the computing resource adjustment instruction set is used to adjust the running position of at least one computing resource in the computer room.
  • the triggering the facility management device to execute the computing resource management instruction set according to the cooling capacity conduction relationship information and the computing resource state information includes: based on the The computing resource status information determines the computing resources to be migrated in the computer room; based on the cooling conduction relationship information and the computing resource status information, a target server is determined for the computing resources to be migrated, and the computing resources to be migrated are migrated to The total power of the computer room after the target server is smaller than the total power of the computer room before the migration; an instruction set for migrating the to-be-migrated computing resources to the target server is generated as the computing resource adjustment instruction set.
  • the determining a target server for the computing resource to be migrated based on the cooling capacity conduction relationship information and the computing resource state information includes: based on the cooling capacity conduction
  • the relationship information and the computing resource status information determine multiple candidate servers and the candidate total power corresponding to each candidate server for the computing resource to be migrated, and the total candidate power is assumed to migrate the computing resource to be migrated to the The total power of the computer room after each candidate server; the candidate server with the lowest corresponding candidate total power among the multiple candidate servers is selected as the target server.
  • determining a plurality of candidate servers for the computing resources to be migrated and the corresponding The candidate total power includes: determining partition cooling capacity absorption relationship information and cooling capacity sharing relationship information according to the cooling capacity conduction relationship information; set; according to the cooling conduction relationship information, the partition combination set to be selected and the computing resource status information, determine multiple candidate servers and the candidate total power corresponding to each candidate server for the computing resources to be migrated, so The plurality of candidate servers belong to the combination set of the to-be-selected partitions; wherein, the partition cooling absorption relationship information is used to indicate the cooling absorption capacity of the cabinets in each partition in the computer room to the temperature adjustment equipment in the partition, and one partition It includes a temperature adjustment device and at least one cabinet; the cooling capacity sharing relationship information is used to indicate between every two partitions in the equipment room, the cooling capacity of the cabinet of one partition to the temperature adjustment device of another partition;
  • the set of partition combinations to be selected includes a plurality of partition combinations to be selected with different number of partitions
  • determining a plurality of A candidate server comprising: determining, based on the computing resource state information, a candidate server on each cabinet in each candidate partition combination in the candidate partition combination set for the computing resource to be migrated; wherein the first The power change amount is not greater than the second power change amount, and the first power change amount is the power change amount of the alternative server assuming that the to-be-migrated computing resource is migrated to the alternative server, and the second power change amount is The power change amount is the power change amount of the other servers assuming that the to-be-migrated computing resources are migrated to other servers except the alternative server on the same cabinet; based on the computing resource status information and the cold The power balance degree corresponding to each cabinet is determined, and the power balance degree is assumed to be the power balance of each cabinet after the computing resources to be migrated are migrated to the candidate servers The power balance degree of the combination of partitions to be selected to which the cabinet belongs; in each combination of partitions to be selected, the candidate servers on the cabinet with the
  • the method further includes: according to the number of CPU cores of the computing resource to be migrated, the CPU utilization rate of the computing resource to be migrated, the CPU of the candidate server The number of cores, one or more of the no-load power consumption or full-load power consumption of the candidate server is determined, assuming that after the to-be-migrated computing resources are migrated to the candidate server, the power consumption of the candidate server amount of change.
  • the method further includes: after migrating the to-be-migrated computing resource to the candidate server on the each cabinet according to the assumption, the power of each cabinet is , Determine the cooling capacity conduction relationship information of each cabinet and the temperature adjustment equipment in the partition where each cabinet is located, the normalized power of each cabinet; determine the normalized power of each cabinet according to the normalized power of each cabinet. The power balance corresponding to each cabinet is described.
  • the number of CPU cores of the candidate server is not less than the number of CPU cores of the computing resources to be migrated and the CPU cores of the computing resources existing in the candidate server. and/or, the memory specification of the candidate server is not less than the sum of the memory specification of the computing resource to be migrated and the memory specification of the computing resources existing in the candidate server.
  • the determining a target server for the computing resource to be migrated based on the cooling capacity conduction relationship information and the computing resource status information includes: based on the computing resource status information, and determine an alternative server on each cabinet in the computer room for the computing resource to be migrated; wherein the first power change is not greater than the second power change, and the first power change is assumed to be the After the to-be-migrated computing resource is migrated to the alternative server, the power change of the alternative server, the second power change is the assumption that the to-be-migrated computing resource is migrated to the same cabinet except the standby server.
  • the power variation of the other servers After selecting other servers other than the server, the power variation of the other servers; based on the computing resource status information and the cooling conduction relationship information, determine the power balance corresponding to each cabinet, the power balance The degree is the power balance degree of the computer room assuming that the computing resources to be migrated are migrated to the candidate servers on each cabinet; the standby server on the cabinet with the smallest corresponding power balance degree is selected among the multiple cabinets. Select the server as the target server. In this way, the alternative servers on each cabinet in the equipment room can be selected, and the multiple alternative servers are the servers with the smallest corresponding power changes. Assuming that the computing resources to be migrated are migrated to these servers, the Computational power will be reduced.
  • the candidate server on the cabinet with the smallest corresponding power balance from the multiple cabinets as the target server, it is possible to reduce the temperature control power of the temperature adjustment device after the assumption that the migration is made, and then it is possible to reduce the temperature control power of the temperature adjustment device after migration.
  • the total power of the computer room is less than the total power of the computer room before the migration.
  • the method further includes: calculating the power of each of the multiple cabinets after the computing resource adjustment instruction set is assumed to be executed; The power of each cabinet in the at least one temperature adjustment device determines the temperature setting parameter of each temperature adjustment device in the at least one temperature adjustment device, and the temperature setting parameter is used to satisfy the requirements of the plurality of cabinets after the computing resource adjustment instruction set is executed. and triggering the facility management device to execute a temperature adjustment instruction set, where the temperature adjustment instruction set is generated according to the temperature setting parameters. In this way, the instruction set can be adjusted according to the computing resources, and the temperature adjustment instruction set of the temperature adjustment device can be determined.
  • the method further includes: calculating the power of each cabinet in the plurality of cabinets after the computing resource adjustment instruction set is assumed to be executed;
  • the power of the cabinets included in the partition where the adjustment equipment is located determines the temperature setting parameters of each temperature adjustment equipment; triggering the facility management equipment to execute a temperature adjustment instruction set, the temperature adjustment instruction set is generated according to the temperature setting parameters of.
  • the computing resource management instruction set is a computing resource allocation instruction set, and the computing resource allocation instruction set is used to allocate at least one computing resource in the computer room.
  • the triggering the facility management device to execute the computing resource management instruction set according to the cooling conduction relationship information and the computing resource state information includes: receiving a service requirement , the business requirements are used to request the operation of computing resources in the computer room; based on the business requirements, the cooling conduction relationship information and the computing resource status information, the target computing resources that meet the business requirements are determined and provided
  • the target server of the target computing resource the increase of the total power of the computer room before and after the target server provides the target computing resource is minimized; an instruction set for allocating the target computing resource on the target server is generated. as the computing resource allocation instruction set.
  • the target computing resources that meet the business requirements are determined based on the business requirements, the cooling capacity conduction relationship information, and the computing resource status information, and the A target server for a target computing resource, including: determining multiple candidate servers and a candidate increment corresponding to each candidate server for the target computing resource based on the business requirements, the cooling conduction relationship information, and the computing resource status information , the candidate increment is the increment of the total power of the computer room after assuming that the target computing resource is migrated to each candidate server; the candidate with the lowest corresponding candidate increment is selected from the multiple candidate server sets server as the target server.
  • the candidate server with the lowest corresponding candidate increment the increment of the total power of the computer room can be minimized before and after the execution of the computing resource allocation instruction set.
  • determining a plurality of candidate servers and each The candidate increment corresponding to the candidate server includes: determining the cooling capacity absorption relationship information and cooling capacity sharing relationship information of the partition according to the cooling capacity conduction relationship information; determining according to the cooling capacity absorption relationship information of the zone and the cooling capacity sharing relationship information A combination set of partitions to be selected; according to the business requirements, the information on the cooling capacity conduction relationship, the combination set of partitions to be selected and the computing resource status information, determine multiple candidate servers and each candidate server for the target computing resource
  • the cooling capacity of a partition includes a temperature adjustment device and at least one cabinet; the cooling capacity sharing relationship information is used to indicate between every two partitions in the equipment room, the cabinet of one partition adjusts the temperature of the other partition
  • Resource determination of multiple candidate servers includes: based on the business requirements and computing resource status information, determining, for the target computing resource, a candidate server on each cabinet in each candidate partition combination set in the candidate partition combination set A server; wherein, the candidate server has the ability to provide the target computing resource, and a third power change is not greater than a fourth power change, and the third power change is assumed to migrate the target computing resources to After the candidate server, the power variation of the candidate server, and the fourth power variation is the power variation assuming that the target computing resource is migrated to other servers other than the candidate server on the same cabinet ; Based on the computing resource status information and the cooling conduction relationship information, determine the power balance degree corresponding to each cabinet, and the power balance degree assumes that the target computing resource is migrated to each cabinet After the candidate server on the cabinet, the power balance degree of the combination of
  • the method further includes: according to the number of CPU cores of the target computing resource, the CPU utilization rate of the target computing resource, and the number of CPU cores of the candidate server , one or more items of no-load power consumption or full-load power consumption of the candidate server are determined, assuming that after the target computing resource is migrated to the candidate server, the amount of power change of the candidate server.
  • the method further includes: after migrating the target computing resource to the candidate server on each cabinet according to the assumption, the power of each cabinet, Determine the cooling capacity conduction relationship information of each cabinet and the temperature adjustment equipment in the partition where each cabinet is located, and the normalized power of each cabinet; determine the normalized power of each cabinet according to the The power balance corresponding to each cabinet.
  • the number of CPU cores of the candidate server is not less than the number of CPU cores of the target computing resource and the number of CPU cores of the computing resources existing in the candidate server. and/or, the memory specification of the candidate server is not less than the sum of the memory specification of the target computing resource and the memory specification of the computing resources existing in the candidate server.
  • the triggering the facility management device to execute the computing resource management instruction set according to the cooling capacity conduction relationship information and the computing resource state information includes: based on the Computing resource status information, determining an alternative server on each cabinet in the computer room for the target computing resource; wherein, the alternative server has the ability to provide the target computing resource, and the third power variation is not is greater than the fourth power change amount, the third power change amount is the power change amount of the alternative server after the target computing resource is assumed to be migrated to the alternative server, and the fourth power change amount is the hypothesis After the target computing resource is migrated to other servers except the alternative server on the same cabinet, the power variation of the other servers; based on the computing resource state information and the cooling conduction relationship information, determine The power balance degree corresponding to each cabinet, the power balance degree is the power balance degree of the computer room assuming that the target computing resources are migrated to the alternative server on each cabinet; The candidate server on the cabinet with the corresponding minimum power balance degree is selected as the target server among the multiple cabinets.
  • the alternative servers on each cabinet in the computer room can be selected, and the multiple alternative servers are the servers with the smallest corresponding power change, then assuming that the target computing resources are allocated to these servers, the computing The increase in power will be smaller. Further, selecting the candidate server on the cabinet with the smallest corresponding power balance from the multiple cabinets as the target server, it is possible to make the temperature control power increment of the temperature adjustment device after the migration is assumed to be small, and further This makes the increment of the total power of the migrated computer room smaller.
  • the method further includes: acquiring the power of each of the multiple cabinets after the computing resource allocation instruction set is assumed to be executed; The power of each cabinet in the at least one temperature adjustment device determines a temperature setting parameter of each temperature adjustment device, and the temperature setting parameter is used to satisfy the requirements of the plurality of cabinets after the computing resource allocation instruction set is executed. the temperature demand; triggering the facility management device to execute the temperature adjustment instruction set generated according to the temperature setting parameter.
  • the method further includes: calculating the power of each cabinet in the plurality of cabinets after the computing resource allocation instruction set is assumed to be executed;
  • the power of the cabinets included in the partition where the adjustment equipment is located determines the temperature setting parameters of each temperature adjustment equipment; triggering the facility management equipment to execute a temperature adjustment instruction set, the temperature adjustment instruction set is generated according to the temperature setting parameters of.
  • the method further includes: calculating the power of each cabinet in the plurality of cabinets after the computing resource allocation instruction set is assumed to be executed;
  • the power of the cabinets included in the partition where the adjustment equipment is located determines the temperature setting parameters of each temperature adjustment equipment; triggering the facility management equipment to execute a temperature adjustment instruction set, the temperature adjustment instruction set is generated according to the temperature setting parameters of.
  • the total power of the equipment room includes the computing power of the multiple cabinets and the total power of the at least one temperature adjustment device.
  • the first cabinet is included in the plurality of cabinets
  • the first temperature adjustment device is included in the plurality of temperature adjustment devices
  • the method further includes: based on the first temperature , the second temperature, the first inlet air temperature and the second inlet air temperature to determine the cooling capacity conduction relationship information between the first cabinet and the first temperature adjustment equipment; wherein, the first inlet air temperature is simulated based on the first set model
  • the inlet air temperature of the first cabinet obtained from the operation of the equipment room, the first setting model is constructed based on the equipment room, and the first setting model includes: the multiple cabinets of the first power,
  • the at least one temperature adjustment device of the first temperature is the second inlet air temperature is the inlet air temperature of the first cabinet obtained by simulating the operation of the computer room based on the second setting model
  • the second inlet air temperature is the inlet air temperature of the first cabinet.
  • a second setting model is constructed based on the equipment room, and the second setting model includes: the plurality of cabinets of the first power, the first
  • the computer room includes a first partition, and the first partition includes a second temperature adjustment device and at least one second cabinet; the method further includes: based on the at least one The cooling capacity of each of the second cabinets for the second temperature adjustment device is determined, and the zone cooling absorption relation information of the first zone is determined.
  • the computer room includes a first partition and a second partition, the first partition includes a second temperature adjustment device and at least one second cabinet, and the second partition includes a third temperature regulating device and at least one third cabinet; the method further includes: based on the cold absorption capacity of each of the at least one second cabinet to the third temperature regulating device, and the at least one The cooling capacity of each of the third cabinets for the second temperature adjustment device is determined, and the cooling capacity of the first partition and the second partition share relationship information.
  • the computer room includes a first partition, and the first partition includes a second temperature adjustment device and at least one second cabinet;
  • the cold energy absorption capacity of the second temperature adjustment device is stronger than the cold energy absorption capacity of each second cabinet for other temperature adjustment devices in the plurality of temperature adjustment devices except the second temperature adjustment device.
  • an embodiment of the present application provides an energy-saving device, the energy-saving device is used to save energy consumption of a computer room, the computer room includes a plurality of cabinets and at least one temperature adjustment device, wherein each cabinet includes at least one server, Each server is used to provide computing resources, and each temperature adjustment device is used to adjust the temperature of the computer room.
  • the computer room is also configured with a facility management device, and the facility management device is used to manage the computing resources and temperature adjustment of the computer room.
  • the energy-saving device includes an acquisition unit and a processing unit: the acquisition unit is used to acquire the cooling conduction relationship information of the computer room and the computing resource status information of the computer room; the processing unit is used to The cooling conduction relationship information and the computing resource status information trigger the facility management device to execute the computing resource management instruction set to minimize the total power of the equipment room while meeting business requirements; wherein the cooling conduction relationship information It is used to determine the cooling capacity of each cabinet in the computer room to each temperature adjustment device in the computer room, and the computing resource status information is used to determine the operating position and operating status of the computing resources contained in the computer room. .
  • the computing resource management instruction set is a computing resource adjustment instruction set
  • the computing resource adjustment instruction set is used to adjust the running position of at least one computing resource in the computer room.
  • the processing unit is specifically configured to: determine the computing resource to be migrated in the computer room based on the computing resource status information;
  • the computing resource status information determines the target server for the computing resource to be migrated, and the total power of the computer room after the computing resource to be migrated is migrated to the target server is smaller than the total power of the computer room before the migration;
  • the instruction set for migrating the computing resource to be migrated to the target server is used as the computing resource adjustment instruction set.
  • the processing unit is specifically configured to: determine multiple candidate servers and The candidate total power corresponding to each candidate server, the candidate total power is the total power of the computer room after assuming that the to-be-migrated computing resources are migrated to each candidate server; The candidate server with the lowest candidate total power is used as the target server.
  • the processing unit is further configured to: calculate the power of each of the multiple cabinets after the computing resource adjustment instruction set is assumed to be executed;
  • the power of each of the plurality of cabinets determines a temperature setting parameter of each temperature adjusting device in the at least one temperature adjusting device, and the temperature setting parameter is used to satisfy the requirements of the multiple temperature adjusting instructions after the computing resource adjusting instruction set is executed.
  • the temperature requirement of each cabinet triggering the facility management device to execute a temperature adjustment instruction set, where the temperature adjustment instruction set is generated according to the temperature setting parameter.
  • the computing resource management instruction set is a computing resource allocation instruction set, and the computing resource allocation instruction set is used to allocate at least one computing resource in the computer room.
  • the processing unit is specifically configured to: receive business requirements, where the business requirements are used to request to run computing resources in the computer room; based on the business requirements, the The cooling conduction relationship information and the computing resource status information are determined, the target computing resources that meet the business requirements and the target server that provides the target computing resources, and the total amount of the computer room before and after the target server provides the target computing resources.
  • the increase in power is minimized; and an instruction set for allocating the target computing resource at the target server is generated as the computing resource allocation instruction set.
  • the processing unit is specifically configured to: determine, based on the business requirement, the cooling capacity conduction relationship information, and the computing resource state information, a multi-level decision for the target computing resource.
  • candidate servers and a candidate increment corresponding to each candidate server the candidate increment is the increment of the total power of the computer room after assuming that the target computing resources are migrated to each candidate server; From the candidate server set, the candidate server with the lowest corresponding candidate increment is selected as the target server.
  • the processing unit is further configured to: obtain the power of each of the multiple cabinets after the computing resource allocation instruction set is assumed to be executed; The power of each of the plurality of cabinets determines a temperature setting parameter of each temperature adjusting device in the at least one temperature adjusting device, and the temperature setting parameter is used to satisfy the requirement that the multiple The temperature requirement of each cabinet; triggering the facility management device to execute a temperature adjustment instruction set, where the temperature adjustment instruction set is generated according to the temperature setting parameter.
  • the present application provides yet another energy-saving device, the energy-saving device is used to save energy consumption of a computer room, the computer room includes a plurality of cabinets and at least one temperature adjustment device, wherein each cabinet includes at least one server, each A server is used to provide computing resources, each temperature adjustment device is used to adjust the temperature of the computer room, and the computer room is also configured with a facility management device, and the facility management device is used to manage the computing resources and temperature adjustment equipment of the computer room , the energy-saving device includes a processor and a memory, the memory is used for storing program instructions, and the processor is configured to invoke the program instructions to perform the following operations: acquiring the cooling conduction relationship information of the equipment room and all The computing resource status information of the computer room; according to the cooling conduction relationship information and the computing resource status information, trigger the facility management device to execute the computing resource management instruction set, so as to meet the business requirements and minimize the total power of the computer room.
  • the cooling conduction relationship information is used to determine the cooling capacity of each cabinet in the equipment
  • the computing resource management instruction set is a computing resource adjustment instruction set
  • the computing resource adjustment instruction set is used to adjust the running position of at least one computing resource in the computer room.
  • the processor is configured to invoke the program instruction to specifically perform the following operations: determine the computing resource to be migrated in the computer room based on the computing resource state information; A target server is determined for the computing resource to be migrated based on the cooling conduction relationship information and the computing resource status information, and the total power of the computer room after the computing resource to be migrated is migrated to the target server is less than the total power before the migration. the total power of the computer room; and generating an instruction set for migrating the to-be-migrated computing resource to the target server as the computing resource adjustment instruction set.
  • the processor is configured to call the program instructions to specifically perform the following operations: based on the cooling capacity conduction relationship information and the computing resource state information
  • the computing resource to be migrated determines a plurality of candidate servers and the candidate total power corresponding to each candidate server, and the total candidate power is the total power of the computer room assuming that the computing resource to be migrated is migrated to each candidate server; From the plurality of candidate servers, a candidate server with the lowest total power of the corresponding candidates is selected as the target server.
  • the processor is configured to call the program instructions and further perform the following operations: obtaining, assuming that the computing resource adjustment instruction set is executed, the following operations are performed in the multiple cabinets.
  • the power of each cabinet; the temperature setting parameter of each temperature regulating device in the at least one temperature regulating device is determined according to the power of each cabinet in the plurality of cabinets, and the temperature setting parameter is used to meet the requirements of executing the After calculating the resource adjustment instruction set, the temperature requirements of the multiple cabinets; triggering the facility management device to execute the temperature adjustment instruction set, the temperature adjustment instruction set being generated according to the temperature setting parameters.
  • the computing resource management instruction set is a computing resource allocation instruction set, and the computing resource allocation instruction set is used to allocate at least one computing resource in the computer room.
  • the processor is configured to invoke the program instructions to specifically perform the following operations: receive a service requirement, where the service requirement is used to request to run the computation in the computer room resources; based on the business requirements, the cooling conduction relationship information, and the computing resource status information, determine the target computing resources that meet the business requirements and the target server that provides the target computing resources, and the target server provides all the Minimizing the increment of the total power of the computer room before and after the target computing resource; generating an instruction set for allocating the target computing resource on the target server as the computing resource allocation instruction set.
  • the processor is configured to invoke the program instructions to specifically perform the following operations: based on the business requirements, the cooling conduction relationship information, and the computing resources
  • the status information determines a plurality of candidate servers and a candidate increment corresponding to each candidate server for the target computing resource, and the candidate increment is assumed to be the change in the computer room after the target computing resource is migrated to each candidate server.
  • the increment of the total power; the candidate server with the lowest corresponding candidate increment is selected from the multiple candidate server sets as the target server.
  • the processor is configured to invoke the program instructions and further perform the following operation: after the calculation assumption is executed, the computing resource allocation instruction set is executed in the multiple cabinets.
  • the power of each cabinet; the temperature setting parameter of each temperature regulating device in the at least one temperature regulating device is determined according to the power of each cabinet in the plurality of cabinets, and the temperature setting parameter is used to meet the requirements of executing the
  • the present application provides a computing resource supply system, the computing resource supply system includes the system including the energy-saving equipment and a computer room as described in any one of the first to third aspects above.
  • the present application provides a computer-readable storage medium, the computer-readable storage medium is used for storing instructions, and when the instructions are executed, the instructions are as described in any one of the first to third aspects. The described method is implemented.
  • the present application provides a computer program or computer program product, comprising codes or instructions, which, when the codes or instructions are run on a computer, cause the computer to execute the method described in any one of the first to third aspects is realized.
  • FIG. 1 is a schematic structural diagram of a computer room provided by an embodiment of the present application.
  • FIG. 2 is a schematic diagram of the architecture of a computing resource provisioning system provided by an embodiment of the present application
  • FIG. 3 is a schematic diagram of the architecture of another computing resource provisioning system provided by an embodiment of the present application.
  • FIG. 4 is a schematic diagram of the architecture of another computing resource provisioning system provided by an embodiment of the present application.
  • FIG. 5 is a schematic flowchart of an energy saving method applied to a computer room provided by an embodiment of the present application
  • FIG. 6 is a schematic structural diagram of an energy-saving device provided by an embodiment of the present application.
  • FIG. 7 is a schematic structural diagram of another energy-saving device provided by an embodiment of the application.
  • At least one (item) means one or more
  • plural means two or more
  • at least two (item) means two or three and three
  • “and/or” is used to describe the corresponding relationship between associated objects, indicating that there can be three kinds of relationships, for example, “A and/or B” can mean: only A exists, only B exists, and both A and B exist three A case where A and B can be singular or plural.
  • the character “/” generally indicates that the associated objects are an “or” relationship.
  • At least one item(s) below” or similar expressions thereof refer to any combination of these items, including any combination of single item(s) or plural items(s).
  • At least one (a) of a, b or c can mean: a, b, c, "a and b", “a and c", “b and c", or "a and b and c" ", where a, b, c can be single or multiple.
  • the energy-saving method provided by the embodiments of the present application is applied to the computer room.
  • the computer room and related concepts are first introduced.
  • a data center (or called a resource center, computing resource center, or computing center) is a set of facilities that provide information technology (Internet technology, IT) services to the outside world. It can include computer software and hardware systems and other supporting equipment (such as communication and storage systems), as well as data communication connections, environmental control equipment, monitoring equipment, various safety devices, etc. From a spatial point of view, a data center can be divided into multiple computer rooms (or hangars); from a logical point of view, it can be divided into three layers, the computer room supporting layer, the infrastructure layer, and the software layer.
  • the supporting layer of the equipment room includes refrigeration equipment, power supply equipment, fire protection system, etc., providing power and cold sources for the infrastructure layer.
  • the infrastructure layer including computing devices, storage devices, and virtualized resources, is used to carry business applications at the software layer.
  • the software layer includes business applications, cloud platforms, etc.
  • the following analyzes the impact of the above three layers on the total power of the data center.
  • the data center software layer (for example, OpenStack) can receive tasks from outside the data center (such as virtual machine creation tasks) in real time, put them in the task queue for scheduling, and assign tasks to one of the servers through the task scheduling algorithm Executing tasks on the server will occupy the CPU resources of the server, thereby increasing the IT load, thereby increasing IT energy consumption and cooling energy consumption. Therefore, how to allocate new tasks received, or new tasks generated by its own data center, will affect the total power of the data center.
  • tasks from outside the data center such as virtual machine creation tasks
  • Executing tasks on the server will occupy the CPU resources of the server, thereby increasing the IT load, thereby increasing IT energy consumption and cooling energy consumption. Therefore, how to allocate new tasks received, or new tasks generated by its own data center, will affect the total power of the data center.
  • virtualization technology is usually used to virtualize a physical server into multiple virtual machines, which are dynamically allocated to users. Adjusting the ownership relationship between servers and virtual machines, that is, migrating virtual machines from one physical server to another physical server (also called virtual machine migration), can realize the redistribution of IT load, which will affect IT energy consumption and cooling energy consumption . Therefore, how to migrate and adjust the virtual machine will affect the total power of the data center.
  • the cooling system On the supporting floor of the data center computer room, the cooling system is responsible for taking away the heat generated by the IT system in the computer room. According to the relationship between energy consumption and heat, the heat generated by the IT system is approximately equal to the IT energy consumption. Therefore, on the premise that the cooling system meets the cooling capacity requirement, how to dynamically adjust the set temperature and air output of the cooling equipment according to the IT energy consumption will affect the total power of the data center.
  • the computer room is the basic unit that constitutes a data center in space.
  • the equipment room includes a plurality of cabinets and at least one temperature adjustment device.
  • each cabinet includes at least one server, and each server is used to provide computing resources to meet external input or business requirements generated inside the computer room (for example, monitoring tasks, security protection tasks, etc.); each temperature adjustment device It is used to adjust the temperature of the computer room, so that the server and other supporting devices in the computer room can work within a suitable temperature/humidity range.
  • the temperature of the computer room can be understood as the indoor temperature of the computer room, the air temperature, and so on.
  • each temperature adjustment device is used to adjust the inlet air temperature of the cabinet, so that the inlet air temperature of the cabinet is not greater than the upper limit of the inlet air temperature of the cabinet (for example, 28°C, 27°C, etc.) .
  • the upper limit of the inlet air temperature of the cabinet may be determined by the normal operating temperature requirements of the servers or other supporting devices placed in the cabinet.
  • a computer room is taken as an application scenario, and an energy saving method applied to a computer room is introduced.
  • FIG. 1 it is a schematic structural diagram of a computer room provided by an embodiment of the present application.
  • the computer room shown in Figure 1 it includes 16 cabinets (at least one server can be placed on one cabinet), namely L1-L8, R1-R8; 4 temperature adjustment devices (for example, air conditioners, temperature control machines, refrigerators) , air-cooled equipment, water-cooled equipment, etc.), namely A1-A4.
  • L1-L8, R1-R8 4 temperature adjustment devices (for example, air conditioners, temperature control machines, refrigerators) , air-cooled equipment, water-cooled equipment, etc.), namely A1-A4.
  • the computer room For the energy consumption of the computer room, it is mainly composed of the IT energy consumption generated by the server and the cooling energy consumption (in some environmental conditions, there may also be heating energy consumption) generated by the temperature adjustment equipment. Therefore, from these two aspects Consider how to save energy in the computer room. Since the server system will receive tasks from the outside (such as virtual machine creation tasks) in real time, the tasks will be assigned to one of the servers for execution through the task scheduling algorithm, and the execution of tasks will occupy the central processing unit (CPU) resources of the server. , thereby increasing the IT load, thereby increasing IT energy consumption and cooling energy consumption. Therefore, how to formulate an appropriate task scheduling algorithm is very important to save the total energy consumption of the computer room.
  • tasks from the outside such as virtual machine creation tasks
  • the tasks will be assigned to one of the servers for execution through the task scheduling algorithm, and the execution of tasks will occupy the central processing unit (CPU) resources of the server. , thereby increasing the IT load, thereby increasing IT energy consumption and cooling energy consumption. Therefore, how to formulate an appropriate task
  • the air conditioning system is responsible for processing the heat generated by the IT system in the computer room. According to the relationship between energy consumption and heat, the heat generated by the IT system is approximately equal to the IT energy consumption. Therefore, on the premise that the air conditioning system meets the cooling capacity requirement, how to dynamically adjust the set temperature and air volume according to the IT energy consumption is very important to save the cooling energy consumption of the equipment room.
  • task scheduling instructions and temperature setting parameters are generally determined by building a predictive model. Specifically, by collecting data such as server resource utilization, computer room air conditioning system parameters, total energy consumption of the computer room, and external environmental parameters for a long period of time, model training is performed to obtain a total energy consumption prediction model and an air conditioning parameter prediction model. Based on the prediction model, the total energy consumption and air-conditioning parameters after each task is scheduled to any server can be predicted in turn, so that the server with the smallest total energy consumption can be selected as the target server for executing each task, and the corresponding air-conditioning parameters are the final air-conditioning parameters. Setting parameters.
  • this method needs to collect a large amount of experimental data under various conditions to train the model, which is highly dependent on the sample data; and the sample data is difficult to meet the complex and changeable needs of the actual computer room operation.
  • the present application provides an energy saving method applied to a computer room, which can minimize the total power of the computer room under the condition of meeting business requirements.
  • FIG. 2 it is a schematic structural diagram of a computing resource provisioning system provided by an embodiment of the present application.
  • the computing resource supply system may include simulation equipment, energy saving equipment, facility management equipment and a computer room.
  • the cooling energy conduction relationship information is used to determine the cooling energy absorption capacity of each cabinet in the equipment room to each temperature regulating device in the equipment room.
  • the cooling capacity here specifically refers to the cooling capacity that is conducted from the temperature adjustment device to the cabinet, or the cabinet absorbs the cooling capacity of the temperature adjustment device.
  • the information on the relationship between cooling and conduction may also be referred to as information on the relationship between heat conduction.
  • the heat conduction relationship information is used to determine the heat absorption capacity of each temperature adjustment device in the equipment room to each cabinet in the equipment room.
  • the heat here specifically refers to the heat that is conducted from the cabinet to the temperature adjustment equipment, or the temperature adjustment equipment absorbs the heat of the cabinet.
  • the cooling capacity specifically refers to the cooling capacity that is conducted from the cabinet to the temperature regulating device, or that the temperature regulating device absorbs the cooling capacity of the cabinet.
  • the heat specifically refers to the heat that is conducted from the temperature adjustment equipment to the cabinet, or the cabinet absorbs the heat of the temperature adjustment equipment.
  • the computing resources in the embodiments of the present application may include servers, virtual machines, tasks, and the like in the computer room.
  • a server is a hardware device placed in a cabinet; virtual machines and tasks are software units that are assigned to a server.
  • at least one computing resource targeted by the computing resource management instruction set may be a virtual machine or a task.
  • the computing resource management instruction set can adjust the running position of the virtual machine or the task; or allocate (or create) the virtual machine or the task in the server for business requirements.
  • the computer resource status information is used to determine the running position and running status of the computing resources contained in the computer room.
  • the running position or referred to as position information, coordinate information
  • each cabinet and each server may be numbered (or named or marked), and a number has a corresponding relationship with a piece of location information.
  • cabinet L1 corresponds to the first cabinet on the left side from top to bottom in the equipment room.
  • create a coordinate system based on the computer room record the coordinates of each cabinet and each server, and a coordinate has a corresponding relationship with a location information.
  • the running state of the server can be described by the number of central processing unit (CPU) cores, memory specifications, no-load power, full-load power, and other running data of the server.
  • the running state of the virtual machine can be described by the number of CPU cores, memory specifications, CPU utilization, memory utilization, and other running data of the virtual machine.
  • the running state of the task can be described by the CPU requirements (ie, how many CPU cores are required) and memory requirements required for the task to run.
  • the computing resource instruction set includes at least one adjustment or allocation instruction for computing resources.
  • the computing resource management instruction set may be a computing resource adjustment instruction set, and the computing resource adjustment instruction set is used to adjust the running position of at least one computing resource in the computer room.
  • the computing resource management instruction set may be a computing resource allocation instruction set, and the computing resource allocation instruction set is used to allocate at least one computing resource in the computer room.
  • Simulation equipment is an electronic equipment that uses data center simulation technology to simulate the airflow field and temperature field of the computer room.
  • the control equation of air flow is solved by computational fluid dynamics software in the data center, in which the law of conservation of mass, the law of momentum change, the first law of thermodynamics, the second law of thermodynamics, Fourier's law of heat transfer, etc. are mainly used, and the The airflow field and temperature field of the computer room are simulated.
  • the simulation equipment can use the computational fluid dynamics software to build the computer room corresponding computer room computational fluid dynamics model (also known as virtual machine room). It should be noted that the cabinet placement, air conditioner size and position, computer room height, wiring rack layout, beam position, column position, etc. should be consistent between the virtual computer room and the real computer room. After the computer room computational fluid dynamics model is established, it can be used to simulate the temperature field and airflow field of the computer room under the distribution of the set parameters (or simulation parameters) of different cabinet power consumption and/or temperature adjustment equipment.
  • the simulation device can construct a virtual computer room corresponding to the actual computer room according to the location information of the constituent objects collected in the actual computer room; The inlet air temperature of each cabinet, so that the energy-saving equipment can conduct information on the cooling capacity of the computer room.
  • the simulation device may also receive simulation parameters sent by the energy-saving device, simulate the operation of the machine room according to the simulation parameters and the computational fluid dynamics model of the machine room, and output the simulation results.
  • a data center (including one or more computer rooms) may be configured with a common facility management device (for example, it may be a data center infrastructure management (DCIM) device).
  • DCIM data center infrastructure management
  • each equipment room is configured with its own corresponding facility management equipment, and the DCIM equipment can manage the facility management equipment of each equipment room.
  • the facility management device can collect computing resource status data of temperature adjustment devices, cabinets, servers, virtual machines, tasks to be scheduled, etc., to generate computing resource status information; it can also control the managed devices, such as adjusting Set parameters of temperature adjustment equipment, migrate virtual machines, schedule tasks to specified servers, etc.
  • facility management can also provide external data acquisition interfaces and equipment control interfaces for external systems to acquire required data and control equipment.
  • the facility management device may acquire computing resource status information in the actual computer room, and send the computing resource status information to the energy-saving device; in addition, the facility management device may be triggered by the energy-saving device to execute the computing resource management instruction set.
  • Energy-saving equipment can interact with facility management equipment and simulation equipment through interfaces.
  • the energy-saving device may collect computing resource status information of the equipment room through the interface with the facility management device, and send the computing resource management instruction set to the facility management device through the interface.
  • the energy-saving device may also send simulation parameters to the simulation device through the interface with the simulation device, or collect simulation results (exemplarily, multiple temperature data) obtained after the simulation device simulates the computer room running.
  • the energy-saving device can obtain the temperature data obtained from the simulation device after simulating the operation of the computer room, and according to the temperature data, the energy-saving device can calculate the cooling conduction relationship information of the computer room. In addition, the energy-saving device can obtain the computing resource status information of the equipment room from the facility management device. Next, the energy-saving device triggers the facility management device to execute the computing resource management instruction set according to the cooling conduction relationship information and the computing resource state information, so as to minimize the total power of the computer room while meeting the service requirements.
  • the energy-saving device may further determine the temperature setting parameter of each temperature adjustment device according to the power of each cabinet in the multiple cabinets in the computer room after the assumed execution of the computing resource management instruction set, The facility management device is then triggered to execute the temperature adjustment instruction set generated according to the temperature setting parameter.
  • FIG. 3 it is a schematic structural diagram of another computing resource provisioning system provided by an embodiment of the present application.
  • the system includes simulation equipment, facility management equipment, and a computer room, and the above-mentioned energy-saving equipment is integrated in the facility management equipment.
  • FIG. 4 it is a schematic structural diagram of another computing resource provisioning system provided by an embodiment of the present application.
  • the system includes facility management equipment and a computer room, and the above-mentioned simulation equipment and energy-saving equipment are integrated in the facility management equipment.
  • FIG. 5 it is a schematic flowchart of an energy saving method applied to a computer room provided by an embodiment of the present application.
  • the method can be applied to the computing resource supply system shown in FIG. 2 (or FIG. 3 and FIG. 4 ), and the energy-saving equipment described below can be the energy-saving equipment shown in FIG. 2 ; the computer room described below can be the energy-saving equipment shown in FIG. 1 or The computer room shown in Figure 2 (or Figure 3, Figure 4).
  • the energy-saving device may be integrated in the facility management device, and the device executing the following method may also be the facility management device.
  • the method includes the following steps:
  • the energy-saving device acquires the cooling conduction relationship information of the computer room and the computing resource state information of the computer room.
  • the energy-saving device may acquire the computing resource status information of the computer room through the facility management device.
  • the running location information of the server included in the computing resource status information can be exemplified as follows:
  • the running location information of the virtual machine included in the computing resource status information can be exemplified as follows:
  • the computing resource status information may also include server running status information: the number of CPU cores, memory specifications, no-load power, full-load power, etc.; virtual machine running status information: the number of CPU cores, memory specifications, CPU cores Utilization, memory utilization, etc.
  • the energy-saving device may collect multiple temperature data obtained after simulating the operation of the computer room through the simulation device, so as to obtain the information on the cooling conduction relationship of the computer room.
  • the cooling conduction relationship information of the first cabinet and the first temperature adjustment device as an example, the method of the cooling conduction relationship information of the computer room is introduced.
  • the first cabinet is any cabinet in the equipment room
  • the first temperature adjustment device is any temperature adjustment equipment in the equipment room.
  • the energy-saving device determines information on the cooling capacity conduction relationship between the first cabinet and the first temperature adjustment device based on the first temperature, the second temperature, the first inlet air temperature, and the second inlet air temperature; wherein the first An inlet air temperature is the inlet air temperature of the first cabinet obtained by simulating the operation of the equipment room based on a first setting model, the first setting model is constructed based on the equipment room, and the first setting model It includes: the plurality of cabinets of the first power, the at least one temperature adjustment device of the first temperature; the second inlet air temperature is obtained by simulating the operation of the computer room based on the second setting model The inlet air temperature of the first cabinet, the second setting model is constructed based on the equipment room, and the second setting model includes: the plurality of cabinets of the first power, all the cabinets of the second temperature the first temperature adjustment device, and other temperature adjustment devices other than the first temperature adjustment device in the at least one temperature adjustment device of the first temperature.
  • the first power, the first temperature, and the second temperature can be sent by the energy-saving device to the simulation device, so that the simulation device can set the operation of the cabinet or the temperature adjustment device in the first setting model and the second setting model based on these data. state.
  • the first setting model and the second setting model can be understood as the computer room computational fluid dynamics model introduced in the above content, and the first setting model and the second setting model are constructed according to the position information of the actual constituent objects in the computer room.
  • the request for setting parameters (that is, including the first power and the first temperature) of the first setting model sent by the energy-saving device to the simulation device may be:
  • the request sent by the energy-saving device to the simulation device for setting parameters of the second setting model (that is, including the first power, the first temperature, and the second temperature) may be:
  • the simulation device obtains the first inlet air temperature of the first cabinet obtained by simulating the operation of the computer room based on the first setting model, and the first cabinet obtained by simulating the operation of the computer room based on the second setting model the second inlet air temperature.
  • the energy-saving device can obtain the first inlet air temperature and the second inlet air temperature from the simulated device. It should be noted that the difference between the first setting model and the second setting model is only that the temperature of the first temperature adjustment device is adjusted, and the temperature of other temperature adjustment devices and the power of each cabinet remain unchanged, which can be understood as , which limits the influence of other factors on the inlet air temperature of the first cabinet; in this way, information on the cooling capacity conduction relationship between the first cabinet and the first temperature regulating device can be determined.
  • the embodiment of the present application uses a temperature correlation index (TCI) to describe the cooling capacity of a cabinet to a temperature conditioning device.
  • TCI temperature correlation index
  • the larger the value the stronger the cooling capacity of a cabinet to a temperature adjustment device; in other words, the stronger the correlation between a cabinet and a temperature adjustment device.
  • calculation method of the TCI of the first cabinet and the first temperature adjustment device may refer to formula 1-1:
  • the cooling capacity conduction relationship information between the first cabinet and the first temperature adjustment device may be an association relationship between the first cabinet, the first temperature adjustment device, and the TCI 11 . In a similar manner as described above, the information on the cooling transfer relationship of other cabinets and other temperature-adjusting equipment can be calculated.
  • the row identifiers "L1", “L2”, “R1”, and “R2” represent the identifiers (or labels, numbers) of the cabinets, and the identifiers correspond to the location information of the cabinets.
  • “L1” represents the first cabinet on the left in the schematic diagram of the computer room in Figure 1
  • “R2” represents the second cabinet on the right in the schematic diagram of the computer room in Figure 1).
  • the column identifiers "A1”, “A2”, “A3”, and “A4" represent the identifiers of the temperature adjustment devices, and the identifiers correspond to the location information of the temperature adjustment devices.
  • “A1" represents the first air conditioner on the left in the schematic diagram of the computer room in Figure 1.
  • each unit represents the TCI value of the corresponding cabinet and the corresponding temperature adjustment device.
  • the TCI value is 0.64, which indicates the TCI value of the cabinet L1 and the temperature adjustment device A1. is 0.64; the size of this value represents the cooling capacity of the cabinet L1 to the temperature adjustment device A1.
  • the attribution relationship (or referred to as a strong association relationship) between the cabinet and the temperature adjustment device may be determined according to the cooling capacity conduction relationship information.
  • the cabinet L1 if the TCI between the cabinet L1 and the temperature adjustment device A1 is greater than the TCI between the cabinet L1 and other temperature adjustment devices, it can be considered that there is an attribution relationship between the cabinet L1 and the temperature adjustment device A1, that is, , the cabinet L1 is the cabinet belonging to the temperature adjustment device A1, and the temperature adjustment device A1 is the belonging temperature adjustment device of the cabinet L1.
  • the energy-saving device triggers the facility management device to execute the computing resource management instruction set according to the cooling conduction relationship information and the computing resource state information, so as to minimize the total power of the computer room while meeting service requirements.
  • the total power of the computer room includes the computing power of multiple cabinets in the computer room (also referred to as IT power, etc.) and the temperature control power of at least one temperature adjustment device in the computer room (also referred to as cooling power or heating power, and many more).
  • the meaning of minimizing the total power of the computer room is to execute the computing resource management instruction set so that the total power of the computer room is smaller than the total power of the computer room after other management instruction sets are executed.
  • the computing resource management instruction set is a computing resource adjustment instruction set, and the computing resource adjustment instruction set is used to adjust the running position of at least one computing resource in the computer room. It can be understood that at least one computing resource can be a virtual machine or a task.
  • the computing resource adjustment instruction set may include one or more adjustment instructions for computing resources. After the facility management device executes the computing resource adjustment instruction set, the operating position of at least one computing resource can be adjusted, and the IT load can be redistributed, which affects computing power and temperature control power. In this embodiment of the present application, the total power of the computer room after the computing resource adjustment instruction set is executed is smaller than the total power of the computer room before execution.
  • the process of triggering the facility management device to execute the computing resource management instruction set according to the cooling conduction relationship information and the computing resource status information by the energy-saving device may include the following steps:
  • the computer room can periodically update the running location of virtual machines and/or tasks. In this case, all virtual machines and/or tasks contained in the computer room can be used as the computing resources to be migrated.
  • each virtual machine or task has a set update period.
  • the set update period of different virtual machines or tasks may be the same or different.
  • these virtual machines or tasks can be used as computing resources to be migrated.
  • the computer room can first allocate temporary computing resources for the business requirements, so as to respond in time. These temporary computing resources should be adjusted after a certain period of time according to business requirements. In this case, these temporary computing resources can be used as computing resources to be migrated.
  • the service requirement in this scenario may be an online service, and this type of service is a delay-sensitive service and has high requirements on the response speed of the service.
  • the real-time end-user-oriented service is an online service.
  • the service requirement may be an e-commerce service, a search service, a recommendation service, and the like.
  • the manner of determining a target server for the computing resource to be migrated based on the cooling conduction relationship information and the computing resource state information may be: based on the cooling conduction relationship information and the computing
  • the resource status information determines a plurality of candidate servers and the candidate total power corresponding to each candidate server for the computing resource to be migrated, and the total candidate power is assumed to be after the computing resource to be migrated is migrated to each candidate server.
  • the total power of the computer room; and the candidate server with the lowest total power of the corresponding candidate servers is selected as the target server among the plurality of candidate servers.
  • the number of CPU cores of the candidate server is required to be not less than the number of CPU cores of the computing resources to be migrated and the number of CPU cores of the candidate server. and/or, the memory specification of the candidate server is not less than the memory specification of the computing resource to be migrated and the memory specification of the existing computing resource in the candidate server Sum.
  • further requirements may be made on the candidate server according to the business characteristics of the computing resources to be migrated. For example, if the computing resources to be migrated are an authentication task, the candidate server needs to have an authentication function. It should be noted that some other requirements may also be added to the candidate server, which are not specifically limited in this embodiment of the present application. In this implementation manner, all servers in the computer room that meet the operating requirements of the computing resources to be migrated can be used as candidate servers.
  • the energy-saving device may first select a server that meets the operating requirements of the computing resources to be migrated (see the introduction of the first possible implementation manner) from multiple servers in the computer room as an alternative server. Afterwards, based on the computing resource status information, determine the power variation corresponding to each alternative server, the power variation is assumed to be after the computing resource to be migrated is migrated to each alternative server, the The amount of power variation for each candidate server.
  • the candidate servers with the smallest corresponding power variation on each cabinet are formed into the multiple candidate servers.
  • the method further includes: according to the number of CPU cores of the computing resources to be migrated, the CPU utilization of the computing resources to be migrated, and the number of CPU cores of the candidate server, the empty number of the candidate server.
  • One or more items of load power consumption or full load power consumption are determined, assuming that after the to-be-migrated computing resource is migrated to the candidate server, the amount of power change of the candidate server.
  • the calculation method of the power variation ⁇ P placed in the server k by the virtual machine j is introduced.
  • This section describes the calculation method of the power variation ⁇ P put into the server k by the virtual machine creation task 1.
  • the energy-saving device may first select a server that meets the operating requirements of the computing resources to be migrated (see the introduction of the first possible implementation manner) from multiple servers in the computer room as an alternative server Afterwards, based on the cooling conduction relationship information and computing resource status information, determine a plurality of candidate cabinets in the computer room with the largest TCI of each temperature adjustment device (for example, take Table 1 as an example, and the temperature adjustment device The server with the largest TCI of A1 is server L4), and the candidate servers on multiple candidate cabinets are used as the multiple candidate servers.
  • multiple candidate servers have the largest TCI with the temperature adjustment device, indicating that the stronger the ability of the candidate servers to absorb cooling, the more computing resources to be migrated are migrated to these servers, and the temperature control power of the temperature adjustment device will be reduced. If it is reduced, the total power of the computer room after the migration can be made smaller than the total power of the computer room before the migration.
  • the computing resources to be migrated may include multiple computing resources, a target server may be determined for each computing resource to be migrated, and the target servers of different computing resources to be migrated may be the same or different.
  • the computing resource instructions for migrating each computing resource to be migrated to its corresponding target server are combined into the computing resource adjustment instruction set.
  • target servers may be determined for the computing resources to be migrated in descending order of CPU utilization. Since the computing resources to be migrated with higher CPU utilization occupy higher power, in this way, the computing resources to be migrated that have a greater impact on the total power of the equipment room can be migrated first, which is beneficial to reduce the total power of the equipment room.
  • the running status information of the target server needs to be updated, because the target server needs to reserve computing resources for the computing resource to be migrated. That is to say, in the process of selecting the target server corresponding to the next computing resource to be migrated, the computing resource state information updated based on the selection result of the previous computing resource to be selected is used.
  • virtual machine 1 is assumed to be on server 1 before the migration, and it is assumed to be on server 2 after the migration. It is assumed that the changes in the data structure (or running location information) of the attribution relationship between the server and the virtual machine before and after the migration are as follows: :
  • the following describes the manner of determining the total candidate power corresponding to each candidate server.
  • the method may include the following steps:
  • Step a1 Obtain the power of each of the plurality of cabinets after assuming that the computing resources to be migrated are migrated to a candidate server.
  • the CPU utilization of the server is equal to the sum of the CPU loads of all virtual machines on the server, plus the sum of the CPU requirements of the virtual machine creation tasks on the server, divided by the number of CPU cores of the server, where the virtual machine CPU load is equal to the number of virtual machine CPU cores Multiply by the virtual machine CPU utilization. Indicates the CPU utilization of server k.
  • the server contains virtual machines, in, represents the power of server k, is the full load power of server k, is the no-load power of server k, the power of the server is linearly related to the CPU utilization of the server.
  • the rack power is equal to the sum of the power of all servers on the rack. Indicates the power of the i-th cabinet.
  • Step b1 Determine the calculated power of the multiple cabinets according to the power of each cabinet. Specifically, the calculated power of the multiple cabinets is the sum of the powers of each cabinet.
  • Step c1 Determine a temperature setting parameter of each temperature adjusting device in the at least one temperature adjusting device according to the power of each cabinet in the plurality of cabinets, where the temperature setting parameter is used to meet the requirements of executing the computing resources to be migrated.
  • the method of determining the temperature setting parameters of each temperature regulating device can be referred to as follows:
  • the energy-saving equipment compares the power of each cabinet (calculated in step S21), the simulated temperature value of the temperature adjustment equipment (initially, the operating state of each temperature adjustment equipment is ON, and the initial temperature is the maximum temperature Z that can be set).
  • °C eg, 35 °C
  • the simulation device simulates the operation of the computer room based on the input parameter. It should be noted that, if the power of all the cabinets that use one temperature adjustment device as the attributable temperature adjustment device is 0, the operating state of the one temperature adjustment device may be set to be off.
  • the request for input parameters sent by the energy-saving device to the simulated device may be:
  • the output result obtained by the simulation device simulating the operation of the computer room based on the input parameters may be:
  • the analog temperature value of the temperature adjustment device with the largest TCI of the cabinet is adjusted down to the preset value (for example, , 1°C), and then again obtain the inlet air temperature of each cabinet according to the adjusted simulated temperature value; until each cabinet T i ⁇ the upper temperature limit of the cabinet.
  • the corresponding simulated temperature value of each temperature regulating device is determined as the temperature setting parameter of each temperature regulating device.
  • the simulated temperature value of each temperature adjustment device is determined as the temperature setting parameter of each temperature adjustment device.
  • Step d1 Calculate the power of each temperature regulating device based on the inlet air temperature and outlet air temperature of each temperature regulating device corresponding to the temperature setting parameters of each temperature regulating device (which can be obtained from step c1).
  • the power of a temperature regulating device is:
  • C is the specific heat capacity of air, which takes the value of 1005.
  • the calculation formula of the coefficient of performance (COP; representing the cooling capacity that can be obtained per unit power consumption) and 0.0068, 0.0008, and 0.458 in the calculation formula are obtained from a large number of experiments. coefficient value. Indicates the power of the jth air conditioner.
  • Step e1 Determine the temperature control power of the at least one temperature adjustment device according to the power of each temperature adjustment device. Specifically, the temperature control power of the at least one temperature adjustment device is the sum of the powers of each temperature adjustment device.
  • Step f1 Determine the candidate total power corresponding to the one candidate server according to the computing power of the multiple cabinets and the temperature control power of at least one temperature adjustment device.
  • the energy-saving device may select a candidate server with the lowest total power among the multiple candidate servers as the target server.
  • the energy-saving device determines the target server for the computing resource to be migrated based on the cooling capacity conduction relationship information and the computing resource status information, and may further be: based on the computing resource status information, for all the computing resources.
  • the to-be-migrated computing resource determines an alternative server on each cabinet in the computer room; wherein the first power change is not greater than the second power change, and the first power change is assumed to be the calculation of the to-be-migrated After the resource is migrated to the alternative server, the power change of the alternative server, and the second power change is the assumption that the computing resource to be migrated is migrated to the same cabinet except the alternative server.
  • the power variation of the other servers based on the computing resource status information and the cooling conduction relationship information, determine the power balance degree corresponding to each cabinet, and the power balance degree is assumed to be
  • the power balance of the computer room After the to-be-migrated computing resources are migrated to the candidate servers on each cabinet, the power balance of the computer room; the candidate server on the cabinet with the smallest corresponding power balance is selected from the plurality of cabinets as all the cabinets. the target server.
  • the energy-saving device can first select some servers that meet the operating requirements of the computing resources to be migrated (see the introduction in the above content) from multiple servers in the computer room, and then select an alternative server from these servers. It is assumed that after the to-be-migrated computing resources are migrated to a certain server, the power variation of a certain server can be referred to the introduction in the above content, and will not be repeated here.
  • the method further includes: according to the number of CPU cores of the target computing resource, the CPU utilization rate of the target computing resource, the number of CPU cores of the candidate server, and the no-load power of the candidate server.
  • One or more items of power consumption or full load power consumption are determined, assuming that after the target computing resource is migrated to the alternative server, the amount of power change of the alternative server.
  • the energy-saving device may determine, according to the computing resource status information, the number of CPU cores of the target computing resource, the CPU utilization rate of the target computing resource, the number of CPU cores of the candidate server, and the empty number of the candidate server.
  • One or more of load power consumption or full load power consumption may be determined, according to the computing resource status information, the number of CPU cores of the target computing resource, the CPU utilization rate of the target computing resource, the number of CPU cores of the candidate server, and the empty number of the candidate server.
  • the power of each rack, the power of each rack and the The cooling capacity conduction relationship information of the corresponding attributable temperature adjustment equipment is determined, and the normalized power of each cabinet is determined; the power balance degree corresponding to each cabinet is determined according to the normalized power of each cabinet.
  • the server k is the server on cabinet i.
  • Step a2 CPU utilization S cpu-u of each server in the computer room:
  • the CPU utilization of the server is equal to the sum of the CPU loads of all virtual machines on the server, plus the sum of the CPU requirements of the virtual machine creation tasks on the server, divided by the number of CPU cores of the server, where the virtual machine CPU load is equal to the number of virtual machine CPU cores Multiply by the virtual machine CPU utilization. Indicates the CPU utilization of server k.
  • Step b2 the power P server of each server in the computer room:
  • the server contains virtual machines, in, represents the power of server k, is the full load power of server k, is the no-load power of server k, the power of the server is linearly related to the CPU utilization of the server.
  • Step c2 the power P rack of each cabinet in the computer room: the power of the cabinet is equal to the sum of the power of all servers on the cabinet. Indicates the power of the i-th cabinet.
  • Step d2 the normalized power ⁇ P rack of each cabinet in the computer room:
  • the attributable temperature adjustment device corresponding to the cabinet is determined according to Table 1.
  • the temperature adjustment device is the temperature adjustment device j.
  • Step e2 normalized power of all cabinets in the computer room
  • the standard deviation of , and the standard deviation (which can also be the variance) is used as the power balance degree corresponding to the cabinet i.
  • the power balance corresponding to each cabinet in the equipment room can be determined. Afterwards, the candidate server on the cabinet with the minimum corresponding power balance degree is selected as the target server among the plurality of cabinets. It should be noted that, the smaller the power balance, the more balanced the power among the cabinets in the equipment room, which will make the temperature control power of the temperature adjustment equipment smaller. In this way, the alternative servers on each cabinet in the equipment room can be selected, and the multiple alternative servers are the servers with the smallest corresponding power changes. Assuming that the computing resources to be migrated are migrated to these servers, the Computational power will be reduced.
  • the candidate server on the cabinet with the smallest corresponding power balance from the multiple cabinets as the target server, it is possible to reduce the temperature control power of the temperature adjustment device after the assumption that the migration is made, and then it is possible to reduce the temperature control power of the temperature adjustment device after migration.
  • the total power of the computer room is less than the total power of the computer room before the migration.
  • the method further includes: calculating the power of each cabinet in the plurality of cabinets after the computing resource adjustment instruction set is assumed to be executed; according to the power of each cabinet in the plurality of cabinets determining a temperature setting parameter of each temperature adjusting device in the at least one temperature adjusting device, where the temperature setting parameter is used to meet the temperature requirements of the multiple cabinets after executing the computing resource adjustment instruction set; triggering all
  • the facility management device executes a temperature adjustment instruction set, and the temperature adjustment instruction set is generated according to the temperature setting parameter.
  • the temperature adjustment instruction set may include one or more temperature adjustment instructions, and the temperature adjustment instructions can adjust the operating state (on or off) of the temperature adjustment device, and can also adjust the set temperature of the temperature adjustment device.
  • step a1 for the manner of determining the temperature setting parameters of each temperature adjusting device, reference may be made to the introduction of step a1 to step c1 in the above content.
  • step c1 for the manner of determining the temperature setting parameters of each temperature adjusting device, reference may be made to the introduction of step a1 to step c1 in the above content.
  • the computing resource adjustment instruction set and the temperature adjustment instruction set may be:
  • the computing resource management instruction set is a computing resource allocation instruction set, and the computing resource allocation instruction set is used to allocate at least one computing resource in the computer room. It can be understood that at least one computing resource can be a virtual machine or a task.
  • the computing resource allocation instruction set may include one or more allocation instructions for computing resources.
  • the increment of the total power of the computer room before and after the execution of the computing resource allocation instruction set is minimized. It should be noted that the meaning of minimizing the increment of the total power of the computer room is to execute the computing resource allocation instruction set, so that the increment of the total power of the computer room is smaller than the increase of the total power of the computer room caused by the execution of other allocation instruction sets. .
  • the process of triggering the facility management device to execute the computing resource management instruction set according to the cooling conduction relationship information and the computing resource status information by the energy-saving device may include the following steps:
  • the business demand may be a business demand from outside the computer room, or may be a business demand generated by the computer room itself.
  • the increment of the total power of the computer room before and after the target server provides the target computing resource is minimized.
  • the computer room can allocate target computing resources for the business requirements in real time, and determine a target server that provides the target computing resources.
  • the service requirement in this scenario may be a computing service, which is generally not sensitive to delay and does not require high service response speed.
  • the service requirement may be a spark service, an mr (mapreduce) service, a graph computing service, an audio and video codec service, etc.; the service requirement may come from a device outside the computer room, or from a device inside the computer room. equipment.
  • the manner of determining the target server for the computing resource to be migrated based on the business requirement, the cooling capacity conduction relationship information and the computing resource state information may be: based on the business requirement, all the The cooling conduction relationship information and the computing resource status information determine multiple candidate servers and a candidate increment corresponding to each candidate server for the target computing resource, and the candidate increment is assumed to migrate the target computing resource to The increment of the total power of the computer room after each candidate server; in the multiple candidate server sets, the candidate server with the lowest corresponding candidate increment is selected as the target server.
  • the number of CPU cores of the candidate server is not less than the number of CPU cores of the target computing resource and the number of CPU cores already in the candidate server.
  • the sum of the number of CPU cores of the existing computing resources; and/or, the memory specification of the candidate server is not less than the sum of the memory specification of the target computing resource and the memory specification of the existing computing resources in the candidate server.
  • further requirements may be made on the candidate server according to business requirements. For example, if the business is an authentication task, the candidate server needs to have an authentication function. It should be noted that some other requirements may also be added to the candidate server, which are not specifically limited in this embodiment of the present application.
  • all servers in the computer room that meet the operating requirements of the target computing resources can be used as candidate servers.
  • the energy-saving device may first select a server that meets the operating requirements of the target computing resource (see the introduction of the first possible implementation manner) from multiple servers in the computer room as an alternative server; After that, based on the computing resource state information, determine the power change amount corresponding to each candidate server, where the power change amount is assumed to be after the target computing resource is migrated to each candidate server. Select the power variation of the server.
  • the candidate servers with the smallest corresponding power variation on each cabinet are formed into the multiple candidate servers. In this way, multiple candidate servers are servers with the smallest corresponding power changes, then the target computing resources are allocated to these servers, and the increment of the computing power of the servers is small, which can increase the total power of the allocated computer room. The amount is low.
  • the method further includes: according to the number of CPU cores of the target computing resource, the CPU utilization rate of the target computing resource, the number of CPU cores of the candidate server, and the no-load power of the candidate server.
  • One or more items of power consumption or full load power consumption are determined, assuming that after the target computing resource is migrated to the alternative server, the amount of power change of the alternative server.
  • the energy-saving device may first select a server that meets the operating requirements of the target computing resources (see the introduction of the first possible implementation manner) from multiple servers in the computer room as an alternative server; After that, based on the cooling conduction relationship information and the computing resource status information, determine a plurality of candidate cabinets in the equipment room with the largest TCI with each temperature adjustment device (for example, using Table 1 as an example, with the temperature adjustment device A1 The server with the largest TCI is server L4), and the candidate servers on multiple candidate cabinets are used as the multiple candidate servers.
  • multiple candidate servers have the largest TCI with the temperature adjustment device, indicating that the stronger the ability of the candidate servers to absorb cooling capacity, the target computing resources are allocated to these servers, and the increase in the temperature control power of the temperature adjustment device Smaller, which can make the increment of the total power of the migrated computer room lower.
  • the target computing resource may include multiple computing resources, and a target server may be determined for each target computing resource, and the target servers of different target computing resources may be the same or different.
  • the computing resource instructions allocated to each target computing resource to its corresponding target server are combined into the computing resource allocation instruction set.
  • the target server may be determined for the target computing resource in descending order of CPU utilization. Since the target computing resources with higher CPU utilization occupy higher power, in this way, the target computing resources that have a greater impact on the total power of the computer room can be migrated first, which is beneficial to lower the increment of the total power of the computer room.
  • the running status information of the target server needs to be updated, because the target server needs to reserve computing resources for the target computing resource. That is to say, in the process of selecting the target server corresponding to the next target computing resource, the computing resource state information updated based on the selection result of the last candidate computing resource is used.
  • the following describes how to determine the candidate increment corresponding to each candidate server.
  • the method may include the following steps:
  • Step a3 Obtain the first total power of the computer room before allocating the target computing resource to a candidate server under the assumption that execution is performed.
  • Step b3 Obtain the power of each of the plurality of cabinets after allocating the target computing resource to a candidate server in a hypothetical execution.
  • Step c3 Determine the calculated power of the multiple cabinets according to the power of each cabinet. Specifically, the calculated power of the multiple cabinets is the sum of the powers of each cabinet.
  • Step d3 Determine, according to the power of each of the plurality of cabinets, a temperature setting parameter of each temperature adjusting device in the at least one temperature adjusting device, where the temperature setting parameter is used to meet the requirements of executing the target computing resource allocation. After reaching a candidate server, the temperature requirements of the plurality of racks.
  • Step e3 Calculate the power of each temperature regulating device based on the inlet air temperature and outlet air temperature of each temperature regulating device corresponding to the temperature setting parameters of each temperature regulating device (which can be obtained from step d3).
  • Step f3 Determine the temperature control power of the at least one temperature adjustment device according to the power of each temperature adjustment device. Specifically, the temperature control power of the at least one temperature adjustment device is the sum of the powers of each temperature adjustment device.
  • Step g3 Determine the second total power of the computer room corresponding to the one candidate server according to the computing power of the multiple cabinets and the temperature control power of the at least one temperature adjustment device, and the candidate increment corresponding to the one candidate server is the second total power The difference from the first total power.
  • step b3-step g3 For the execution manner of step b3-step g3, reference may be made to the execution manner of step a1-step f1 introduced in the above content.
  • the candidate increment corresponding to each candidate server can be determined. Afterwards, the candidate server with the lowest corresponding candidate increment is selected from the multiple candidate server sets as the target server.
  • the method of determining the target server for the target computing resource based on the cooling capacity conduction relationship information and the computing resource status information may further be: based on the computing resource status information, determine the target server for the target computing resource.
  • the target computing resource determines an alternative server on each cabinet in the computer room; wherein, the third power change amount is not greater than the fourth power change amount, and the third power change amount is assumed to migrate the target computing resources to After the candidate server, the power variation of the candidate server, the fourth power variation is assumed to be after the target computing resource is migrated to other servers other than the candidate server on the same cabinet,
  • the energy-saving device can first select some servers that meet the operating requirements of the target computing resources (see the introduction of the first possible implementation) from multiple servers in the computer room, and then select the server from this part of the server. Select the server. It is assumed that after the target computing resource is migrated to a certain server, the power variation of a certain server can be referred to the introduction in the above content, and will not be repeated here.
  • the method further includes: according to the number of CPU cores of the target computing resource, the CPU utilization rate of the target computing resource, the number of CPU cores of the candidate server, and the no-load power of the candidate server.
  • One or more items of power consumption or full load power consumption are determined, assuming that after the target computing resource is migrated to the alternative server, the amount of power change of the alternative server.
  • the energy-saving device may determine, according to the computing resource status information, the number of CPU cores of the target computing resource, the CPU utilization rate of the target computing resource, the number of CPU cores of the candidate server, and the empty number of the candidate server.
  • One or more of load power consumption or full load power consumption may be determined, according to the computing resource status information, the number of CPU cores of the target computing resource, the CPU utilization rate of the target computing resource, the number of CPU cores of the candidate server, and the empty number of the candidate server.
  • the power of each rack, the power of each rack and the The cooling capacity conduction relationship information of the corresponding attributable temperature adjustment equipment is determined, and the normalized power of each cabinet is determined; the power balance degree corresponding to each cabinet is determined according to the normalized power of each cabinet.
  • the power balance corresponding to each cabinet in the equipment room can be determined. Afterwards, the candidate server on the cabinet with the minimum corresponding power balance degree is selected as the target server among the plurality of cabinets. It should be noted that, the smaller the power balance, the more balanced the power among the cabinets in the equipment room, which will make the temperature control power of the temperature adjustment equipment smaller. In this way, the alternative servers on each cabinet in the computer room can be selected, and the multiple alternative servers are the servers with the smallest corresponding power change, then assuming that the target computing resources are allocated to these servers, the computing The increase in power will be smaller.
  • the method further includes: calculating, assuming that the computing resource allocation instruction set is executed, the power of each cabinet in the plurality of cabinets; according to the power of each cabinet in the plurality of cabinets determining a temperature setting parameter of each temperature adjusting device in the at least one temperature adjusting device, where the temperature setting parameter is used to meet the temperature requirements of the multiple cabinets after executing the computing resource allocation instruction set; triggering all
  • the facility management device executes a temperature adjustment instruction set, and the temperature adjustment instruction set is generated according to the temperature setting parameter.
  • the energy-saving equipment may partition the cabinets and the temperature adjustment equipment included in the equipment room.
  • a partition includes a temperature conditioning device and at least one cabinet. Afterwards, from these partitions, the candidate partition combination with stronger cooling capacity absorption and cooling capacity is determined. Migrate the computing resources to be migrated to the servers in the combination of partitions to be selected, or allocate the target computing resources to the servers in the combination of partitions to be selected; because the combination of partitions to be selected has strong cooling capacity and cooling capacity Sharing capability, in this way, the temperature control power of the temperature regulation equipment can be reduced, thereby minimizing the total power of the equipment room.
  • the cabinets and temperature adjustment equipment included in the computer room may be partitioned based on the cooling conduction relationship information of the computer room.
  • the equipment room includes a first partition, and the first partition includes a second temperature adjustment device and at least one second cabinet; wherein, each second cabinet has a cooling capacity of the second temperature adjustment device , which is stronger than the cold energy absorption capacity of each second cabinet to other temperature adjustment devices in the plurality of temperature adjustment devices except the second temperature adjustment device.
  • a partition includes a temperature adjustment device and a cabinet with the strongest absorption capacity for the cooling energy of the temperature adjustment device. It can be understood from the attribution relationship between the cabinet and the temperature adjustment device described in the above content, a partition includes one temperature adjustment device, and the cabinet with this temperature adjustment device as the attributable temperature adjustment device.
  • Table 2 for the partition mode of the equipment room.
  • each unit under the "Partition” column is the number of the partition
  • the value of each unit under the "Temperature adjustment equipment” column is the number of the temperature adjustment equipment in the corresponding partition
  • the value of each unit under the "Cabinet” column Indicates the number of a group of cabinets in the corresponding partition.
  • the information on the partition cooling capacity absorption relationship of the zones and the cooling capacity sharing relationship information between the two zones may be determined according to the cooling capacity conduction relationship information.
  • the partition cooling energy absorption relationship information is used to indicate the cooling energy absorption capacity of the cabinets in each partition in the equipment room to the temperature adjustment equipment in this partition, and one partition includes one temperature adjustment equipment and at least one cabinet.
  • the cooling capacity sharing relationship information is used to indicate, between every two partitions in the equipment room, the cooling capacity sharing capability of a cabinet in one partition to a temperature regulating device in another partition.
  • the computer room includes a first partition
  • the first partition includes a second temperature adjustment device and at least one second cabinet.
  • the method of the embodiment of the present application further includes: determining, based on the cooling capacity of each of the at least one second cabinet, the cooling capacity of the second temperature regulating device, the zone cooling absorption relationship information of the first zone.
  • the partition cooling capacity absorption relationship information of the first partition is an association relationship between the cabinets in the first partition, the temperature adjustment equipment in the first partition, and a partition temperature index (partition temperature index, PTI).
  • PTI partition temperature index
  • PTI partition temperature index
  • the value of the PTI of the first partition is the sum of the values of the TCIs of each cabinet in the first partition and the temperature regulating devices in the first partition.
  • the value of each cell under the column "PTI" represents the PTI value of the corresponding partition.
  • the equipment room further includes a second partition, where the second partition includes a third temperature adjustment device and at least one third cabinet; the method in this embodiment of the present application further includes: based on each of the at least one second cabinets The cold absorption capacity of the third temperature conditioning equipment and the cold absorption capacity of each of the at least one third cabinet to the second temperature conditioning equipment are determined, and the first partition is related to the The cooling capacity sharing relationship information of the second partition.
  • the cooling capacity sharing relationship information of the first partition and the second partition is the association relationship between the first partition, the second partition and a partition correlation index (partition correlation index, PCI).
  • PCI is used to describe the cooling capacity sharing capability of a cabinet in one partition to a temperature adjustment device in another partition between every two partitions.
  • the PTI value of the first partition and the second partition is the sum of the TCI values of each cabinet in the first partition and the temperature adjustment equipment in the second partition, and the value of each cabinet in the second partition and the second partition. The sum of the TCI values of the temperature regulating equipment, and half of the sum.
  • the cooling capacity sharing relationship information of the equipment room can be expressed as Table 4.
  • Partition 1 Partition 2 PCI P1 P2 0.4 P1 P3 1.13 P1 P4 0.29 P2 P3 0.27 P2 P4 1.21 P3 P4 0.35
  • each unit under the column “Partition 1" is the number of the first partition in the two partitions sharing cooling capacity
  • the value of each unit under the column “Partition 2" is the number of the first partition in the two partitions sharing cooling capacity.
  • the number of the two partitions, the value of each cell under the "PCI" column is the PCI between the first partition and the second partition.
  • the first row in the above-mentioned inter-partition PCI table indicates that the capacity value of the cooling capacity shared by zone P1 and zone P2 is 0.4, that is, the capacity value of all cabinets in zone P1 to absorb the cooling capacity provided by the temperature adjustment equipment in zone P2 is the same as that in zone P2.
  • the half of the sum of the capacity of all the cabinets to absorb the cooling capacity provided by the temperature adjustment equipment in the partition P1 is 0.4.
  • the combination set of partitions to be selected may be determined according to the partition cooling capacity absorption relationship information and the cooling capacity sharing relationship information.
  • the set of partition combinations to be selected includes a plurality of partition combinations to be selected with different numbers of partitions, wherein one partition combination to be selected is one or more partition combinations with the same number of partitions in cooling energy absorption capacity and cooling capacity sharing The most powerful partition combination.
  • a partition combination may include one or more partitions. Exemplarily, when a partition combination includes one partition, the partition combination may be P1, P2, P3, and P4.
  • the partition combination can be: partition combination of P1 and P2, partition combination of P1 and P3, partition combination of P1 and P4, partition combination of P2 and P3, partition combination of P2 and P4, partition combination of P3 and P4 .
  • the partition combination may be: the partition combination of P1, P2 and P3, the partition combination of P1, P3 and P4, and the partition combination of P2, P3 and P4.
  • the partition combination is the partition combination of P1, P2, P3 and P4.
  • the partition combination with the largest sum of the corresponding PTI and PCI is taken as the candidate partition combination. It should be noted that when there is only one partition in the partition combination, since the partition combination does not have the cooling capacity sharing capability, the PTI of the partition combination can be compared.
  • the combination set of partitions to be selected in the equipment room can be expressed as Table 5.
  • each unit under the column “Number of Partitions” is the number of partitions included in the combination of partitions to be selected
  • the value of each unit under the column of “Combination of Partitions to be selected” is the combination of partitions to be selected under the corresponding number of partitions
  • "PTI The value of each cell under the "sum of PCI” column is the sum of the intra-partition PTI and the inter-partition PCI of the corresponding partition combination to be selected.
  • the combination of partitions to be selected is the combination of partitions P2, P3 and P4.
  • the energy-saving device may determine multiple candidate servers and each candidate server for the to-be-migrated computing resource according to the cooling conduction relationship information, the partition combination set to be selected, and the computing resource state information The corresponding candidate total power, the multiple candidate servers belong to the candidate partition combination set.
  • the energy-saving device may determine multiple candidate servers for the to-be-migrated computing resource according to the cooling conduction relationship information, the partition combination set to be selected, and the computing resource state information, and may include the following steps:
  • the first power change is not greater than the second power change
  • the first power change is the power change of the alternative server assuming that the computing resource to be migrated is migrated to the alternative server
  • the second power variation assumes the power variation of the other servers after the to-be-migrated computing resources are migrated to other servers except the candidate server on the same cabinet. It should be noted that you can first select some servers that meet the operating requirements of the computing resources to be migrated (see the introduction in the above content) from the servers in each partition combination to be selected, and then select candidates from these servers. server.
  • the calculation assumes that after the to-be-migrated computing resource is migrated to a certain server, the power variation of a certain server can be referred to the introduction in the above content, and will not be repeated here.
  • the candidate server on each cabinet in each partition combination to be selected can be selected, and the multiple candidate servers are the servers with the smallest corresponding power change, then it is assumed that the computing resources to be migrated are migrated to these servers , the computing power of the server will be reduced, so that the total power of the computer room after the migration can be made smaller than the total power of the computer room before the migration.
  • S32 Determine, based on the computing resource state information and the cooling capacity conduction relationship information, a power balance degree corresponding to each cabinet, where the power balance degree assumes that the to-be-migrated computing resources are migrated to each of the cabinets. After the candidate servers on each cabinet, the power balance of the combination of the partitions to be selected to which each cabinet belongs.
  • the method further includes: according to the number of CPU cores of the target computing resource, the CPU utilization rate of the target computing resource, the number of CPU cores of the candidate server, and the no-load power of the candidate server.
  • One or more items of power consumption or full load power consumption are determined, assuming that after the target computing resource is migrated to the alternative server, the amount of power change of the alternative server.
  • the energy-saving device may determine, according to the computing resource status information, the number of CPU cores of the target computing resource, the CPU utilization rate of the target computing resource, the number of CPU cores of the candidate server, and the empty number of the candidate server.
  • One or more of load power consumption or full load power consumption may be determined, according to the computing resource status information, the number of CPU cores of the target computing resource, the CPU utilization rate of the target computing resource, the number of CPU cores of the candidate server, and the empty number of the candidate server.
  • the power of each rack, the power of each rack and the The cooling conduction relationship information of the temperature adjustment equipment in the partition (which can also be expressed as the cooling conduction relationship information of each cabinet and the temperature adjustment equipment corresponding to each cabinet) is determined, and each cabinet is determined.
  • the normalized power of each cabinet is determined; the power balance corresponding to each cabinet is determined according to the normalized power of each cabinet.
  • the server k is a server on the cabinet i, and the combination of partitions to be selected to which the cabinet belongs is the combination of partitions to be selected m.
  • Step a4 Calculate the CPU utilization S cpu-u of each server in the partition combination m to be selected:
  • the CPU utilization of the server is equal to the sum of the CPU load of all virtual machines on the server, plus the sum of the CPU requirements of the virtual machine creation tasks on the server, divided by the number of CPU cores of the server, where the virtual machine CPU load is equal to the number of virtual machine CPU cores Multiply by the virtual machine CPU utilization. Indicates the CPU utilization of server k.
  • Step b4 calculate the power P server of each server in the partition combination m to be selected:
  • the server contains virtual machines, in, represents the power of server k, is the full load power of server k, is the no-load power of server k, the power of the server is linearly related to the CPU utilization of the server.
  • Step c4 Calculate the power P rack of each cabinet in the partition combination m to be selected: the power of the cabinet is equal to the sum of the power of all servers on the cabinet. Indicates the power of the i-th cabinet.
  • Step d4 calculate the normalized power ⁇ P rack of each cabinet in the partition combination m to be selected:
  • the temperature adjustment equipment of the partition where the cabinet is located is obtained through the PTI table in the partition, and the temperature adjustment equipment number is denoted as j, and the TCI of the i-th cabinet and the j-th air conditioner is obtained through the TCI table, denoted as is tci ij , then the normalized power of cabinet i In the same way, the normalized power ⁇ P rack of each cabinet in the partition combination m to be selected can be calculated.
  • Step e4 calculate the normalized power of all cabinets in the partition combination m to be selected
  • the standard deviation of , and the standard deviation (which can also be the variance) is used as the power balance degree corresponding to the cabinet i.
  • the power balance degree corresponding to each cabinet in each combination of partitions to be selected can be determined.
  • the candidate servers on the cabinet with the corresponding minimum power balance are formed into the multiple candidate servers.
  • a candidate server on the cabinet with the smallest corresponding power balance may be determined, and these four servers constitute the multiple candidate servers. server.
  • the smaller the power balance is the more balanced the power among the cabinets in the partition combination to be selected is, which will reduce the temperature control power of the temperature adjustment device.
  • the computing resources to be migrated are migrated to these servers, and the computing power of the servers is reduced, so that the total power of the migrated computer room is smaller than the total power of the computer room before the migration.
  • the energy-saving device may then determine the candidate total power corresponding to each candidate server, and select the candidate server with the lowest corresponding candidate total power among the multiple candidate servers. as the target server.
  • the computing resources to be migrated may include multiple computing resources, a target server may be determined for each computing resource to be migrated, and the target servers of different computing resources to be migrated may be the same or different.
  • the computing resource instructions for migrating each computing resource to be migrated to its corresponding target server are combined into the computing resource adjustment instruction set.
  • the method further includes: calculating, assuming that the computing resource adjustment instruction set is executed, the power of each cabinet in the plurality of cabinets; adjusting the partition where the equipment is located according to each temperature
  • the power of the included cabinet determines the temperature setting parameters of each temperature adjustment device; triggering the facility management device to execute a temperature adjustment instruction set, the temperature adjustment instruction set is generated according to the temperature setting parameters.
  • the temperature setting parameter of a temperature adjustment device is determined by the power of the cabinet included in the partition where the temperature adjustment device is located. It can also be expressed as a temperature adjustment device that needs to meet the temperature requirements of the cabinet in its own partition.
  • step c1 Combined with the content in the method for determining the temperature setting parameters of each temperature adjustment device (step c1) described in the above content, in this implementation, if there is a cabinet's T i > the upper limit of the inlet air temperature of the cabinet (where the cabinet enters the temperature limit). If the upper limit of the air temperature is W°C, such as 28°C), the simulated temperature value of the temperature adjustment device in the partition where the cabinet is located can be lowered to a preset value (eg, 1°C), and then obtained again according to the lowered simulated temperature value. Inlet air temperature of each cabinet; until each cabinet T i ⁇ the upper temperature limit of the cabinet. Under the condition that each cabinet T i ⁇ the upper temperature limit of the cabinet, the corresponding simulated temperature value of each temperature regulating device is determined as the temperature setting parameter of each temperature regulating device.
  • a preset value eg, 1°C
  • the set of candidate computing resource adjustment instructions (examples include virtual machine migration instructions and task scheduling instructions) determined based on different combinations of candidate partitions output by the energy-saving device, and the set of instructions determined according to these candidate computing resource adjustment instruction sets.
  • An example of the calculation results of the calculated power, temperature control power, and temperature setting parameters of each temperature regulating device is as follows:
  • the candidate computing resource adjustment instruction set with the lowest total power of the corresponding candidate can be selected as the computer resource adjustment instruction set;
  • a temperature adjustment instruction set is generated by adjusting the temperature setting parameters of the device.
  • the energy-saving device may determine a plurality of candidate servers for the target computing resource according to the service requirement, the cooling conduction relationship information, the partition combination set to be selected, and the computing resource state information and a candidate increment corresponding to each candidate server, where the multiple candidate servers belong to the partition combination set to be selected.
  • the energy-saving device determines a plurality of candidate servers for the target computing resource according to the service requirement, the cooling conduction relationship information, the combination set of partitions to be selected, and the computing resource status information, which may include the following steps: :
  • Some servers capable of providing the target computing resources are selected from the servers in each candidate partition combination, and then candidate servers are selected from these servers.
  • the third power change amount is not greater than the fourth power change amount
  • the third power change amount is the power change amount of the alternative server assuming that the target computing resource is migrated to the alternative server.
  • the fourth power variation is the power variation of the other servers assuming that the target computing resource is migrated to other servers except the candidate server on the same cabinet.
  • the power variation of a certain server can be calculated by referring to the introduction in the above content, which will not be repeated here.
  • the candidate server on each cabinet in each partition combination to be selected can be selected, and the multiple candidate servers are the servers with the smallest corresponding power change, then the target computing resources are migrated to these servers, and the server The computing power is reduced, so that the total power of the computer room after the migration is smaller than the total power of the computer room before the migration.
  • the method further includes: according to the number of CPU cores of the target computing resource, the CPU utilization rate of the target computing resource, the number of CPU cores of the candidate server, and the no-load power of the candidate server.
  • One or more items of power consumption or full load power consumption are determined, assuming that after the target computing resource is migrated to the alternative server, the amount of power change of the alternative server.
  • the power of each rack, the power of each rack and the The cooling conduction relationship information of the temperature adjustment equipment in the partition (which can also be expressed as the cooling conduction relationship information of each cabinet and the temperature adjustment equipment corresponding to each cabinet) is determined, and each cabinet is determined.
  • the normalized power of each cabinet is determined; the power balance corresponding to each cabinet is determined according to the normalized power of each cabinet.
  • the candidate servers on the cabinet with the corresponding minimum power balance are formed into the multiple candidate servers.
  • a candidate server on the cabinet with the smallest corresponding power balance may be determined, and these four servers constitute the multiple candidate servers. server.
  • the smaller the power balance is the more balanced the power among the cabinets in the partition combination to be selected is, which will reduce the temperature control power of the temperature adjustment device.
  • the target computing resources are allocated to these servers, and the increment of the computing power of the servers is small, so that the increment of the total power of the computer room before and after the allocation is smaller than that of other computing resources. The increment of the total power of the computer room before and after the execution of the resource allocation instruction set.
  • the energy-saving device may then determine the increment of the total candidate power corresponding to each candidate server, and select the corresponding candidate total power among the multiple candidate servers.
  • the candidate server with the lowest increment is used as the target server.
  • the target computing resource may include multiple computing resources, and a target server may be determined for each target computing resource, and the target servers of different target computing resources may be the same or different.
  • the computing resource instructions allocated to each target computing resource to its corresponding target server are combined into the computing resource allocation instruction set.
  • the method further includes: calculating the power of each cabinet in the plurality of cabinets after the computing resource allocation instruction set is assumed to be executed; adjusting the partition where the equipment is located according to each temperature
  • the power of the included cabinet determines the temperature setting parameters of each temperature adjustment device; triggers the facility management device to execute a temperature adjustment instruction set, and the temperature adjustment instruction is generated according to the temperature setting parameters.
  • the temperature setting parameter of a temperature adjustment device is determined by the power of the cabinet included in the partition where the temperature adjustment device is located. It can also be expressed as, a temperature adjustment device needs to meet the temperature requirements of the cabinet in the partition where it is located.
  • step c1 Combined with the content in the method for determining the temperature setting parameters of each temperature adjustment device (step c1) described in the above content, in this implementation, if there is a cabinet's T i > the upper limit of the inlet air temperature of the cabinet (where the cabinet enters the temperature limit) If the upper limit of the wind temperature is W°C, such as 28°C), the simulated temperature value of the temperature adjustment device in the partition where the cabinet is located can be lowered to a preset value (eg, 1°C), and then obtained again according to the lowered simulated temperature value. Inlet air temperature of each cabinet; until each cabinet T i ⁇ the upper temperature limit of the cabinet. Under the condition that each cabinet T i ⁇ the upper temperature limit of the cabinet, the corresponding simulated temperature value of each temperature regulating device is determined as the temperature setting parameter of each temperature regulating device.
  • a preset value eg, 1°C
  • the energy-saving device may include a hardware structure and a software module, and implement the above functions in the form of a hardware structure, a software module, or a hardware structure plus a software module.
  • a certain function among the above functions may be implemented in the form of a hardware structure, a software module, or a hardware structure plus a software module.
  • FIG. 6 it is a schematic structural diagram of an energy-saving device provided by an embodiment of the present application.
  • the energy-saving device is used to save energy consumption of a computer room
  • the computer room includes a plurality of cabinets and at least one temperature adjustment device, wherein each cabinet includes at least one server, each server is used for providing computing resources, and each temperature adjustment device is used for adjusting The temperature of the room.
  • the computer room is also configured with facility management equipment, the facility management equipment is used to manage the computing resources and temperature adjustment equipment of the computer room, and the energy-saving equipment 60 includes an acquisition unit 601 and a processing unit 602, wherein:
  • the obtaining unit 601 is configured to obtain cooling conduction relationship information of the computer room and computing resource status information of the computer room. Specifically, for the operations performed by the obtaining unit 601, reference may be made to the description in step S101 in the method shown in FIG. 5 above.
  • a processing unit 602 configured to trigger the facility management device to execute a computing resource management instruction set according to the cooling conduction relationship information and the computing resource status information, so as to minimize the total power of the computer room while meeting business requirements change. Specifically, for the operations performed by the processing unit 602, reference may be made to the description in step S102 in the method shown in FIG. 5 above.
  • the cooling conduction relationship information is used to determine the cooling capacity of each cabinet in the equipment room to each temperature adjustment device in the equipment room
  • the computing resource status information is used to determine that the equipment room contains The running location and running status of the computing resources.
  • the computing resource management instruction set is a computing resource adjustment instruction set, and the computing resource adjustment instruction set is used to adjust the running position of at least one computing resource in the computer room.
  • the processing unit is specifically configured to: determine the computing resource to be migrated in the computer room based on the computing resource status information;
  • the target server is determined for the computing resources to be migrated, and the total power of the computer room after the computing resources to be migrated are migrated to the target server is less than the total power of the computer room before the migration; a generation for migrating the computing resources to be migrated to the target server is generated.
  • the instruction set of the target server is used as the computing resource adjustment instruction set.
  • the processing unit is specifically configured to: determine a plurality of candidate servers and a candidate total corresponding to each candidate server for the computing resource to be migrated based on the cooling capacity conduction relationship information and the computing resource state information power, the candidate total power is the total power of the computer room after assuming that the computing resources to be migrated are migrated to each candidate server; the candidate server with the lowest corresponding candidate total power is selected as the candidate server among the multiple candidate servers. the target server.
  • the processing unit is further configured to: obtain the power of each cabinet in the plurality of cabinets after the computing resource adjustment instruction set is assumed to be executed; according to the power of each cabinet in the plurality of cabinets determining a temperature setting parameter of each temperature adjusting device in the at least one temperature adjusting device, where the temperature setting parameter is used to meet the temperature requirements of the multiple cabinets after executing the computing resource adjustment instruction set; triggering all The facility management device executes a temperature adjustment instruction set, and the temperature adjustment instruction set is generated according to the temperature setting parameter.
  • the computing resource management instruction set is a computing resource allocation instruction set, and the computing resource allocation instruction set is used to allocate at least one computing resource in the computer room.
  • the processing unit is specifically configured to: receive business requirements, where the business requirements are used to request to run computing resources in the computer room; based on the business requirements, the cooling conduction relationship information, and the The computing resource status information is determined, the target computing resources that meet the business requirements and the target server that provides the target computing resources, and the target server provides the target computing resources.
  • the increment of the total power of the computer room is minimized; generating The instruction set for allocating the target computing resource on the target server is used as the computing resource allocation instruction set.
  • the processing unit is specifically configured to: determine multiple candidate servers and each candidate server for the target computing resource based on the business requirement, the cooling conduction relationship information, and the computing resource state information
  • the corresponding candidate increment, the candidate increment is the increment of the total power of the computer room after assuming that the target computing resource is migrated to each candidate server; select the corresponding candidate from the multiple candidate server sets The candidate server with the lowest increment is used as the target server.
  • the processing unit is further configured to: obtain the power of each cabinet in the plurality of cabinets after assuming that the computing resource allocation instruction set is executed; according to the power of each cabinet in the plurality of cabinets determining a temperature setting parameter of each temperature adjusting device in the at least one temperature adjusting device, where the temperature setting parameter is used to meet the temperature requirements of the multiple cabinets after executing the computing resource allocation instruction set; triggering all The facility management device executes a temperature adjustment instruction set, and the temperature adjustment instruction set is generated according to the temperature setting parameter.
  • each unit of the energy-saving device shown in FIG. 6 may be related to the foregoing method embodiments. It will not be described in detail here.
  • the above-mentioned units may be implemented in hardware, software or a combination of software and hardware.
  • the functions of the obtaining unit 601 and the processing unit 602 in the above content may be implemented by one or more processors in the energy-saving device 60 .
  • the facility management device can be triggered to execute the computing resource management instruction set according to the acquired cooling capacity conduction relationship information and computing resource status information, so as to meet the business requirements and make the total power of the equipment room minimize.
  • the energy-saving device 70 is used to save energy consumption of a computer room, the computer room includes a plurality of cabinets and at least one temperature adjustment device, wherein each cabinet includes at least one server, each server is used for providing computing resources, and each temperature adjustment device is used for The temperature of the machine room is adjusted.
  • the computer room is also configured with a facility management device, which is used to manage computing resources and temperature adjustment devices of the computer room.
  • the energy-saving device 70 may include: one or more processors 701; one or more input devices 702, one or more output devices 703 and memory 704.
  • the above-mentioned processor 701 , input device 702 , output device 703 and memory 704 are connected through a bus 705 .
  • Memory 704 is used to store instructions.
  • the so-called processor 701 may be a central processing unit, and the processor may also be other general-purpose processors, digital signal processors, application-specific integrated circuits, or other programmable logic devices.
  • a general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
  • the input device 702 may include a communication interface, a data line, etc.
  • the output device 703 may include a display (LCD, etc.), a data line, a communication interface, and the like.
  • the memory 704 which may include read-only memory and random access memory, provides instructions and data to the processor 701 .
  • a portion of memory 704 may also include non-volatile random access memory.
  • memory 704 may also store device type information.
  • the processor 701 is configured to execute the instructions stored in the memory 704 to perform the following operations:
  • the cooling conduction relationship information is used to determine the cooling capacity of each cabinet in the equipment room to each temperature adjustment device in the equipment room
  • the computing resource status information is used to determine that the equipment room contains The running location and running status of the computing resources.
  • processor 701 For operations performed by the processor 701, reference may be made to the related content of the foregoing method embodiments. It will not be described in detail here.
  • the facility management device can be triggered to execute the computing resource management instruction set according to the obtained cooling capacity conduction relationship information and computing resource status information, so as to meet the business requirements and make the total power of the equipment room minimize.
  • the present application also provides a computer-readable storage medium on which a computer program is stored, and when the computer-readable storage medium is executed by a computer, implements the functions of any of the foregoing method embodiments.
  • the present application also provides a computer program product, which implements the functions of any of the above method embodiments when the computer program product is executed by a computer.
  • the embodiment of the present application further provides a computing resource supply system, and the computing resource supply system may include the energy-saving equipment and the computer room in the embodiment corresponding to FIG. 5 .
  • the energy-saving equipment and the computer room reference may be made to the description of any of the foregoing method embodiments.
  • the above-mentioned embodiments it may be implemented in whole or in part by software, hardware, firmware or any combination thereof.
  • software it can be implemented in whole or in part in the form of a computer program product.
  • the computer program product includes one or more computer instructions. When the computer instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present application are generated.
  • the computer may be a general purpose computer, special purpose computer, computer network, or other programmable device.
  • the computer instructions may be stored in or transmitted from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions may be downloaded from a website site, computer, server or data center Transmission to another website site, computer, server, or data center by wire (eg, coaxial cable, optical fiber, digital subscriber line, DSL) or wireless (eg, infrared, wireless, microwave, etc.).
  • the computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. that includes an integration of one or more available media.
  • the available media may be magnetic media (eg, floppy disks, hard disks, magnetic tapes), optical media (eg, high-density digital video discs (DVDs)), or semiconductor media (eg, solid state disks, SSD)) etc.
  • the corresponding relationships shown in each table in this application may be configured or predefined.
  • the values of the information in each table are only examples, and can be configured with other values, which are not limited in this application.
  • the corresponding relationships shown in some rows may not be configured.
  • appropriate deformation adjustments can be made based on the above table, for example, splitting, merging, and so on.
  • the names of the parameters shown in the headings in the above tables may also adopt other names that can be understood by the communication device, and the values or representations of the parameters may also be other values or representations that the communication device can understand.
  • other data structures can also be used, such as arrays, queues, containers, stacks, linear lists, pointers, linked lists, trees, graphs, structures, classes, heaps, hash tables, or hash tables. Wait.
  • Predefined in this application may be understood as defining, predefining, storing, pre-storing, pre-negotiating, pre-configuring, curing, or pre-firing.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Microelectronics & Electronic Packaging (AREA)
  • Human Computer Interaction (AREA)
  • Thermal Sciences (AREA)
  • Software Systems (AREA)
  • Computer Hardware Design (AREA)
  • Power Sources (AREA)
  • Cooling Or The Like Of Electrical Apparatus (AREA)

Abstract

本申请实施例公开了一种应用于机房的节能方法、装置及系统，该机房包括多个机柜和至少一个温度调节设备，其中每一个机柜包括至少一个服务器，每一个服务器用于提供计算资源，每一个温度调节设备用于调节机房的温度，所述机房还配置有设施管理设备，所述设施管理设备用于管理所述机房的计算资源和温度调节设备，该方法包括：获取所述机房的冷量传导关系信息和所述机房的计算资源状态信息；根据所述冷量传导关系信息和所述计算资源状态信息，触发设施管理设备执行计算资源管理指令集，以在满足业务需求的同时使所述机房的总功率最小化。通过这种方式，可以在满足业务需求的同时使机房的总功率最小化。

Description

一种应用于机房的节能方法、装置及系统
本申请要求于2021年3月25日提交中国国家知识产权局、申请号为202110321118.4、申请名称为“一种应用于机房的节能方法、装置及系统”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及计算机技术领域，尤其涉及一种应用于机房的节能方法、装置及系统。
背景技术
机房中通常存放有服务器和空调，服务器为用户提供信息技术(Internet technology，IT)服务，空调用于控制机房的空气温度，使得服务器和机房中的其他配套装置能够工作在适宜的温度/湿度范围内。
机房的能耗主要由服务器所产生的IT能耗和空调所产生的制冷能耗构成，因此主要从这两个方面来考虑如何节省机房的能耗。由于服务器系统将实时接收来自外部的任务(如虚拟机创建任务)，通过任务调度算法将任务分配到其中一台服务器上执行，执行任务会占用服务器的中央处理器(central processing unit，CPU)资源，进而增加IT负载，从而增加IT能耗和制冷能耗。因此如何制定合适的任务调度算法，对于节省机房总能耗非常关键。另外，空调系统负责处理机房IT系统所产生的热量，按照能耗和热量的关系，IT系统所产生的热量约等于IT能耗。因此空调系统在满足制冷量需求的前提下，如何根据IT能耗动态的调节设定温度、出风量，对于节省机房制冷能耗非常关键。
现阶段，一般通过构建预测模型来确定出任务调度指令和温度设定参数。具体的，通过采集较长一段时间内服务器的资源利用率、机房空调系统参数、机房总能耗、外部环境参数等数据，据此进行模型训练得到总能耗预测模型、空调参数预测模型。基于预测模型依次预测每个任务调度到任意一台服务器上后的总能耗和空调参数，从而可以选取总能耗最小的服务器作为执行每个任务的目标服务器，对应的空调参数为空调的最终设定参数。但这种方式需要采集大量、多种情况下的实验数据对模型进行训练，对样本数据的依赖性较大；且样本数据较难满足实际机房运行中复杂多变的需求。如何降低机房的整体能耗是本领域技术人员亟待解决的问题。
发明内容
本申请提供一种应用于机房的节能方法及相关装置，可以在满足业务需求的同时使机房的总功率最小化。
第一方面，本申请提供了一种应用于机房的节能方法，所述机房包括多个机柜和至少一个温度调节设备，其中每一个机柜包括至少一个服务器，每一个服务器用于提供计算资源，每一个温度调节设备用于调节所述机房的温度，所述机房还配置有设施管理设备，所述设施管理设备用于管理所述机房的计算资源和温度调节设备，所述方法包括：获取所述机房的冷量传导关系信息和所述机房的计算资源状态信息；根据所述冷量传导关系信息和所述计算资 源状态信息，触发所述设施管理设备执行计算资源管理指令集，以在满足业务需求的同时使所述机房的总功率最小化；其中，所述冷量传导关系信息用于确定所述机房中的每一个机柜对所述机房中的每一个温度调节设备的冷量吸收能力，所述计算资源状态信息用于确定所述机房包含的计算资源的运行位置和运行状态。通过这种方法，可以在满足业务需求的同时使机房的总功率最小化。
结合第一方面，在一种可能的实现方式中，所述计算资源管理指令集为计算资源调整指令集，所述计算资源调整指令集用于调整所述机房中至少一个计算资源的运行位置。
结合第一方面，在一种可能的实现方式中，所述根据所述冷量传导关系信息和所述计算资源状态信息，触发所述设施管理设备执行计算资源管理指令集，包括：基于所述计算资源状态信息在所述机房中确定待迁移计算资源；基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定目标服务器，所述待迁移计算资源被迁移至所述目标服务器后所述机房的总功率小于迁移前所述机房的总功率；生成用于将所述待迁移计算资源迁移到所述目标服务器的指令集作为所述计算资源调整指令集。
结合第一方面，在一种可能的实现方式中，所述基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定目标服务器，包括：基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定多个候选服务器及每一个候选服务器对应的候选总功率，所述候选总功率为假设将所述待迁移计算资源迁移到所述每一个候选服务器后所述机房的总功率；在所述多个候选服务器中选择相应候选总功率最低的候选服务器作为所述目标服务器。通过选择相应候选总功率最低的候选服务器的方式，可以使得迁移后机房的总功率，低于迁移前该机房的总功率。
结合第一方面，在一种可能的实现方式中，所述基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定多个候选服务器及每一个候选服务器对应的候选总功率，包括：根据所述冷量传导关系信息确定分区冷量吸收关系信息和冷量共享关系信息；根据所述分区冷量吸收关系信息和所述冷量共享关系信息确定待选分区组合集合；根据所述冷量传导关系信息、所述待选分区组合集合和所述计算资源状态信息，为所述待迁移计算资源确定多个候选服务器及每一个候选服务器对应的候选总功率，所述多个候选服务器属于所述待选分区组合集合；其中，所述分区冷量吸收关系信息用于指示所述机房中各个分区中机柜对本分区中的温度调节设备的冷量吸收能力，一个分区包括一个温度调节设备以及至少一个机柜；所述冷量共享关系信息用于指示所述机房中每两个分区之间，一个分区的机柜对另一个分区的温度调节设备的冷量共享能力；所述待选分区组合集合包括具有不同分区数量的多个待选分区组合，其中，一个待选分区组合为具有相同分区数量的一个或多个分区组合中冷量吸收能力和冷量共享能力最强的分区组合。
结合第一方面，在一种可能的实现方式中，所述基于所述冷量传导关系信息、所述待选分区组合集合和所述计算资源状态信息，为所述待迁移计算资源确定多个候选服务器，包括：基于所述计算资源状态信息，为所述待迁移计算资源确定所述待选分区组合集合中每个待选分区组合中的每个机柜上的备选服务器；其中，第一功率变化量不大于第二功率变化量，所述第一功率变化量为假设将所述待迁移计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量，所述第二功率变化量为假设将所述待迁移计算资源迁移至同个机柜上除所述备选服务器外的其他服务器后，所述其他服务器的功率变化量；基于所述计算资源状态信息和所述冷量传导关系信息，确定所述每个机柜所对应的功率均衡度，所述功率均衡度为假设将所述待迁移计算资源迁移到所述每个机柜上的备选服务器后，所述每个机柜所属的待选分 区组合的功率均衡度；将所述每个待选分区组合中，相应功率均衡度最小的机柜上的备选服务器，组成所述多个候选服务器。其中，功率均衡度越小，表示该待选分区组合内的各个机柜之间的功率越均衡，这样会使得温度调节设备的温控功率较小。又因为备选服务器对应的功率变化量也较小，那么待迁移计算资源迁移到这些服务器中，服务器的计算功率减少，从而能使得迁移后的机房的总功率小于迁移前的机房的总功率。
结合第一方面，在一种可能的实现方式中，所述方法还包括：根据所述待迁移计算资源的CPU核数、所述待迁移计算资源的CPU利用率、所述备选服务器的CPU核数，所述备选服务器的空载功耗或者满载功耗中的一项或多项确定，假设将所述待迁移计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量。
结合第一方面，在一种可能的实现方式中，所述方法还包括：根据假设将所述待迁移计算资源迁移至所述每个机柜上的备选服务器后，所述每个机柜的功率、所述每个机柜与所述每个机柜所在分区的温度调节设备的冷量传导关系信息确定，所述每个机柜的归一化功率；根据所述每个机柜的归一化功率确定所述每个机柜所对应的功率均衡度。
结合第一方面，在一种可能的实现方式中，所述备选服务器的CPU核数不小于，所述待迁移计算资源的CPU核数与所述备选服务器中已存在的计算资源的CPU核数之和；和/或，所述备选服务器的内存规格不小于，所述待迁移计算资源的内存规格与所述备选服务器中已存在的计算资源的内存规格之和。
结合第一方面，在一种可能的实现方式中，所述基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定目标服务器，包括：基于所述计算资源状态信息，为所述待迁移计算资源确定所述机房中的每个机柜上的备选服务器；其中，第一功率变化量不大于第二功率变化量，所述第一功率变化量为假设将所述待迁移计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量，所述第二功率变化量为假设将所述待迁移计算资源迁移至同个机柜上除所述备选服务器外的其他服务器后，所述其他服务器的功率变化量；基于所述计算资源状态信息和所述冷量传导关系信息，确定所述每个机柜所对应的功率均衡度，所述功率均衡度为假设将所述待迁移计算资源迁移到所述每个机柜上的备选服务器后，所述机房的功率均衡度；在所述多个机柜中选择相应功率均衡度最小的机柜上的备选服务器作为所述目标服务器。通过这种方式，可以选取出机房中的每个机柜上的备选服务器，该多个备选服务器为相应功率变化量最小的服务器，那么假设将待迁移计算资源迁移到这些服务器中，服务器的计算功率将减少。进一步的，再从多个机柜中选择相应功率均衡度最小的机柜上的备选服务器作为所述目标服务器，又可以使得假设迁移后的温度调节设备的温控功率减小，进而可以使得迁移后的机房的总功率小于迁移前的机房的总功率。
结合第一方面，在一种可能的实现方式中，所述方法还包括：计算假设执行所述计算资源调整指令集后，所述多个机柜中每个机柜的功率；根据所述多个机柜中每个机柜的功率确定所述至少一个温度调节设备中每个温度调节设备的温度设定参数，所述温度设定参数用于满足执行所述计算资源调整指令集之后，所述多个机柜的温度需求；触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。通过这种方式，可以根据计算资源调整指令集，确定温度调节设备的温度调节指令集。
结合第一方面，在一种可能的实现方式中，所述方法还包括：计算假设执行所述计算资源调整指令集后，所述多个机柜中每个机柜的功率；根据所述每个温度调节设备所在的分区所包含的机柜的功率确定每个温度调节设备的温度设定参数；触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。
结合第一方面，在一种可能的实现方式中，所述计算资源管理指令集为计算资源分配指令集，所述计算资源分配指令集用于在所述机房中分配至少一个计算资源。
结合第一方面，在一种可能的实现方式中，所述根据所述冷量传导关系信息和所述计算资源状态信息，触发所述设施管理设备执行计算资源管理指令集，包括：接收业务需求，所述业务需求用于请求运行所述机房中的计算资源；基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息确定，满足所述业务需求的目标计算资源及提供所述目标计算资源的目标服务器，所述目标服务器提供所述目标计算资源前后所述机房的总功率的增量最小化；生成用于将在所述目标服务器分配所述目标计算资源的指令集作为所述计算资源分配指令集。
结合第一方面，在一种可能的实现方式中，所述基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息确定满足所述业务需求的目标计算资源及提供所述目标计算资源的目标服务器，包括：基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息为所述目标计算资源确定多个候选服务器及每一个候选服务器对应的候选增量，所述候选增量为假设将所述目标计算资源迁移到所述每一个候选服务器后所述机房的总功率的增量；在所述多个候选服务器集合中选择相应候选增量最低的候选服务器作为所述目标服务器。通过选择相应候选增量最低的候选服务器的方式，可以使得执行计算资源分配指令集前后，该机房的总功率的增量最小化。
结合第一方面，在一种可能的实现方式中，所述基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息为所述目标计算资源确定多个候选服务器及每一个候选服务器对应的候选增量，包括：根据所述冷量传导关系信息确定分区冷量吸收关系信息和冷量共享关系信息；根据所述分区冷量吸收关系信息和所述冷量共享关系信息确定待选分区组合集合；根据所述业务需求、所述冷量传导关系信息、所述待选分区组合集合和所述计算资源状态信息，为所述目标计算资源确定多个候选服务器及每一个候选服务器对应的候选增量，所述多个候选服务器属于所述待选分区组合集合；其中，所述分区冷量吸收关系信息用于指示所述机房中各个分区中机柜对本分区中的温度调节设备的冷量吸收能力，一个分区包括一个温度调节设备以及至少一个机柜；所述冷量共享关系信息用于指示所述机房中每两个分区之间，一个分区的机柜对另一个分区的温度调节设备的冷量共享能力；所述待选分区组合集合包括具有不同分区数量的多个待选分区组合，其中，一个待选分区组合为具有相同分区数量的一个或多个分区组合中冷量吸收能力和冷量共享能力最强的分区组合。
结合第一方面，在一种可能的实现方式中，所述根据所述业务需求、所述冷量传导关系信息、所述待选分区组合集合和所述计算资源状态信息，为所述目标计算资源确定多个候选服务器，包括：基于所述业务需求、计算资源状态信息，为所述目标计算资源确定所述待选分区组合集合中每个待选分区组合中的每个机柜上的备选服务器；其中，所述备选服务器具备提供所述目标计算资源的能力，且第三功率变化量不大于第四功率变化量，所述第三功率变化量为假设将所述目标计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量，第四功率变化量为假设将所述目标计算资源迁移至同个机柜上除所述备选服务器外的其他服务器的功率变化量；基于所述计算资源状态信息和所述冷量传导关系信息，确定所述每个机柜所对应的功率均衡度，所述功率均衡度为假设将所述目标计算资源迁移到所述每个机柜上的备选服务器后，所述每个机柜所属的待选分区组合的功率均衡度；将所述每个待选分区组合中，相应功率均衡度最小的机柜上的备选服务器，组成所述多个候选服务器。其中，功率均衡度越小，表示该待选分区组合内的各个机柜之间的功率越均衡，这样会使得温度调节设 备的温控功率较小。又因为备选服务器对应的功率变化量也较小，那么目标计算资源分配到这些服务器中，服务器的计算功率的增量较小，从而能使得执行计算资源分配指令集前后，该机房的总功率的增量最小化。
结合第一方面，在一种可能的实现方式中，所述方法还包括：根据所述目标计算资源的CPU核数、所述目标计算资源的CPU利用率、所述备选服务器的CPU核数，所述备选服务器的空载功耗或者满载功耗中的一项或多项确定，假设将所述目标计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量。
结合第一方面，在一种可能的实现方式中，所述方法还包括：根据假设将所述目标计算资源迁移至所述每个机柜上的备选服务器后，所述每个机柜的功率、所述每个机柜与所述每个机柜所在分区的温度调节设备的冷量传导关系信息确定，所述每个机柜的归一化功率；根据所述每个机柜的归一化功率确定所述每个机柜所对应的功率均衡度。
结合第一方面，在一种可能的实现方式中，所述备选服务器的CPU核数不小于，所述目标计算资源的CPU核数与所述备选服务器中已存在的计算资源的CPU核数之和；和/或，所述备选服务器的内存规格不小于，所述目标计算资源的内存规格与所述备选服务器中已存在的计算资源的内存规格之和。
结合第一方面，在一种可能的实现方式中，所述根据所述冷量传导关系信息和所述计算资源状态信息，触发所述设施管理设备执行计算资源管理指令集，包括：基于所述计算资源状态信息，为所述目标计算资源确定所述机房中的每个机柜上的备选服务器；其中，所述备选服务器具备提供所述目标计算资源的能力，且第三功率变化量不大于第四功率变化量，所述第三功率变化量为假设将所述目标计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量，所述第四功率变化量为假设将所述目标计算资源迁移至同个机柜上除所述备选服务器外的其他服务器后，所述其他服务器的功率变化量；基于所述计算资源状态信息和所述冷量传导关系信息，确定所述每个机柜所对应的功率均衡度，所述功率均衡度为假设将所述目标计算资源迁移到所述每个机柜上的备选服务器后，所述机房的功率均衡度；在所述多个机柜中选择相应功率均衡度最小的机柜上的备选服务器作为所述目标服务器。通过这种方式，可以选取出机房中的每个机柜上的备选服务器，该多个备选服务器为相应功率变化量最小的服务器，那么假设将目标计算资源分配到这些服务器中，服务器的计算功率的增量将较小。进一步的，再从多个机柜中选择相应功率均衡度最小的机柜上的备选服务器作为所述目标服务器，又可以使得假设迁移后的温度调节设备的温控功率的增量较小，进而可以使得迁移后的机房的总功率的增量较小。结合第一方面，在一种可能的实现方式中，所述方法还包括：获取假设执行所述计算资源分配指令集后，所述多个机柜中每个机柜的功率；根据所述多个机柜中每个机柜的功率确定所述至少一个温度调节设备中每个温度调节设备的温度设定参数，所述温度设定参数用于满足执行所述计算资源分配指令集之后，所述多个机柜的温度需求；触发所述设施管理设备执行根据所述温度设定参数生成的温度调节指令集。
结合第一方面，在一种可能的实现方式中，所述方法还包括：计算假设执行所述计算资源分配指令集后，所述多个机柜中每个机柜的功率；根据所述每个温度调节设备所在的分区所包含的机柜的功率确定每个温度调节设备的温度设定参数；触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。
结合第一方面，在一种可能的实现方式中，所述方法还包括：计算假设执行所述计算资源分配指令集后，所述多个机柜中每个机柜的功率；根据所述每个温度调节设备所在的分区所包含的机柜的功率确定每个温度调节设备的温度设定参数；触发所述设施管理设备执行温 度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。
结合第一方面，在一种可能的实现方式中，所述机房的总功率，包括所述多个机柜的计算功率和所述至少一个温度调节设备的总功率。
结合第一方面，在一种可能的实现方式中，第一机柜包含于所述多个机柜，第一温度调节设备包含于所述多个温度调节设备，所述方法还包括：基于第一温度、第二温度、第一进风温度和第二进风温度确定第一机柜与第一温度调节设备的冷量传导关系信息；其中，所述第一进风温度为基于第一设定模型模拟所述机房运行而得出的所述第一机柜的进风温度，所述第一设定模型基于所述机房构建，所述第一设定模型包括：第一功率的所述多个机柜、所述第一温度的所述至少一个温度调节设备；所述第二进风温度为基于第二设定模型模拟所述机房运行而得出的所述第一机柜的进风温度，所述第二设定模型基于所述机房构建，所述第二设定模型包括：所述第一功率的所述多个机柜、所述第二温度的所述第一温度调节设备，以及所述第一温度的所述至少一个温度调节设备中除所述第一温度调节设备的其他温度调节设备。
结合第一方面，在一种可能的实现方式中，所述机房包括第一分区，所述第一分区包括第二温度调节设备和至少一个第二机柜；所述方法还包括：基于所述至少一个第二机柜中的每一个机柜对所述第二温度调节设备的冷量吸收能力确定，所述第一分区的分区冷量吸收关系信息。
结合第一方面，在一种可能的实现方式中，所述机房包括第一分区和第二分区，所述第一分区包括第二温度调节设备和至少一个第二机柜，所述第二分区包括第三温度调节设备和至少一个第三机柜；所述方法还包括：基于所述至少一个第二机柜中的每一个机柜对所述第三温度调节设备的冷量吸收能力，以及所述至少一个第三机柜中的每一个机柜对所述第二温度调节设备的冷量吸收能力确定，所述第一分区与所述第二分区的冷量共享关系信息。
结合第一方面，在一种可能的实现方式中，所述机房包括第一分区，所述第一分区包括第二温度调节设备和至少一个第二机柜；其中，每一个第二机柜对所述第二温度调节设备的冷量吸收能力，强于所述每一个第二机柜对所述多个温度调节设备中除所述第二温度调节设备外的其他温度调节设备的冷量吸收能力。
第二方面，本申请实施例提供了一种节能设备，所述节能设备用于节省机房的能耗，所述机房包括多个机柜和至少一个温度调节设备，其中每一个机柜包括至少一个服务器，每一个服务器用于提供计算资源，每一个温度调节设备用于调节所述机房的温度，所述机房还配置有设施管理设备，所述设施管理设备用于管理所述机房的计算资源和温度调节设备，所述节能设备包括获取单元和处理单元：所述获取单元，用于获取所述机房的冷量传导关系信息和所述机房的计算资源状态信息；所述处理单元，用于根据所述冷量传导关系信息和所述计算资源状态信息，触发设施管理设备执行计算资源管理指令集，以在满足业务需求的同时使所述机房的总功率最小化；其中，所述冷量传导关系信息用于确定所述机房中的每一个机柜对所述机房中的每一个温度调节设备的冷量吸收能力，所述计算资源状态信息用于确定所述机房包含的计算资源的运行位置和运行状态。
结合第二方面，在一种可能的实现方式中，所述计算资源管理指令集为计算资源调整指令集，所述计算资源调整指令集用于调整所述机房中至少一个计算资源的运行位置。
结合第二方面，在一种可能的实现方式中，所述处理单元具体用于：基于所述计算资源状态信息在所述机房中确定待迁移计算资源；基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定目标服务器，所述待迁移计算资源被迁移至所述目标服务 器后所述机房的总功率小于迁移前所述机房的总功率；生成用于将所述待迁移计算资源迁移到所述目标服务器的指令集作为所述计算资源调整指令集。
结合第二方面，在一种可能的实现方式中，所述处理单元具体用于：基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定多个候选服务器及每一个候选服务器对应的候选总功率，所述候选总功率为假设将所述待迁移计算资源迁移到所述每一个候选服务器后所述机房的总功率；在所述多个候选服务器中选择相应候选总功率最低的候选服务器作为所述目标服务器。
结合第二方面，在一种可能的实现方式中，所述处理单元还用于：计算假设执行所述计算资源调整指令集后，所述多个机柜中每个机柜的功率；根据所述多个机柜中每个机柜的功率确定所述至少一个温度调节设备中每个温度调节设备的温度设定参数，所述温度设定参数用于满足执行所述计算资源调整指令集之后，所述多个机柜的温度需求；触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。
结合第二方面，在一种可能的实现方式中，所述计算资源管理指令集为计算资源分配指令集，所述计算资源分配指令集用于在所述机房中分配至少一个计算资源。
结合第二方面，在一种可能的实现方式中，所述处理单元具体用于：接收业务需求，所述业务需求用于请求运行所述机房中的计算资源；基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息确定，满足所述业务需求的目标计算资源及提供所述目标计算资源的目标服务器，所述目标服务器提供所述目标计算资源前后所述机房的总功率的增量最小化；生成用于将在所述目标服务器分配所述目标计算资源的指令集作为所述计算资源分配指令集。
结合第二方面，在一种可能的实现方式中，所述处理单元具体用于：基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息为所述目标计算资源确定多个候选服务器及每一个候选服务器对应的候选增量，所述候选增量为假设将所述目标计算资源迁移到所述每一个候选服务器后所述机房的总功率的增量；在所述多个候选服务器集合中选择相应候选增量最低的候选服务器作为所述目标服务器。
结合第二方面，在一种可能的实现方式中，所述处理单元还用于：获取假设执行所述计算资源分配指令集后，所述多个机柜中每个机柜的功率；根据所述多个机柜中每个机柜的功率确定所述至少一个温度调节设备中每个温度调节设备的温度设定参数，所述温度设定参数用于满足执行所述计算资源分配指令集之后，所述多个机柜的温度需求；触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。
第三方面，本申请提供了又一种节能设备，所述节能设备用于节省机房的能耗，所述机房包括多个机柜和至少一个温度调节设备，其中每一个机柜包括至少一个服务器，每一个服务器用于提供计算资源，每一个温度调节设备用于调节所述机房的温度，所述机房还配置有设施管理设备，所述设施管理设备用于管理所述机房的计算资源和温度调节设备，所述节能设备包括处理器和存储器，所述存储器用于存储程序指令，所述处理器被配置用于调用所述程序指令来执行如下操作：获取所述机房的冷量传导关系信息和所述机房的计算资源状态信息；根据所述冷量传导关系信息和所述计算资源状态信息，触发设施管理设备执行计算资源管理指令集，以在满足业务需求的同时使所述机房的总功率最小化；其中，所述冷量传导关系信息用于确定所述机房中的每一个机柜对所述机房中的每一个温度调节设备的冷量吸收能力，所述计算资源状态信息用于确定所述机房包含的计算资源的运行位置和运行状态。
结合第三方面，在一种可能的实现方式中，所述计算资源管理指令集为计算资源调整指 令集，所述计算资源调整指令集用于调整所述机房中至少一个计算资源的运行位置。
结合第三方面，在一种可能的实现方式中，所述处理器被配置用于调用所述程序指令具体执行如下操作：基于所述计算资源状态信息在所述机房中确定待迁移计算资源；基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定目标服务器，所述待迁移计算资源被迁移至所述目标服务器后所述机房的总功率小于迁移前所述机房的总功率；生成用于将所述待迁移计算资源迁移到所述目标服务器的指令集作为所述计算资源调整指令集。
结合第三方面，在一种可能的实现方式中，所述处理器被配置用于调用所述程序指令具体执行如下操作：基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定多个候选服务器及每一个候选服务器对应的候选总功率，所述候选总功率为假设将所述待迁移计算资源迁移到所述每一个候选服务器后所述机房的总功率；在所述多个候选服务器中选择相应候选总功率最低的候选服务器作为所述目标服务器。
结合第三方面，在一种可能的实现方式中，所述处理器被配置用于调用所述程序指令还执行如下操作：获取假设执行所述计算资源调整指令集后，所述多个机柜中每个机柜的功率；根据所述多个机柜中每个机柜的功率确定所述至少一个温度调节设备中每个温度调节设备的温度设定参数，所述温度设定参数用于满足执行所述计算资源调整指令集之后，所述多个机柜的温度需求；触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。
结合第三方面，在一种可能的实现方式中，所述计算资源管理指令集为计算资源分配指令集，所述计算资源分配指令集用于在所述机房中分配至少一个计算资源。
结合第三方面，在一种可能的实现方式中，所述处理器被配置用于调用所述程序指令具体执行如下操作：接收业务需求，所述业务需求用于请求运行所述机房中的计算资源；基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息确定，满足所述业务需求的目标计算资源及提供所述目标计算资源的目标服务器，所述目标服务器提供所述目标计算资源前后所述机房的总功率的增量最小化；生成用于将在所述目标服务器分配所述目标计算资源的指令集作为所述计算资源分配指令集。
结合第三方面，在一种可能的实现方式中，所述处理器被配置用于调用所述程序指令具体执行如下操作：基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息为所述目标计算资源确定多个候选服务器及每一个候选服务器对应的候选增量，所述候选增量为假设将所述目标计算资源迁移到所述每一个候选服务器后所述机房的总功率的增量；在所述多个候选服务器集合中选择相应候选增量最低的候选服务器作为所述目标服务器。
结合第三方面，在一种可能的实现方式中，所述处理器被配置用于调用所述程序指令还执行如下操作：计算假设执行所述计算资源分配指令集后，所述多个机柜中每个机柜的功率；根据所述多个机柜中每个机柜的功率确定所述至少一个温度调节设备中每个温度调节设备的温度设定参数，所述温度设定参数用于满足执行所述计算资源分配指令集之后，所述多个机柜的温度需求；触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。
第四方面，本申请提供了一种计算资源供应系统，所述计算资源供应系统包括该系统包括如上述第一方面至第三方面中任意一项所描述的节能设备和机房。
第五方面，本申请提供了一种计算机可读存储介质，所述计算机可读存储介质用于存储指令，当所述指令被执行时，使得如第一方面至第三方面中任意一项所述的方法被实现。
第六方面，本申请提供一种计算机程序或计算机程序产品，包括代码或指令，当代码或指令在计算机上运行时，使得计算机执行如第一方面至第三方面中任意一项所述的方法被实现。
附图说明
为了更清楚地说明本申请实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍。
图1是本申请实施例提供的一种机房的架构示意图；
图2是本申请实施例提供的一种计算资源供应系统的架构示意图；
图3是本申请实施例提供的又一种计算资源供应系统的架构示意图；
图4是本申请实施例提供的又一种计算资源供应系统的架构示意图；
图5是本申请实施例提供的一种应用于机房的节能方法的流程示意图；
图6是本申请实施例提供的一种节能设备的结构示意图；
图7是申请实施例提供的另一种节能设备的结构示意图。
具体实施方式
下面结合附图对本申请实施例中的技术方案进行更详细地描述。
本申请的说明书、权利要求书及附图中的术语“第一”和“第二”等是用于区别不同对象，而不是用于描述特定顺序。此外，术语“包括”和“具有”以及它们任何变形，意图在于覆盖不排他的包含。例如包含了一系列步骤或单元的过程、方法、系统、产品或设备没有限定于已列出的步骤或单元，而是可选地还包括没有列出的步骤或单元，或可选地还包括对于这些过程、方法、产品或设备固有的其它步骤或单元。
在本文中提及“实施例”意味着，结合实施例描述的特定特征、结构或特性可以包含在本申请的至少一个实施例中。在说明书中的各个位置出现该短语并不一定均是指相同的实施例，也不是与其它实施例互斥的独立的或备选的实施例。本领域技术人员显式地和隐式地理解的是，本文所描述的实施例可以与其它实施例相结合。
在本申请中，“至少一个(项)”是指一个或者多个，“多个”是指两个或两个以上，“至少两个(项)”是指两个或三个及三个以上，“和/或”，用于描述关联对象的对应关系，表示可以存在三种关系，例如，“A和/或B”可以表示：只存在A，只存在B以及同时存在A和B三种情况，其中A，B可以是单数或者复数。字符“/”一般表示前后关联对象是一种“或”的关系。“以下至少一项(个)”或其类似表达，是指这些项中的任意组合，包括单项(个)或复数项(个)的任意组合。例如，a，b或c中的至少一项(个)，可以表示：a，b，c，“a和b”，“a和c”，“b和c”，或“a和b和c”，其中a，b，c可以是单个，也可以是多个。
本申请实施例所提供的节能方法应用于机房，为了更好地理解本方案，首先对机房以及其相关的概念进行介绍。
数据中心(或者称为资源中心、计算资源中心、计算中心)，是一套集中对外提供信息技术(Internet technology，IT)服务的设施。它可以包括计算机软硬件系统和其它与之配套的设备(例如通信和存储系统)，还包含数据通信连接、环境控制设备、监控设备、各种安全装置等。从空间角度，一个数据中心可以划分为多个机房(或称为机库)；从逻辑角度，可划 分为3层，机房配套层、基础设施层、软件层。机房配套层包含制冷设备、供电设备、消防系统等，为基础设施层提供电力、冷源等。基础设施层，包含计算设备、存储设备、虚拟化资源等，用来承载软件层的业务应用。软件层包含业务应用、云平台等。
以下对上述3层对于数据中心总功率的影响进行分析。
数据中心软件层(例如，OpenStack)，可以实时接收来自数据中心外部的任务(如虚拟机创建任务)，并将其放入任务队列中等待调度，通过任务调度算法将任务分配到其中一台服务器上执行，执行任务会占用服务器的CPU资源，进而增加IT负载，从而增加IT能耗和制冷能耗。因此如何对接收的新任务，或者自身数据中心产生的新任务进行分配，会影响数据中心总功率。
数据中心基础设施层，通常会使用虚拟化技术将一台物理服务器虚拟为多台虚拟机，动态分配给用户使用。调整服务器和虚拟机的归属关系，即将虚拟机从一台物理服务器迁移到另一台物理服务器(也叫做虚拟机迁移)，可以实现IT负载的重新分配，这会影响IT能耗和制冷能耗。因此如何对虚拟机进行迁移调整，会影响数据中心总功率。数据中心机房配套层，制冷系统负责带走机房IT系统所产生的热量，按照能耗和热量的关系，IT系统所产生的热量约等于IT能耗。因此在制冷系统在满足制冷量需求的前提下，如何根据IT能耗动态地调节制冷设备的设定温度、出风量等，会影响数据中心的总功率。
机房，是在空间上构成数据中心的基本单元。机房中包括多个机柜和至少一个温度调节设备。其中，每一个机柜包括至少一个服务器，每一个服务器用于提供计算资源，以满足外部输入或者机房内部产生(例如，监控任务、安全性保护任务，等等)的业务需求；每一个温度调节设备用于调节机房的温度，以使得服务器和机房中的其他配套装置能够工作在适宜的温度/湿度范围内。其中，机房的温度，可以理解为机房的室内温度、空气温度，等等。具体的，在本申请实施例中，每一个温度调节设备用于调节机柜的进风温度，以使得机柜的进风温度不大于机柜进风温度上限(例如，28℃、27℃，等等)。该机柜进风温度上限可以由机柜中放置的服务器或者其他配套装置的正常工作温度要求来确定。本申请实施例中，以机房为应用场景，介绍应用于机房的节能方法。
参见图1，是本申请实施例提供的一种机房的架构示意图。在图1所示的机房中，包括16个机柜(1个机柜上可以放置至少一个服务器)，即L1-L8、R1-R8；4个温度调节设备(例如，空调、温度控制机、冷冻机、风冷设备、水冷设备，等等)，即A1-A4。
对于机房的能耗而言，主要由服务器所产生的IT能耗和温度调节设备所产生的制冷能耗(在一些环境条件下，也可能存在制热能耗)构成，因此从这两个方面来考虑如何节省机房的能耗。由于服务器系统将实时接收来自外部的任务(如虚拟机创建任务)，通过任务调度算法将任务分配到其中一台服务器上执行，执行任务会占用服务器的中央处理器(central processing unit，CPU)资源，进而增加IT负载，从而增加IT能耗和制冷能耗。因此如何制定合适的任务调度算法，对于节省机房总能耗非常关键。另外，空调系统负责处理机房IT系统所产生的热量，按照能耗和热量的关系，IT系统所产生的热量约等于IT能耗。因此空调系统在满足制冷量需求的前提下，如何根据IT能耗动态的调节设定温度、出风量，对于节省机房制冷能耗非常关键。
现阶段，一般通过构建预测模型来确定出任务调度指令和温度设定参数。具体的，通过采集较长一段时间内服务器的资源利用率、机房空调系统参数、机房总能耗、外部环境参数等数据，据此进行模型训练得到总能耗预测模型、空调参数预测模型。基于预测模型依次预测每个任务调度到任意一台服务器上后的总能耗和空调参数，从而可以选取总能耗最小的服 务器作为执行每个任务的目标服务器，对应的空调参数为空调的最终设定参数。但这种方式需要采集大量、多种情况下的实验数据对模型进行训练，对样本数据的依赖性较大；且样本数据较难满足实际机房运行中复杂多变的需求。
鉴于上述介绍的现有情况中存在的问题，本申请提供了一种应用于机房的节能方法，该方法可以在满足业务需求的情况下，使所述机房的总功率最小化。
参见图2，是本申请实施例提供的一种计算资源供应系统的架构示意图。该计算资源供应系统可以包括仿真设备、节能设备、设施管理设备和机房。
为了便于理解，首先对本申请实施例所涉及的相关名词进行解释。
1、冷量传导关系信息
冷量传导关系信息用于确定所述机房中的每一个机柜对所述机房中的每一个温度调节设备的冷量吸收能力。此处的冷量具体是指从温度调节设备传导到机柜，或者机柜吸收温度调节设备的冷量。可选的，冷量传导关系信息也可以称为热量传导关系信息。该热量传导关系信息用于确定所述机房中的每一个温度调节设备对所述机房中的每一个机柜的热量吸收能力。此处的热量具体是指从机柜传导到温度调节设备，或者温度调节设备吸收机柜的热量。需要说明的是，若温度调节设备的功能为制热，则冷量具体是指从机柜传导到温度调节设备，或者温度调节设备吸收机柜的冷量。热量具体是指从温度调节设备传导到机柜，或者机柜吸收温度调节设备的热量。
2、计算资源
本申请实施例中的计算资源可以包括机房中的服务器、虚拟机、任务等。可以理解的是，服务器是被放置在机柜中的硬件设备；虚拟机和任务是被分配到服务器中的软件单元。需要说明的是，本申请实施例中计算资源管理指令集所针对的至少一个计算资源可以为虚拟机或者任务。示例性的，该计算资源管理指令集可以调整虚拟机或者任务的运行位置；或者为业务需求在服务器中分配(或者称为创建)虚拟机或者任务。
3、计算资源状态信息
计算机资源状态信息用于确定所述机房包含的计算资源的运行位置和运行状态。需要说明的是，在一个实际的机房创建完成之后，服务器相对于机柜的运行位置(或者称为位置信息、坐标信息)可以确定。示例性的，为了记录服务器的运行位置，可以对每个机柜、每个服务器进行编号(或者称为命名、标记)，一个编号与一个位置信息具有对应关系。例如，参见图1，机柜L1即对应机房中左侧由上至下的第一个机柜。或者基于机房创建坐标系，对每个机柜、每个服务器的坐标进行记录，一个坐标与一个位置信息具有对应关系。人为或者机械性地移动服务器的位置，会改变服务器的位置信息；一般而言，服务器的运行位置较为固定。而对于虚拟机和任务而言，是基于业务需求动态分配给用户使用的，可以通过计算资源管理指令分配和移动它们的运行位置。
示例性的，服务器的运行状态可以用服务器的中央处理器(central processing unit，CPU)核数、内存规格、空载功率、满载功率，等等运行数据来描述。虚拟机的运行状态可以用虚拟机的CPU核数、内存规格、CPU利用率、内存利用率，等等运行数据来描述。任务的运行状态可以用任务运行所需的CPU需求(即需要几个CPU核)和内存需求，等等数据来描述。
4、计算资源管理指令集
计算资源指令集包括至少一条针对计算资源的调整或者分配指令。在一种可能的实现方式中，计算资源管理指令集可以为计算资源调整指令集，所述计算资源调整指令集用于调整所述机房中至少一个计算资源的运行位置。在一种可能的实现方式中，计算资源管理指令集 可以为计算资源分配指令集，所述计算资源分配指令集用于在所述机房中分配至少一个计算资源。
接下来，对本申请实施例应用的系统架构中的组成部分一一进行介绍。
仿真设备，是一种应用数据中心仿真技术对机房的气流场、温度场进行模拟的电子设备。示例性的，例如通过数据中心的计算流体力学软件求解空气流动的控制方程，其中主要利用质量守恒定律、动量变化定律、热力学第一定律、热力学第二定律、傅里叶传热定律等，对机房的气流场、温度场进行模拟。
仿真设备可以根据实际机房中每个构成对象(电源组件、温度调节设备、机柜、地板、立柱、管道，等等)的位置信息，利用计算流体力学软件构建该机房对应的机房计算流体力学模型(也可以称为虚拟机房)。需要说明的是，虚拟机房和真实机房的机柜摆放、空调大小及位置、机房高度、走线架布置、横梁位置、立柱位置等等保持一致。在机房计算流体力学模型建立好之后，可以用于模拟不同机柜功耗和/或温度调节设备的设定参数(或称为仿真参数)分布下，机房的温度场、气流场。在本申请实施例中，仿真设备可以根据实际机房中采集到构成对象的位置信息，构建实际机房对应的虚拟机房；之后，通过该虚拟机房，来获取不同温度调节设备的设定参数分布下，各个机柜的进风温度，以使得节能设备可以计算机房的冷量传导关系信息。可选的，仿真设备还可以接收节能设备发送的仿真参数，根据该仿真参数和该机房计算流体力学模型，模拟机房的运行，输出仿真结果。
设施管理设备，用于管理机房的计算资源和温度调节设备。可选的，数据中心(包括一个或多个机房)可以配置有共用的设施管理设备(示例性的，可以为数据中心基础设施管理(data center infrastructure management，DCIM)设备)。可选的，每个机房均配置有自身对应的设施管理设备，DCIM设备可以对每个机房的设施管理设备进行管理。在本申请实施例中，设施管理设备可以采集温度调节设备、机柜、服务器、虚拟机、待调度任务等的计算资源状态数据，以生成计算资源状态信息；也可以控制所管理的设备，如调节温度调节设备的设定参数、迁移虚拟机、将任务调度到指定服务器上等。另外，设施管理还可以对外提供数据获取接口和设备控制接口，供外部系统来获取所需数据和控制设备。在本申请实施例中，设施管理设备可以获取实际机房中的计算资源状态信息，并向节能设备发送该计算资源状态信息；另外，设施管理设备还可以被节能设备触发执行计算资源管理指令集。
节能设备，可以通过接口和设施管理设备、仿真设备进行交互。示例性的，节能设备可以通过与设施管理设备的接口可以采集机房的计算资源状态信息，并通过该接口向设施管理设备发送计算资源管理指令集。节能设备还可以通过与仿真设备的接口向仿真设备发送仿真参数，或者采集仿真设备模拟机房运行后得到的仿真结果(示例性的，可以为多个温度数据)。
机房，可以参照上述图1对应的内容的介绍，此处不再赘述。
在本申请实施例中，首先，节能设备可以根据从仿真设备中获取仿真设备模拟机房运行后得到的多个温度数据，根据这些温度数据，节能设备计算出机房的冷量传导关系信息。另外，节能设备可以从设施管理设备获取机房的计算资源状态信息。接下来，节能设备根据所述冷量传导关系信息和所述计算资源状态信息，触发设施管理设备执行计算资源管理指令集，以在满足业务需求的同时使所述机房的总功率最小化。在一些可能的实现方式中，节能设备还可以根据假设的执行所述计算资源管理指令集之后，机房中所述多个机柜中每个机柜的功率确定每个温度调节设备的温度设定参数，再触发设施管理设备执行根据所述温度设定参数生成的温度调节指令集。
参见图3，是本申请实施例提供的又一种计算资源供应系统的架构示意图。在这种可选 的系统架构中，该系统包括仿真设备、设施管理设备和机房，上述节能设备集成在设施管理设备中。
参见图4，是本申请实施例提供的又一种计算资源供应系统的架构示意图。在这种可选的系统架构中，该系统包括设施管理设备和机房，上述仿真设备和节能设备均集成在设施管理设备中。
下面对本申请实施例进行详细的描述。
参见图5，是本申请实施例提供的一种应用于机房的节能方法的流程示意图。其中，该方法可以应用于图2(或者图3、图4)所示的计算资源供应系统，下面描述的节能设备可以是图2中所示的节能设备；下面描述的机房可以是图1或者图2(或者图3、图4)中所示的机房。可选的，在另一些可能的系统架构中(例如，图3、图4)，节能设备可以集成在设施管理设备中，执行下述方法的设备也可以是设施管理设备。该方法包括以下步骤：
S101、节能设备获取机房的冷量传导关系信息和所述机房的计算资源状态信息。
其中，冷量传导关系信息和计算资源状态信息的概念，可以参照上述内容中的介绍，此处不再赘述。
可选的，节能设备可以通过设施管理设备获取所述机房的计算资源状态信息。
示例性的，计算资源状态信息中包括的服务器的运行位置信息，可以示例如下：
Figure PCTCN2022082480-appb-000001
示例性的，计算资源状态信息中包括的虚拟机的运行位置信息，可以示例如下：
Figure PCTCN2022082480-appb-000002
Figure PCTCN2022082480-appb-000003
示例性的，计算资源状态信息中还可以包括服务器的运行状态信息：CPU核数、内存规格、空载功率、满载功率，等等；虚拟机的运行状态信息：CPU核数、内存规格、CPU利用率、内存利用率，等等。
可选的，节能设备可以通过仿真设备采集模拟机房运行后得到的多个温度数据，来计算机房的冷量传导关系信息。以第一机柜和第一温度调节设备的冷量传导关系信息为例，对计算机房的冷量传导关系信息的方式进行介绍。其中，第一机柜为机房中的任意一个机柜，第一温度调节设备为机房中的任意一个温度调节设备。
在一些实施例中，节能设备基于第一温度、第二温度、第一进风温度和第二进风温度确定第一机柜与第一温度调节设备的冷量传导关系信息；其中，所述第一进风温度为基于第一设定模型模拟所述机房运行而得出的所述第一机柜的进风温度，所述第一设定模型基于所述机房构建，所述第一设定模型包括：第一功率的所述多个机柜、所述第一温度的所述至少一个温度调节设备；所述第二进风温度为基于第二设定模型模拟所述机房运行而得出的所述第一机柜的进风温度，所述第二设定模型基于所述机房构建，所述第二设定模型包括：所述第一功率的所述多个机柜、所述第二温度的所述第一温度调节设备，以及所述第一温度的所述至少一个温度调节设备中除所述第一温度调节设备的其他温度调节设备。
其中，第一功率、第一温度、第二温度可以由节能设备发送给仿真设备，以使得仿真设备基于这些数据设定第一设定模型和第二设定模型中机柜或者温度调节设备的运行状态。第一设定模型和第二设定模型可以理解为上述内容中介绍的机房计算流体力学模型，第一设定模型和第二设定模型是按照机房中实际构成对象的位置信息进行构建的。
示例性的，节能设备发送给仿真设备第一设定模型的设定参数(即包括第一功率和第一温度)的请求可以为：
Figure PCTCN2022082480-appb-000004
Figure PCTCN2022082480-appb-000005
示例性的，节能设备发送给仿真设备第二设定模型的设定参数(即包括第一功率、第一温度和第二温度)的请求可以为：
Figure PCTCN2022082480-appb-000006
Figure PCTCN2022082480-appb-000007
仿真设备基于第一设定模型模拟所述机房运行而得出的所述第一机柜的第一进风温度，以及基于第二设定模型模拟所述机房运行而得出的所述第一机柜的第二进风温度。
示例性的，针对一个设定模型，仿真设备运行得出的结果如下：
Figure PCTCN2022082480-appb-000008
节能设备可以从仿真设备获取第一进风温度和第二进风温度。需要说明的是，第一设定模型和第二设定模型不同之处仅在于调整了第一温度调节设备的温度，其他温度调节设备的温度以及每个机柜的功率保持不变，可以理解为，限制了其他因素对第一机柜的进风温度的影响；通过这种方式，可以确定出第一机柜与第一温度调节设备的冷量传导关系信息。
在一种可能的实现方式中，本申请实施例用温度相关性指数(temperature correlation index，TCI)来描述一个机柜对一个温度调节设备的冷量吸收能力。示例性的，这个值越大，表示一个机柜对一个温度调节设备的冷量吸收能力越强；换句话说，也可以表示一个机柜和一个温度调节设备的关联性越强。
可选的，第一机柜与第一温度调节设备的TCI的计算方式可以参见公式1-1：
Figure PCTCN2022082480-appb-000009
其中，
Figure PCTCN2022082480-appb-000010
为所述第二进风温度，
Figure PCTCN2022082480-appb-000011
为所述第一进风温度，Y为所述第二温度，X为所述第一温度。第一机柜与第一温度调节设备的冷量传导关系信息可以为第一机柜、第一温度调节 设备与TCI 11的关联关系。按照上述介绍的相似的方式，可以计算出其他机柜和其他温度调节设备的冷量传导关系信息。
参见表1，表1中记录了机房中的部分冷量传导关系信息。
表1
  A1 A2 A3 A4
L1 0.64 0.08 0.33 0.07
L2 0.72 0.08 0.28 0.06
L3 0.75 0.11 0.31 0.08
L4 0.76 0.13 0.29 0.09
R1 0.31 0.07 0.66 0.07
R2 0.24 0.06 0.78 0.08
R3 0.24 0.09 0.78 0.09
R4 0.27 0.1 0.74 0.11
… … … … …
其中，行标识“L1”“L2”“R1”“R2”表示机柜的标识(或者标号、编号)，该标识与机柜的位置信息相对应。例如，“L1”表示图1机房示意图中左边第一个机柜，“R2”表示图1机房示意图中右边第二个机柜)。列标识“A1”“A2”“A3”“A4”表示温度调节设备的标识，该标识与温度调节设备的位置信息相对应。例如，“A1”表示图1机房示意图中左边第一个空调。每个单元的值表示相应机柜和相应温度调节设备的TCI的值，例如，当行标识为L1，列标识为A1时，TCI的取值为0.64，表示机柜L1与温度调节设备A1的TCI的值为0.64；这个值的大小表示了机柜L1对温度调节设备A1的冷量吸收能力的大小。
可选的，在一种可能的实现方式中，可以通过冷量传导关系信息确定机柜和温度调节设备之间的归属关系(或者称为强关联关系)。示例性的，对于机柜L1而言，它与温度调节设备A1的TCI，大于机柜L1与其他温度调节设备的TCI，则可以视为机柜L1与温度调节设备A1之间具有归属关系，也即是，机柜L1为归属于温度调节设备A1的机柜，温度调节设备A1是机柜L1的归属温度调节设备。
S102、节能设备根据所述冷量传导关系信息和所述计算资源状态信息，触发设施管理设备执行计算资源管理指令集，以在满足业务需求的同时使所述机房的总功率最小化。
其中，机房的总功率包括机房中多个机柜的计算功率(也可以称为IT功率，等等)和机房中至少一个温度调节设备的温控功率(也可以称为制冷功率或者制热功率，等等)。需要说明的是，机房的总功率的最小化的含义是，执行该计算资源管理指令集，使得机房的总功率小于其他管理指令集被执行之后该机房的总功率。在一些实施例中，所述计算资源管理指令集为计算资源调整指令集，所述计算资源调整指令集用于调整所述机房中至少一个计算资源的运行位置。可以理解的是，至少一个计算资源可以为虚拟机或者任务。计算资源调整指令集可以包括一条或者多条针对计算资源的调整指令。设施管理设备执行计算资源调整指令集之后，可以调整至少一个计算资源的运行位置，可以实现IT负载的重新分配，这样影响计算功率和温控功率。在本申请实施例中，执行计算资源调整指令集之后所述机房的总功率小于执行前所述机房的总功率。
可选的，节能设备根据所述冷量传导关系信息和所述计算资源状态信息，触发设施管理设备执行计算资源管理指令集的过程可以包括如下步骤：
S11、基于所述计算资源状态信息在所述机房中确定待迁移计算资源。
在一些应用场景中，机房可以周期性地更新虚拟机和/或任务的运行位置。在这种情况下，机房中所包含的全部虚拟机和/或任务，可以作为该待迁移计算资源。
在一些应用场景中，每个虚拟机或者任务均有设定的更新周期，可选的，不同的虚拟机或者任务所设定的更新周期可以相同，也可以不同。在当前时刻到达某些虚拟机或者任务的更新周期的情况下，这些虚拟机或者任务可以作为待迁移计算资源。
在一些应用场景中，对于机房新获取到的业务需求(可能是来自机房外部的业务需求，也可能是机房自身产生的业务需求)，机房可以先为该业务需求分配临时计算资源，以便及时响应该业务需求，在一定时间后再对这些临时计算资源进行调整。在这种情况下，这些临时计算资源可以作为待迁移计算资源。示例性的，这种场景下的业务需求可以是在线业务，这类业务为时延敏感型业务，对业务的响应速度的要求较高。例如，实时面向终端用户的业务为一种在线业务，示例性的，该业务需求可以是电商业务、搜索业务、推荐业务等等。
S12、基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定目标服务器。
具体的，所述待迁移计算资源被迁移至所述目标服务器后所述机房的总功率小于迁移前所述机房的总功率。在一些实施例中，所述基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定目标服务器的方式可以为：基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定多个候选服务器及每一个候选服务器对应的候选总功率，所述候选总功率为假设将所述待迁移计算资源迁移到所述每一个候选服务器后所述机房的总功率；在所述多个候选服务器中选择相应候选总功率最低的候选服务器作为所述目标服务器。
以下介绍确定候选服务器的一些方法。
第一种可能的实现方式中，为了使得待迁移计算资源可以在目标服务器中正常运行，要求所述候选服务器的CPU核数不小于，所述待迁移计算资源的CPU核数与所述候选服务器中已存在的计算资源的CPU核数之和；和/或，所述候选服务器的内存规格不小于，所述待迁移计算资源的内存规格与所述候选服务器中已存在的计算资源的内存规格之和。可选的，还可以根据待迁移计算资源的业务特性，对候选服务器作出进一步的要求，比如说，待迁移计算资源为一种鉴权任务，那么候选服务器需要具备鉴权的功能。需要说明的是，还可能对候选服务器增加一些其他的要求，本申请实施例不作具体限制。在这种实现方式中，机房中满足待迁移计算资源的运行要求的服务器均可以作为候选服务器。
第二种可能的实现方式中，节能设备可以首先在机房的多个服务器中选取出满足待迁移计算资源的运行要求(可参见第一种可能的实现方式的介绍)的服务器，作为备选服务器；之后，基于所述计算资源状态信息，确定每个备选服务器对应的功率变化量，该功率变化量为假设将所述待迁移计算资源迁移至所述每个备选服务器后，所述每个备选服务器的功率变化量。将每个机柜上相应功率变化量最小的备选服务器，组成该多个候选服务器。在这种方式中，多个候选服务器为相应功率变化量最小的服务器，那么待迁移计算资源迁移到这些服务器中，服务器的计算功率减少，从而能使得迁移后的机房的总功率小于迁移前的机房的总功率。
可选的，所述方法还包括：根据所述待迁移计算资源的CPU核数、所述待迁移计算资源的CPU利用率、所述备选服务器的CPU核数，所述备选服务器的空载功耗或者满载功耗中的一项或多项确定，假设将所述待迁移计算资源迁移至所述备选服务器后，所述备选服务器 的功率变化量。
示例性的，介绍虚拟机j放入服务器k的功率变化量ΔP的计算方式。
如果在假设迁移前，服务器k上没有虚拟机，则：
Figure PCTCN2022082480-appb-000012
如果在假设迁移前，服务器k上包含虚拟机，则：
Figure PCTCN2022082480-appb-000013
其中，
Figure PCTCN2022082480-appb-000014
表示虚拟机j的CPU核数，
Figure PCTCN2022082480-appb-000015
表示虚拟机j的CPU利用率，
Figure PCTCN2022082480-appb-000016
表示服务器k的CPU核数，
Figure PCTCN2022082480-appb-000017
为服务器k的空载功耗，
Figure PCTCN2022082480-appb-000018
为服务器k的满载功耗。这些数据可以从计算资源状态信息中获取。
介绍虚拟机创建任务l放入服务器k的功率变化量ΔP的计算方式。
如果在假设迁移前，服务器上本身没有虚拟机和虚拟机创建任务，则：
Figure PCTCN2022082480-appb-000019
如果在假设迁移前，服务器上本身包含虚拟机或虚拟机创建任务，则：
Figure PCTCN2022082480-appb-000020
其中，
Figure PCTCN2022082480-appb-000021
表示任务l的CPU需求，
Figure PCTCN2022082480-appb-000022
表示服务器k的CPU核数，
Figure PCTCN2022082480-appb-000023
为服务器k的空载功耗，
Figure PCTCN2022082480-appb-000024
为服务器k的满载功耗。这些数据可以从计算资源状态信息中获取。
需要说明的是，还可以存在其他计算功率变化量的方式，本申请实施例不再赘述。
第三种可能的实现方式中，节能设备可以首先在机房的多个服务器中选取出满足待迁移计算资源的运行要求(可参见第一种可能的实现方式的介绍)的服务器，作为备选服务器；之后，基于所述冷量传导关系信息和计算资源状态信息，确定所述机房中与每个温度调节设备的TCI最大的多个备选机柜(例如，以表1为示例，与温度调节设备A1的TCI最大的服务器为服务器L4)，将处于多个备选机柜上的备选服务器作为该多个候选服务器。在这种方式中，多个候选服务器为与温度调节设备的TCI最大，表明候选服务器吸收冷量的能力越强，那么待迁移计算资源迁移到这些服务器中，温度调节设备的温控功率将会减小，能使得迁移后的机房的总功率小于迁移前的机房的总功率。
需要说明的是，待迁移计算资源可以包含多个计算资源，可以为每个待迁移计算资源确定一个目标服务器，不同的待迁移计算资源的目标服务器可以相同，也可以不同。每个待迁移计算资源迁移到其对应的目标服务器的计算资源指令，组合成该计算资源调整指令集。
可选的，可以按照CPU利用率由高到低的顺序依次为待迁移计算资源确定目标服务器。由于CPU利用率较高的待迁移计算资源所占用的功率较高，通过这种方式，可以先迁移对机房总功率影响较大的待迁移计算资源，有利于减小机房的总功率。
另外，需要说明的是，在一个待迁移计算资源确定了其对应的目标服务器之后，该目标服务器的运行状态信息则需要更新，因为该目标服务器需要为该待迁移计算资源预留计算资源。也即是说，在选定下一个待迁移计算资源对应的目标服务器的过程中，利用的是基于上一个待选计算资源的选择结果更新后的计算资源状态信息。
示例性的，虚拟机1假设迁移之前在服务器1上，假设迁移之后在服务器2上，假设迁移前和假设迁移后服务器和虚拟机的归属关系数据结构(或者称为运行位置信息)的变化如下：
Figure PCTCN2022082480-appb-000025
以下介绍确定每一个候选服务器对应的候选总功率的方式。该方法可以包括以下步骤：
步骤a1、获取假设执行将待迁移计算资源迁移至一个候选服务器后，所述多个机柜中每个机柜的功率。
可选的，首先计算假设迁移后，每个服务器的CPU利用率S cpu-u：
服务器的CPU利用率等于服务器上所有虚拟机CPU负载之和，再加上服务器上虚拟机创建任务的CPU需求之和，除以服务器的CPU核数，其中虚拟机CPU负载等于虚拟机CPU核数乘以虚拟机CPU利用率。
Figure PCTCN2022082480-appb-000026
表示服务器k的CPU利用率。
之后，计算每个服务器的功率P server：
如果服务器上包含虚拟机，
Figure PCTCN2022082480-appb-000027
其中，
Figure PCTCN2022082480-appb-000028
表示服务器k的功率，
Figure PCTCN2022082480-appb-000029
为服务器k的满载功率，
Figure PCTCN2022082480-appb-000030
为服务器k的空载功率，服务器的功率和服务器的CPU利用率线性相关。
如果服务器上没有虚拟机，
Figure PCTCN2022082480-appb-000031
表示服务器k的功率为0。
接下来，计算每个机柜的功率P rack：机柜功率等于机柜上所有服务器功率之和。
Figure PCTCN2022082480-appb-000032
表示第i个机柜的功率。
步骤b1、根据每个机柜的功率确定所述多个机柜的计算功率。具体的，多个机柜的计算功率为每个机柜的功率之和。
步骤c1、根据所述多个机柜中每个机柜的功率确定所述至少一个温度调节设备中每个温度调节设备的温度设定参数，所述温度设定参数用于满足执行将待迁移计算资源迁移至一个候选服务器之后，所述多个机柜的温度需求。
确定每个温度调节设备的温度设定参数的方式可参考如下：
首先，节能设备将每个机柜的功率(步骤S21计算得到)、温度调节设备的模拟温度值(初始时，每个温度调节设备的运行状态为开启、初始温度均为可设定的最大温度Z℃(例如35℃))作为输入参数发送给仿真设备，以使得仿真设备基于该输入参数模拟机房运行。需要说明的是，若将一个温度调节设备作为归属温度调节设备的所有机柜的功率均为0，则可以设定该一个温度调节设备的运行状态为关闭。
之后，从仿真设备获取每个机柜的进风温度T i、每个空调的进风温度
Figure PCTCN2022082480-appb-000033
每个空调的出风温度
Figure PCTCN2022082480-appb-000034
每个空调的空气流量M j。其中T i表示第i个机柜的进风温度、
Figure PCTCN2022082480-appb-000035
表示第j个空调的进风温度、
Figure PCTCN2022082480-appb-000036
表示第j个空调的出风温度、M j表示第j个空调的空气流量。
示例性的，节能设备发送给仿真设备的输入参数的请求可以为：
Figure PCTCN2022082480-appb-000037
Figure PCTCN2022082480-appb-000038
示例性的，仿真设备基于输入参数模拟所述机房运行而得出的输出结果可以为：
Figure PCTCN2022082480-appb-000039
如果存在机柜的T i>机柜进风温度上限，其中机柜进风温度上限为W℃，例如28℃)，则将与该机柜的TCI最大的温度调节设备的模拟温度值下调预设值(例如，1℃)，并再根据下调后的模拟温度值再次得到每个机柜的进风温度；直到每个机柜T i≤机柜进行温度上限。将满足每个机柜T i≤机柜进行温度上限这个条件下，相应的每个温度调节设备的模拟温度值确定为每个温度调节设备的温度设定参数。
如果每个机柜的T i≤机柜进风温度上限，则将每个温度调节设备的模拟温度值确定为每个温度调节设备的温度设定参数。
步骤d1、基于每个温度调节设备的温度设定参数对应的每个温度调节设备的进风温度、出风温度(可从步骤c1中获取)计算每个温度调节设备的功率。
示例性的，一个温度调节设备的功率为：
Figure PCTCN2022082480-appb-000040
其中，
Figure PCTCN2022082480-appb-000041
其中C为空气的比热容，取值1005，制冷系数(coefficient of performance，COP；表示单位功耗所能获得的冷量)的计算公式及计算公式中的0.0068、0.0008、0.458是大量实验所得出的系数值。
Figure PCTCN2022082480-appb-000042
表示第j个空调的功率。
步骤e1、根据每个温度调节设备的功率确定所述至少一个温度调节设备的温控功率。具体的，至少一个温度调节设备的温控功率为每个温度调节设备的功率之和。
步骤f1、根据所述多个机柜的计算功率和至少一个温度调节设备的温控功率确定该一个候选服务器对应的候选总功率。
按照相同的方式，可以确定出每一个候选服务器对应的候选总功率。之后，节能设备可以在所述多个候选服务器中选择相应候选总功率最低的候选服务器作为所述目标服务器。
在另一些实施例中，节能设备基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定目标服务器的方式还可以为：基于所述计算资源状态信息，为所述待迁移计算资源确定所述机房中的每个机柜上的备选服务器；其中，第一功率变化量不大于第二功率变化量，所述第一功率变化量为假设将所述待迁移计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量，所述第二功率变化量为假设将所述待迁移计算资源迁移至同个机柜上除所述备选服务器外的其他服务器后，所述其他服务器的功率变化量；基于所述计算资源状态信息和所述冷量传导关系信息，确定所述每个机柜所对应的功率均衡度，所述功率均衡度为假设将所述待迁移计算资源迁移到所述每个机柜上的备选服务器后，所述机房的功率均衡度；在所述多个机柜中选择相应功率均衡度最小的机柜上的备选服务器作为所述目标服务器。
需要说明的是，节能设备可以先从机房的多个服务器中选取出满足待迁移计算资源的运行要求(可参见上述内容中的介绍)的部分服务器，再在这部分服务器中选取备选服务器。计算假设将所述待迁移计算资源迁移至某个服务器后，某个服务器的功率变化量的方式可以参照上述内容中的介绍，此处不再赘述。
可选的，所述方法还包括：根据所述目标计算资源的CPU核数、所述目标计算资源的CPU利用率、所述备选服务器的CPU核数，所述备选服务器的空载功耗或者满载功耗中的一项或多项确定，假设将所述目标计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量。其中，该节能设备可以根据计算资源状态信息，确定所述目标计算资源的CPU核数、所述目标计算资源的CPU利用率、所述备选服务器的CPU核数，所述备选服务器的空载功耗或者满载功耗中的一项或多项。
在一种可能的实现方式中，根据假设将所述目标计算资源迁移至所述每个机柜上的备选服务器后，所述每个机柜的功率、所述每个机柜与所述每个机柜所对应的归属温度调节设备的冷量传导关系信息确定，所述每个机柜的归一化功率；根据所述每个机柜的归一化功率确定所述每个机柜所对应的功率均衡度。
示例性的，介绍待迁移计算资源迁移至服务器k的功率均衡度的计算方式。该服务器k是机柜i上的服务器。
步骤a2、计算机房内每个服务器的CPU利用率S cpu-u：
服务器的CPU利用率等于服务器上所有虚拟机CPU负载之和，再加上服务器上虚拟机创建任务的CPU需求之和，除以服务器的CPU核数，其中虚拟机CPU负载等于虚拟机CPU核数乘以虚拟机CPU利用率。
Figure PCTCN2022082480-appb-000043
表示服务器k的CPU利用率。
步骤b2、计算机房内每个服务器的功率P server：
如果服务器上包含虚拟机，
Figure PCTCN2022082480-appb-000044
其中，
Figure PCTCN2022082480-appb-000045
表示服务器k的功率，
Figure PCTCN2022082480-appb-000046
为服务器k的满载功率，
Figure PCTCN2022082480-appb-000047
为服务器k的空载功率，服务器的功率和服务器的CPU利用率线性相关。
如果服务器上没有虚拟机，
Figure PCTCN2022082480-appb-000048
表示服务器k的功率为0。
步骤c2、计算机房内每个机柜的功率P rack：机柜功率等于机柜上所有服务器功率之和。
Figure PCTCN2022082480-appb-000049
表示第i个机柜的功率。
步骤d2、计算机房内每个机柜的归一化功率～P rack：
示例性的，对于机柜i，根据表1确定该机柜所对应的归属温度调节设备，示例性的，该温度调节设备为温度调节设备j。通过TCI表获取第i个机柜和第j个空调的TCI，记为tci ij，则机柜i的归一化功率
Figure PCTCN2022082480-appb-000050
通过相同的方式，可以计算出机房内每个机柜的归一化功率～P rack。
步骤e2、计算机房内所有机柜的归一化功率
Figure PCTCN2022082480-appb-000051
的标准差，将该标准差(还可以是方差)作为机柜i对应的功率均衡度。
通过上述相似的方式，可以确定出机房中的每个机柜对应的功率均衡度。之后，在所述多个机柜中选择相应功率均衡度最小的机柜上的备选服务器作为所述目标服务器。需要说明的是，功率均衡度越小，表示该机房内的各个机柜之间的功率越均衡，这样会使得温度调节设备的温控功率较小。通过这种方式，可以选取出机房中的每个机柜上的备选服务器，该多个备选服务器为相应功率变化量最小的服务器，那么假设将待迁移计算资源迁移到这些服务器中，服务器的计算功率将减少。进一步的，再从多个机柜中选择相应功率均衡度最小的机柜上的备选服务器作为所述目标服务器，又可以使得假设迁移后的温度调节设备的温控功率减小，进而可以使得迁移后的机房的总功率小于迁移前的机房的总功率。
S13、生成用于将所述待迁移计算资源迁移到所述目标服务器的指令集作为所述计算资源调整指令集。
在一种可能的实现方式中，该方法还包括：计算假设执行所述计算资源调整指令集后，所述多个机柜中每个机柜的功率；根据所述多个机柜中每个机柜的功率确定所述至少一个温度调节设备中每个温度调节设备的温度设定参数，所述温度设定参数用于满足执行所述计算资源调整指令集之后，所述多个机柜的温度需求；触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。该温度调节指令集可以包括一条或者多条温度调节指令，该温度调节指令可以调节温度调节设备的运行状态(开启或者关闭)，也可以调节温度调节设备的设定温度。需要说明的是，如何确定每个温度调节设备的温度设定参数的方式，可以参照上述内容中的步骤a1-步骤c1的介绍。另外，在具体应用过程中， 可以无需再次执行确定的步骤，可以从之前的确定出的温度设定参数中选取出该目标服务器对应的每个温度调节设备的温度设定参数。
示例性的，所述计算资源调整指令集和温度调节指令集可以为：
Figure PCTCN2022082480-appb-000052
在另一些实施例中，所述计算资源管理指令集为计算资源分配指令集，所述计算资源分配指令集用于在所述机房中分配至少一个计算资源。可以理解的是，至少一个计算资源可以为虚拟机或者任务。计算资源分配指令集可以包括一条或者多条针对计算资源的分配指令。在本申请实施例中，执行计算资源分配指令集前后所述机房的总功率的增量最小化。需要说明的是，机房的总功率的增量最小化的含义是，执行该计算资源分配指令集，使得机房的总功率的增量小于其他分配指令集被执行所导致机房的总功率的增量。
可选的，节能设备根据所述冷量传导关系信息和所述计算资源状态信息，触发设施管理设备执行计算资源管理指令集的过程可以包括如下步骤：
S21、接收业务需求，所述业务需求用于请求运行所述机房中的计算资源。
需要说明的是，该业务需求，可能是来自机房外部的业务需求，也可能是机房自身产生的业务需求。
S22、基于所述业务需求、所述计算资源状态信息在所述机房中确定，满足所述业务需求的目标计算资源及提供所述目标计算资源的目标服务器。
其中，所述目标服务器提供所述目标计算资源前后所述机房的总功率的增量最小化。
在一些应用场景中，对于机房新获取到的业务需求，机房可以实时为该业务需求分配目标计算资源，以及确定提供所述目标计算资源的目标服务器。示例性的，这种场景下的业务需求可以是计算业务，这类业务一般对时延不敏感，对业务的响应速度要求不高。示例性的， 该业务需求可以是spark业务、mr(mapreduce)业务、图计算业务、音视频编解码业务的业务需求，等等；该业务需求可以来自机房外部的设备，也可以来自机房内部的设备。
在一些实施例中，所述基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定目标服务器的方式可以为：基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息为所述目标计算资源确定多个候选服务器及每一个候选服务器对应的候选增量，所述候选增量为假设将所述目标计算资源迁移到所述每一个候选服务器后所述机房的总功率的增量；在所述多个候选服务器集合中选择相应候选增量最低的候选服务器作为所述目标服务器。
以下介绍确定候选服务器的一些方法。
第一种可能的实现方式中，为了使得目标计算资源可以在目标服务器中正常运行，要求所述候选服务器的CPU核数不小于，所述目标计算资源的CPU核数与所述候选服务器中已存在的计算资源的CPU核数之和；和/或，所述候选服务器的内存规格不小于，所述目标计算资源的内存规格与所述候选服务器中已存在的计算资源的内存规格之和。另外，还可以根据业务需求，对候选服务器作出进一步的要求，比如说，该业务为一种鉴权任务，那么候选服务器需要具备鉴权的功能。需要说明的是，还可能对候选服务器增加一些其他的要求，本申请实施例不作具体限制。在这种实现方式中，机房中满足目标计算资源的运行要求(或者表达为满足业务需求)的服务器均可以作为候选服务器。
第二种可能的实现方式中，节能设备可以首先在机房的多个服务器中选取出满足目标计算资源的运行要求(可参见第一种可能的实现方式的介绍)的服务器，作为备选服务器；之后，基于所述计算资源状态信息，确定每个备选服务器对应的功率变化量，该功率变化量为假设将所述目标计算资源迁移至所述每个备选服务器后，所述每个备选服务器的功率变化量。将每个机柜上相应功率变化量最小的备选服务器，组成该多个候选服务器。在这种方式中，多个候选服务器为相应功率变化量最小的服务器，那么目标计算资源分配到这些服务器中，服务器的计算功率的增量较小，能使得分配后的机房的总功率的增量较低。
可选的，所述方法还包括：根据所述目标计算资源的CPU核数、所述目标计算资源的CPU利用率、所述备选服务器的CPU核数，所述备选服务器的空载功耗或者满载功耗中的一项或多项确定，假设将所述目标计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量。计算每个备选服务器对应的功率变化量的方式，可以参照上述实施例中的介绍，此处不再赘述。
第三种可能的实现方式中，节能设备可以首先在机房的多个服务器中选取出满足目标计算资源的运行要求(可参见第一种可能的实现方式的介绍)的服务器，作为备选服务器；之后，基于所述冷量传导关系信息和计算资源状态信息，确定所述机房中与每个温度调节设备的TCI最大的多个备选机柜(例如，以表1为示例，与温度调节设备A1的TCI最大的服务器为服务器L4)，将处于多个备选机柜上的备选服务器作为该多个候选服务器。在这种方式中，多个候选服务器为与温度调节设备的TCI最大，表明候选服务器吸收冷量的能力越强，那么目标计算资源分配到这些服务器中，温度调节设备的温控功率的增量较小，能使得迁移后的机房的总功率的增量较低。
需要说明的是，目标计算资源可以包含多个计算资源，可以为每个目标计算资源确定一个目标服务器，不同的目标计算资源的目标服务器可以相同，也可以不同。每个目标计算资源分配到其对应的目标服务器的计算资源指令，组合成该计算资源分配指令集。
可选的，可以按照CPU利用率由高到低的顺序依次为目标计算资源确定目标服务器。由 于CPU利用率较高的目标计算资源所占用的功率较高，通过这种方式，可以先迁移对机房总功率影响较大的目标计算资源，有利于使得机房的总功率的增量较低。
另外，需要说明的是，在一个目标计算资源确定了其对应的目标服务器之后，该目标服务器的运行状态信息则需要更新，因为该目标服务器需要为该目标计算资源预留计算资源。也即是说，在选定下一个目标计算资源对应的目标服务器的过程中，利用的是基于上一个待选计算资源的选择结果更新后的计算资源状态信息。
以下介绍确定每一个候选服务器对应的候选增量的方式。该方法可以包括以下步骤：
步骤a3、获取假设执行将目标计算资源分配至一个候选服务器之前，机房的第一总功率。
步骤b3、获取假设执行将目标计算资源分配至一个候选服务器后，所述多个机柜中每个机柜的功率。
步骤c3、根据每个机柜的功率确定所述多个机柜的计算功率。具体的，多个机柜的计算功率为每个机柜的功率之和。
步骤d3、根据所述多个机柜中每个机柜的功率确定所述至少一个温度调节设备中每个温度调节设备的温度设定参数，所述温度设定参数用于满足执行将目标计算资源分配至一个候选服务器之后，所述多个机柜的温度需求。
步骤e3、基于每个温度调节设备的温度设定参数对应的每个温度调节设备的进风温度、出风温度(可从步骤d3中获取)计算每个温度调节设备的功率。
步骤f3、根据每个温度调节设备的功率确定所述至少一个温度调节设备的温控功率。具体的，至少一个温度调节设备的温控功率为每个温度调节设备的功率之和。
步骤g3、根据所述多个机柜的计算功率和至少一个温度调节设备的温控功率确定该一个候选服务器对应的机房的第二总功率，该一个候选服务器对应的候选增量为第二总功率与第一总功率的差值。
需要说明的是，步骤b3-步骤g3的执行方式，可以参照上述内容中介绍的步骤a1-步骤f1的执行方式。按照相同的方式，可以确定出每一个候选服务器对应的候选增量。之后，在所述多个候选服务器集合中选择相应候选增量最低的候选服务器作为所述目标服务器。
在另一些实施例中，所述基于所述冷量传导关系信息和所述计算资源状态信息为所述目标计算资源确定目标服务器的方式还可以为：基于所述计算资源状态信息，为所述目标计算资源确定所述机房中的每个机柜上的备选服务器；其中，第三功率变化量不大于第四功率变化量，所述第三功率变化量为假设将所述目标计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量，所述第四功率变化量为假设将所述目标计算资源迁移至同个机柜上除所述备选服务器外的其他服务器后，所述其他服务器的功率变化量；基于所述计算资源状态信息和所述冷量传导关系信息，确定所述每个机柜所对应的功率均衡度，所述功率均衡度为假设将所述目标计算资源迁移到所述每个机柜上的备选服务器后，所述机房的功率均衡度；在所述多个机柜中选择相应功率均衡度最小的机柜上的备选服务器作为所述目标服务器。
需要说明的是，节能设备可以先从机房的多个服务器中选取出满足目标计算资源的运行要求(可参见第一种可能的实现方式的介绍)的部分服务器，再在这部分服务器中选取备选服务器。计算假设将所述目标计算资源迁移至某个服务器后，某个服务器的功率变化量的方式可以参照上述内容中的介绍，此处不再赘述。
可选的，所述方法还包括：根据所述目标计算资源的CPU核数、所述目标计算资源的CPU利用率、所述备选服务器的CPU核数，所述备选服务器的空载功耗或者满载功耗中的一项或多项确定，假设将所述目标计算资源迁移至所述备选服务器后，所述备选服务器的功率 变化量。其中，该节能设备可以根据计算资源状态信息，确定所述目标计算资源的CPU核数、所述目标计算资源的CPU利用率、所述备选服务器的CPU核数，所述备选服务器的空载功耗或者满载功耗中的一项或多项。
在一种可能的实现方式中，根据假设将所述目标计算资源迁移至所述每个机柜上的备选服务器后，所述每个机柜的功率、所述每个机柜与所述每个机柜所对应的归属温度调节设备的冷量传导关系信息确定，所述每个机柜的归一化功率；根据所述每个机柜的归一化功率确定所述每个机柜所对应的功率均衡度。
需要说明的是，目标计算资源迁移至服务器k的功率均衡度的计算方式，可以参照上述步骤a2-步骤e2的介绍，此处不再赘述。
通过上述相似的方式，可以确定出机房中的每个机柜对应的功率均衡度。之后，在所述多个机柜中选择相应功率均衡度最小的机柜上的备选服务器作为所述目标服务器。需要说明的是，功率均衡度越小，表示该机房内的各个机柜之间的功率越均衡，这样会使得温度调节设备的温控功率较小。通过这种方式，可以选取出机房中的每个机柜上的备选服务器，该多个备选服务器为相应功率变化量最小的服务器，那么假设将目标计算资源分配到这些服务器中，服务器的计算功率的增量将较小。进一步的，再从多个机柜中选择相应功率均衡度最小的机柜上的备选服务器作为所述目标服务器，又可以使得假设迁移后的温度调节设备的温控功率的增量较小，进而可以使得迁移后的机房的总功率的增量较小。S13、生成用于将在所述目标服务器分配所述目标计算资源的指令集作为所述计算资源分配指令集。
在一种可能的实现方式中，该方法还包括：计算假设执行所述计算资源分配指令集后，所述多个机柜中每个机柜的功率；根据所述多个机柜中每个机柜的功率确定所述至少一个温度调节设备中每个温度调节设备的温度设定参数，所述温度设定参数用于满足执行所述计算资源分配指令集之后，所述多个机柜的温度需求；触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。需要说明的是，如何确定每个温度调节设备的温度设定参数的方式，可以参照上述内容中的步骤a2-步骤c2的介绍。另外，在具体应用过程中，可以无需再次执行确定的步骤，可以从之前的确定出的温度设定参数中选取出该目标服务器对应的每个温度调节设备的温度设定参数。
在本申请另一些实施例中，考虑到机房中各个位置处的机柜对温度调节设备的冷量的吸收能力的差异较大，节能设备可以对机房中包括的机柜和温度调节设备进行分区。一个分区中包括一个温度调节设备和至少一个机柜。之后，从这些分区中确定出冷量吸收能力和冷量共享能力更强的待选分区组合。将待迁移计算资源迁移至该待选分区组合中的服务器上，或者将目标计算资源分配至该待选分区组合中的服务器上；由于待选分区组合具有较强的冷量吸收能力和冷量共享能力，通过这种方式，可以降低温度调节设备的温控功率，从而使得机房的总功率最小化。
在一种可能的方式中，可以基于机房的冷量传导关系信息对机房中包括的机柜和温度调节设备进行分区。示例性的，所述机房包括第一分区，所述第一分区包括第二温度调节设备和至少一个第二机柜；其中，每一个第二机柜对所述第二温度调节设备的冷量吸收能力，强于所述每一个第二机柜对所述多个温度调节设备中除所述第二温度调节设备外的其他温度调节设备的冷量吸收能力。可以理解为，一个分区中包括一个温度调节设备，以及对这个温度调节设备的冷量的吸收能力最强的机柜。以上述内容中介绍的机柜和温度调节设备的归属关系来理解，一个分区中包括一个温度调节设备，以及以这个温度调节设备作为归属温度调节设备的机柜。
以上述表1所示的示例为例，机房的分区方式可以参见表2。
表2
分区 温度调节设备 机柜
P1 A1 L1、L2、L3、L4
P2 A2 L5、L6、L7、L8
P3 A3 R1、R2、R3、R4
P4 A4 R5、R6、R7、R8
其中，“分区”列下每个单元的值为分区的编号，“温度调节设备”列下每个单元的值为对应分区内的温度调节设备的编号，“机柜”列下每个单元的值表示对应分区内的一组机柜的编号。
可选的，可以根据冷量传导关系信息确定分区的分区冷量吸收关系信息，以及两个分区之间的冷量共享关系信息。其中，所述分区冷量吸收关系信息用于指示所述机房中各个分区中机柜对本分区中的温度调节设备的冷量吸收能力，一个分区包括一个温度调节设备以及至少一个机柜。所述冷量共享关系信息用于指示所述机房中每两个分区之间，一个分区的机柜对另一个分区的温度调节设备的冷量共享能力。
示例性的，机房包括第一分区，所述第一分区包括第二温度调节设备和至少一个第二机柜。本申请实施例的方法还包括：基于所述至少一个第二机柜中的每一个机柜对所述第二温度调节设备的冷量吸收能力确定，所述第一分区的分区冷量吸收关系信息。具体的，第一分区的分区冷量吸收关系信息为第一分区中的机柜、第一分区中的温度调节设备与分区温度指数(partition temperature index，PTI)的关联关系。本申请实施例用PTI来描述一个分区内的机柜对本分区内的温度调节设备的冷量吸收能力。示例性的，这个值越大，表示一个分区内的机柜对本分区内的温度调节设备的冷量吸收能力越强。第一分区的PTI的值为第一分区中的每一个机柜与第一分区中的温度调节设备的TCI的值的和。
以表1和表2为例，机房的分区冷量吸收关系信息可以表示为表3。
表3
分区 温度调节设备 机柜 PTI
P1 A1 L1、L2、L3、L4 2.87
P2 A2 L5、L6、L7、L8 2.89
P3 A3 R1、R2、R3、R4 2.96
P4 A4 R5、R6、R7、R8 2.94
其中，“PTI”列下每个单元的值表示对应分区的PTI值。以分区P1为例，分区P1的PTI为温度调节设备A1和机柜L1、温度调节设备A1和机柜L2、温度调节设备A1和机柜L3、温度调节设备A1和机柜L4的TCI之和，即，0.64+0.72+0.75+0.76＝2.87。
示例性的，机房还包括第二分区，该第二分区包括第三温度调节设备和至少一个第三机柜；本申请实施例的方法还包括：基于所述至少一个第二机柜中的每一个机柜对所述第三温度调节设备的冷量吸收能力，以及所述至少一个第三机柜中的每一个机柜对所述第二温度调节设备的冷量吸收能力确定，所述第一分区与所述第二分区的冷量共享关系信息。具体的，第一分区与第二分区的冷量共享关系信息为第一分区、第二分区与分区相关性指数(partition correlation index，PCI)的关联关系。本申请实施例用PCI来描述每两个分区之间，一个分区的机柜对另一个分区的温度调节设备的冷量共享能力。示例性的，这个值越大，表示一个分 区内的机柜对另一个内的温度调节设备的冷量吸收能力越强。第一分区与第二分区的PTI的值为第一分区中的每一个机柜与第二分区中的温度调节设备的TCI的值的和、与第二分区中的每一个机柜与第二分区中的温度调节设备的TCI的值的和，的和的一半。
示例性的，以表1和表2为例，分区P1和P3之间的PCI＝((0.33+0.28+0.31+0.29)+(0.31+0.24+0.24+0.27))/2＝1.13，其中0.33、0.28、0.31、0.29分别为分区P1内的机柜L1、L2、L3、L4和分区P3的温度调节设备A3的TCI，0.31、0.24、0.24、0.27分别为分区P3内的机柜R1、R2、R3、R4和分区P1的温度调节设备A1的TCI。按照这种方式，可以计算机房内每两个分区之间的PCI。
以表1和表2为例，机房的冷量共享关系信息可以表示为表4。
表4
分区1 分区2 PCI
P1 P2 0.4
P1 P3 1.13
P1 P4 0.29
P2 P3 0.27
P2 P4 1.21
P3 P4 0.35
其中，“分区1”列下每个单元的值为共享冷量的两个分区中第一个分区的编号，“分区2”列下每个单元的值为共享冷量的两个分区中第二个分区的编号，“PCI”列下每个单元的值为第一个分区和第二个分区之间的PCI。例如，上述分区间PCI表中第一行表示分区P1与分区P2共享冷量的能力值为0.4，即分区P1内所有机柜吸收分区P2内温度调节设备提供的冷量的能力值与分区P2内所有机柜吸收分区P1内温度调节设备提供的冷量的能力值之和的一半为0.4。
在一种可能的实现方式中，可以根据所述分区冷量吸收关系信息和所述冷量共享关系信息确定待选分区组合集合。其中，所述待选分区组合集合包括具有不同分区数量的多个待选分区组合，其中，一个待选分区组合为具有相同分区数量的一个或多个分区组合中冷量吸收能力和冷量共享能力最强的分区组合。需要说明的是，一个分区组合中可以包括一个或者多个分区，示例性的，当一个分区组合中包括一个分区时，分区组合可以为P1、P2、P3、P4，当一个分区组合中包括两个分区时，分区组合可以为：P1和P2的分区组合、P1和P3的分区组合、P1和P4的分区组合、P2和P3的分区组合、P2和P4的分区组合、P3和P4的分区组合。分区组合中包括三个分区时，分区组合可以为：P1、P2和P3的分区组合，P1、P3和P4的分区组合，P2、P3和P4的分区组合。分区组合中包括四个分区时，分区组合为：P1、P2、P3和P4的分区组合。
可选的，计算具有相同分区数量的一个或多个分区组合中每个分区组合对应的PTI与PCI之和。将对应PTI与PCI之和最大的分区组合作为待选分区组合。需要说明的是，在分区组合中只有一个分区的情况下，由于分区组合不存在冷量共享能力，可以比较分区组合的PTI。
以上述机房为例，在分区数量为2的情况下，P2和P4的分区组合的PTI和PCI之和为2.89+2.94+1.21＝7.04，大于P1和P2的分区组合的6.16、P1和P4的分区组合的6.1、P2和P3的分区组合的6.12、P1和P3的分区组合的6.96、P3和P4的分区组合的6.25。因此，在分区数量为2的情况下，待选分区组合为P1和P2的分区组合。
以表2、表3和表4为例，机房的待选分区组合集合可以表示为表5。
表5
分区数量 待选分区组合 PTI和PCI之和
1 P3 2.96
2 P2、P4 7.04
3 P2、P3、P4 10.62
4 P1、P2、P3、P4 15.31
其中，“分区数量”列下每个单元的值为待选分区组合中包含的分区数量，“待选分区组合”列下每个单元的值为对应分区数量下的待选分区组合，“PTI和PCI之和”列下每个单元的值为对应待选分区组合的分区内PTI和分区间PCI之和。例如，可以从表5中查看到，上在分区数量为3的情况下，待选分区组合是P2、P3和P4的分区组合。
基于这部分内容的论述，结合前述方案，可以提出本方案中又一些确定候选服务器的方法，以下对这些方式进行介绍。
在一些实施例中，节能设备可以根据所述冷量传导关系信息、所述待选分区组合集合和所述计算资源状态信息，为所述待迁移计算资源确定多个候选服务器及每一个候选服务器对应的候选总功率，所述多个候选服务器属于所述待选分区组合集合。
可选的，节能设备可以根据所述冷量传导关系信息、所述待选分区组合集合和所述计算资源状态信息，为所述待迁移计算资源确定多个候选服务器的方式可以包含以下步骤：
S31、基于所述计算资源状态信息，为所述待迁移计算资源确定所述待选分区组合集合中每个待选分区组合中的每个机柜上的备选服务器。
其中，第一功率变化量不大于第二功率变化量，所述第一功率变化量为假设将所述待迁移计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量，所述第二功率变化量假设将所述待迁移计算资源迁移至同个机柜上除所述备选服务器外的其他服务器后，所述其他服务器的功率变化量。需要说明的是，可以先从每个待选分区组合中的服务器中筛选出满足待迁移计算资源的运行要求(可参见上述内容中的介绍)的部分服务器，再在这部分服务器中选取备选服务器。
另外，计算假设将所述待迁移计算资源迁移至某个服务器后，某个服务器的功率变化量的方式可以参照上述内容中的介绍，此处不再赘述。通过这种方式，可以选取出每个待选分区组合中的每个机柜上的备选服务器，多个备选服务器为相应功率变化量最小的服务器，那么假设将待迁移计算资源迁移到这些服务器中，服务器的计算功率将减少，从而可以使得迁移后的机房的总功率小于迁移前的机房的总功率。
S32、基于所述计算资源状态信息和所述冷量传导关系信息，确定所述每个机柜所对应的功率均衡度，所述功率均衡度为假设将所述待迁移计算资源迁移到所述每个机柜上的备选服务器后，所述每个机柜所属的待选分区组合的功率均衡度。
可选的，所述方法还包括：根据所述目标计算资源的CPU核数、所述目标计算资源的CPU利用率、所述备选服务器的CPU核数，所述备选服务器的空载功耗或者满载功耗中的一项或多项确定，假设将所述目标计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量。其中，该节能设备可以根据计算资源状态信息，确定所述目标计算资源的CPU核数、所述目标计算资源的CPU利用率、所述备选服务器的CPU核数，所述备选服务器的空载功耗或者满载功耗中的一项或多项。
在一种可能的实现方式中，根据假设将所述目标计算资源迁移至所述每个机柜上的备选服务器后，所述每个机柜的功率、所述每个机柜与所述每个机柜所在分区的温度调节设备的 冷量传导关系信息(也可以表达为，所述每个机柜与所述每个机柜所对应的归属温度调节设备的冷量传导关系信息)确定，所述每个机柜的归一化功率；根据所述每个机柜的归一化功率确定所述每个机柜所对应的功率均衡度。
示例性的，介绍待迁移计算资源迁移至服务器k的功率均衡度的计算方式。该服务器k是机柜i上的服务器，这个机柜所属的待选分区组合为待选分区组合m。
步骤a4、计算待选分区组合m内每个服务器的CPU利用率S cpu-u：
服务器的CPU利用率等于服务器上所有虚拟机CPU负载之和，再加上服务器上虚拟机创建任务的CPU需求之和，除以服务器的CPU核数，其中虚拟机CPU负载等于虚拟机CPU核数乘以虚拟机CPU利用率。
Figure PCTCN2022082480-appb-000053
表示服务器k的CPU利用率。
步骤b4、计算待选分区组合m内每个服务器的功率P server：
如果服务器上包含虚拟机，
Figure PCTCN2022082480-appb-000054
其中，
Figure PCTCN2022082480-appb-000055
表示服务器k的功率，
Figure PCTCN2022082480-appb-000056
为服务器k的满载功率，
Figure PCTCN2022082480-appb-000057
为服务器k的空载功率，服务器的功率和服务器的CPU利用率线性相关。
如果服务器上没有虚拟机，
Figure PCTCN2022082480-appb-000058
表示服务器k的功率为0。
步骤c4、计算待选分区组合m内每个机柜的功率P rack：机柜功率等于机柜上所有服务器功率之和。
Figure PCTCN2022082480-appb-000059
表示第i个机柜的功率。
步骤d4、计算待选分区组合m内每个机柜的归一化功率～P rack：
示例性的，对于机柜i，通过分区内PTI表获取到该机柜所在分区的温度调节设备，该温度调节设备编号记为j，通过TCI表获取第i个机柜和第j个空调的TCI，记为tci ij，则机柜i的归一化功率
Figure PCTCN2022082480-appb-000060
通过相同的方式，可以计算出待选分区组合m内每个机柜的归一化功率～P rack。
步骤e4、计算待选分区组合m内所有机柜的归一化功率
Figure PCTCN2022082480-appb-000061
的标准差，将该标准差(还可以是方差)作为机柜i对应的功率均衡度。
通过上述相似的方式，可以确定出每个待选分区组合中的每个机柜对应的功率均衡度。
S33、将所述每个待选分区组合中，相应功率均衡度最小的机柜上的备选服务器，组成所述多个候选服务器。
示例性的，以表5为例，可以在4个待选分区组合的每个分区组合中，确定出一个相应功率均衡度最小的机柜上的备选服务器，这4个服务器组成该多个候选服务器。需要说明的是，功率均衡度越小，表示该待选分区组合内的各个机柜之间的功率越均衡，这样会使得温度调节设备的温控功率较小。又因为备选服务器对应的功率变化量也较小，那么待迁移计算资源迁移到这些服务器中，服务器的计算功率减少，从而能使得迁移后的机房的总功率小于迁移前的机房的总功率。
需要说明的是，按照上述介绍的方式确定出多个候选服务器之后，节能设备可以接着确定每一个候选服务器对应的候选总功率，在所述多个候选服务器中选择相应候选总功率最低的候选服务器作为所述目标服务器。
需要说明的是，待迁移计算资源可以包含多个计算资源，可以为每个待迁移计算资源确 定一个目标服务器，不同的待迁移计算资源的目标服务器可以相同，也可以不同。每个待迁移计算资源迁移到其对应的目标服务器的计算资源指令，组合成该计算资源调整指令集。
在一种可能的实现方式中，所述方法还包括：计算假设执行所述计算资源调整指令集后，所述多个机柜中每个机柜的功率；根据所述每个温度调节设备所在的分区所包含的机柜的功率确定每个温度调节设备的温度设定参数；触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。其中，一个温度调节设备的温度设定参数由这个温度调节设备所在的分区所包含的机柜的功率来确定，也可以表达为，一个温度调节设备需要满足自身所在分区的机柜的温度需求。结合上述内容中介绍的确定每个温度调节设备的温度设定参数的方法(步骤c1)中的内容，在这种实现方式中，如果存在机柜的T i>机柜进风温度上限(其中机柜进风温度上限为W℃，例如28℃)的情况，可以将该机柜所在分区的温度调节设备的模拟温度值下调预设值(例如，1℃)，并再根据下调后的模拟温度值再次得到每个机柜的进风温度；直到每个机柜T i≤机柜进行温度上限。将满足每个机柜T i≤机柜进行温度上限这个条件下，相应的每个温度调节设备的模拟温度值确定为每个温度调节设备的温度设定参数。
示例性的，节能设备的输出的基于不同待选分区组合确定的待选计算资源调整指令集(示例包括虚拟机迁移指令和任务调度指令)，以及依据这些待选计算资源调整指令集确定出的计算功率、温控功率和每个温度调节设备的温度设定参数的计算结果的示例如下：
Figure PCTCN2022082480-appb-000062
Figure PCTCN2022082480-appb-000063
参见上述输出结果，可以选取相应候选总功率最低的待选计算资源调整指令集作为该计算机资源调整指令集；并且根据相应候选总功率最低的待选计算资源调整指令集所对应的，每个温度调节设备的温度设定参数生成温度调节指令集。
在一些实施例中，该节能设备可以根据所述业务需求、所述冷量传导关系信息、所述待选分区组合集合和所述计算资源状态信息，为所述目标计算资源确定多个候选服务器及每一个候选服务器对应的候选增量，所述多个候选服务器属于所述待选分区组合集合。
可选的，节能设备根据所述业务需求、所述冷量传导关系信息、所述待选分区组合集合和所述计算资源状态信息，为所述目标计算资源确定多个候选服务器可以包含以下步骤：
S41、基于所述业务需求、计算资源状态信息，为所述目标计算资源确定所述待选分区组合集合中每个待选分区组合中的每个机柜上的备选服务器。
先从每个待选分区组合中的服务器中筛选出具备提供所述目标计算资源的能力的部分服务器，再在这部分服务器中选取备选服务器。
其中，第三功率变化量不大于第四功率变化量，所述第三功率变化量为假设将所述目标计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量，第四功率变化量为假设将所述目标计算资源迁移至同个机柜上除所述备选服务器外的其他服务器后，所述其他服务器的功率变化量。
另外，计算假设将所述目标计算资源迁移至某个服务器后，某个服务器的功率变化量的方式可以参照上述内容中的介绍，此处不再赘述。通过这种方式，可以选取出每个待选分区组合中的每个机柜上的备选服务器，多个备选服务器为相应功率变化量最小的服务器，那么目标计算资源迁移到这些服务器中，服务器的计算功率减少，从而能使得迁移后的机房的总功率小于迁移前的机房的总功率。
S42、基于所述计算资源状态信息和所述冷量传导关系信息，确定所述每个机柜所对应的功率均衡度，所述功率均衡度为假设将所述目标计算资源迁移到所述每个机柜上的备选服务器后，所述每个机柜所属的待选分区组合的功率均衡度。
可选的，所述方法还包括：根据所述目标计算资源的CPU核数、所述目标计算资源的CPU利用率、所述备选服务器的CPU核数，所述备选服务器的空载功耗或者满载功耗中的一 项或多项确定，假设将所述目标计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量。
在一种可能的实现方式中，根据假设将所述目标计算资源迁移至所述每个机柜上的备选服务器后，所述每个机柜的功率、所述每个机柜与所述每个机柜所在分区的温度调节设备的冷量传导关系信息(也可以表达为，所述每个机柜与所述每个机柜所对应的归属温度调节设备的冷量传导关系信息)确定，所述每个机柜的归一化功率；根据所述每个机柜的归一化功率确定所述每个机柜所对应的功率均衡度。
需要说明的是，计算每个机柜所对应的功率均衡度的方式可以参照上述步骤a4-步骤e4内容中的介绍。
S43、将所述每个待选分区组合中，相应功率均衡度最小的机柜上的备选服务器，组成所述多个候选服务器。
示例性的，以表5为例，可以在4个待选分区组合的每个分区组合中，确定出一个相应功率均衡度最小的机柜上的备选服务器，这4个服务器组成该多个候选服务器。需要说明的是，功率均衡度越小，表示该待选分区组合内的各个机柜之间的功率越均衡，这样会使得温度调节设备的温控功率较小。又因为备选服务器对应的功率变化量也较小，那么目标计算资源分配到这些服务器中，服务器的计算功率的增量较少，从而能使得分配前后的机房的总功率的增量小于其他计算资源分配指令集执行前后机房的总功率的增量。
需要说明的是，按照上述介绍的方式确定出多个候选服务器之后，节能设备可以接着确定每一个候选服务器对应的候选总功率的增量，在所述多个候选服务器中选择相应候选总功率的增量最低的候选服务器作为所述目标服务器。
需要说明的是，目标计算资源可以包含多个计算资源，可以为每个目标计算资源确定一个目标服务器，不同的目标计算资源的目标服务器可以相同，也可以不同。每个目标计算资源分配到其对应的目标服务器的计算资源指令，组合成该计算资源分配指令集。
在一种可能的实现方式中，所述方法还包括：计算假设执行所述计算资源分配指令集后，所述多个机柜中每个机柜的功率；根据所述每个温度调节设备所在的分区所包含的机柜的功率确定每个温度调节设备的温度设定参数；触发所述设施管理设备执行温度调节指令集，所述温度调节指令是根据所述温度设定参数生成的。其中，一个温度调节设备的温度设定参数由这一个温度调节设备所在的分区所包含的机柜的功率来确定，也可以表达为，一个温度调节设备需要满足自身所在分区的机柜的温度需求。结合上述内容中介绍的确定每个温度调节设备的温度设定参数的方法(步骤c1)中的内容，在这种实现方式中，如果存在机柜的T i>机柜进风温度上限(其中机柜进风温度上限为W℃，例如28℃)的情况，可以将该机柜所在分区的温度调节设备的模拟温度值下调预设值(例如，1℃)，并再根据下调后的模拟温度值再次得到每个机柜的进风温度；直到每个机柜T i≤机柜进行温度上限。将满足每个机柜T i≤机柜进行温度上限这个条件下，相应的每个温度调节设备的模拟温度值确定为每个温度调节设备的温度设定参数。
为了实现上述本申请实施例提供的方法中的各功能，节能设备可以包括硬件结构、软件模块，以硬件结构、软件模块、或硬件结构加软件模块的形式来实现上述各功能。上述各功能中的某个功能可以以硬件结构、软件模块、或者硬件结构加软件模块的方式来执行。
参见图6，是本申请实施例提供的一种节能设备的结构示意图。该节能设备用于节省机房的能耗，该机房包括多个机柜和至少一个温度调节设备，其中每一个机柜包括至少一个服务器，每一个服务器用于提供计算资源，每一个温度调节设备用于调节该机房的温度。该机 房还配置有设施管理设备，所述设施管理设备用于管理该机房的计算资源和温度调节设备，该节能设备60包括获取单元601和处理单元602，其中：
获取单元601，用于获取所述机房的冷量传导关系信息和所述机房的计算资源状态信息。具体的，该获取单元601所执行的操作可以参照上述图5所示方法中的步骤S101中的介绍。
处理单元602，用于根据所述冷量传导关系信息和所述计算资源状态信息，触发所述设施管理设备执行计算资源管理指令集，以在满足业务需求的同时使所述机房的总功率最小化。具体的，该处理单元602所执行的操作可以参照上述图5所示方法中的步骤S102中的介绍。
其中，所述冷量传导关系信息用于确定所述机房中的每一个机柜对所述机房中的每一个温度调节设备的冷量吸收能力，所述计算资源状态信息用于确定所述机房包含的计算资源的运行位置和运行状态。
在一些实施例中，所述计算资源管理指令集为计算资源调整指令集，所述计算资源调整指令集用于调整所述机房中至少一个计算资源的运行位置。
在一些实施例中，所述处理单元具体用于：基于所述计算资源状态信息在所述机房中确定待迁移计算资源；基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定目标服务器，所述待迁移计算资源被迁移至所述目标服务器后所述机房的总功率小于迁移前所述机房的总功率；生成用于将所述待迁移计算资源迁移到所述目标服务器的指令集作为所述计算资源调整指令集。
在一些实施例中，所述处理单元具体用于：基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定多个候选服务器及每一个候选服务器对应的候选总功率，所述候选总功率为假设将所述待迁移计算资源迁移到所述每一个候选服务器后所述机房的总功率；在所述多个候选服务器中选择相应候选总功率最低的候选服务器作为所述目标服务器。
在一些实施例中，所述处理单元还用于：获取假设执行所述计算资源调整指令集后，所述多个机柜中每个机柜的功率；根据所述多个机柜中每个机柜的功率确定所述至少一个温度调节设备中每个温度调节设备的温度设定参数，所述温度设定参数用于满足执行所述计算资源调整指令集之后，所述多个机柜的温度需求；触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。
在一些实施例中，所述计算资源管理指令集为计算资源分配指令集，所述计算资源分配指令集用于在所述机房中分配至少一个计算资源。
在一些实施例中，所述处理单元具体用于：接收业务需求，所述业务需求用于请求运行所述机房中的计算资源；基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息确定，满足所述业务需求的目标计算资源及提供所述目标计算资源的目标服务器，所述目标服务器提供所述目标计算资源前后所述机房的总功率的增量最小化；生成用于将在所述目标服务器分配所述目标计算资源的指令集作为所述计算资源分配指令集。
在一些实施例中，所述处理单元具体用于：基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息为所述目标计算资源确定多个候选服务器及每一个候选服务器对应的候选增量，所述候选增量为假设将所述目标计算资源迁移到所述每一个候选服务器后所述机房的总功率的增量；在所述多个候选服务器集合中选择相应候选增量最低的候选服务器作为所述目标服务器。
在一些实施例中，所述处理单元还用于：获取假设执行所述计算资源分配指令集后，所述多个机柜中每个机柜的功率；根据所述多个机柜中每个机柜的功率确定所述至少一个温度调节设备中每个温度调节设备的温度设定参数，所述温度设定参数用于满足执行所述计算资 源分配指令集之后，所述多个机柜的温度需求；触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。
需要说明的是，图6所示的节能设备的各个单元执行的操作可以上述方法实施例的相关内容。此处不再详述。上述各个单元可以以硬件，软件或者软硬件结合的方式来实现。各个实施方式所能达到的技术效果也可以参照上述内容中的描述。在一个实施例中，上述内容中的获取单元601以及处理单元602的功能可以由节能设备60中的一个或多个处理器来实现。
通过图6所示的节能设备，可以根据获取到的冷量传导关系信息和计算资源状态信息，触发设施管理设备执行计算资源管理指令集，以在满足业务需求的同时使所述机房的总功率最小化。
参见图7，是本申请实施例提供的又一种节能设备的结构示意图。该节能设备70用于节省机房的能耗，该机房包括多个机柜和至少一个温度调节设备，其中每一个机柜包括至少一个服务器，每一个服务器用于提供计算资源，每一个温度调节设备用于调节所述机房的温度。所述机房还配置有设施管理设备，所述设施管理设备用于管理所述机房的计算资源和温度调节设备，该节能设备70可以包括：一个或多个处理器701；一个或多个输入设备702，一个或多个输出设备703和存储器704。上述处理器701、输入设备702、输出设备703和存储器704通过总线705连接。存储器704用于存储指令。
所称处理器701可以是中央处理单元，该处理器还可以是其他通用处理器、数字信号处理器、专用集成电路或者其他可编程逻辑器件等。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。
输入设备702可以包括通信接口，数据线等，输出设备703可以包括显示器(LCD等)，数据线，通信接口等。
该存储器704可以包括只读存储器和随机存取存储器，并向处理器701提供指令和数据。存储器704的一部分还可以包括非易失性随机存取存储器。例如，存储器704还可以存储设备类型的信息。
处理器701用于运行存储器704存储的指令来执行如下操作：
获取所述机房的冷量传导关系信息和所述机房的计算资源状态信息；根据所述冷量传导关系信息和所述计算资源状态信息，触发所述设施管理设备执行计算资源管理指令集，以在满足业务需求的同时使所述机房的总功率最小化。
其中，所述冷量传导关系信息用于确定所述机房中的每一个机柜对所述机房中的每一个温度调节设备的冷量吸收能力，所述计算资源状态信息用于确定所述机房包含的计算资源的运行位置和运行状态。
处理器701执行的操作可以参照上述方法实施例的相关内容。此处不再详述。
通过图7所示的节能设备，可以根据获取到的冷量传导关系信息和计算资源状态信息，触发设施管理设备执行计算资源管理指令集，以在满足业务需求的同时使所述机房的总功率最小化。
本领域技术人员还可以了解到本申请实施例列出的各种说明性逻辑块(illustrative logical block)和步骤(step)可以通过电子硬件、电脑软件，或两者的结合进行实现。这样的功能是通过硬件还是软件来实现取决于特定的应用和整个系统的设计要求。本领域技术人员可以对于每种特定的应用，可以使用各种方法实现所述的功能，但这种实现不应被理解为超出本申请实施例保护的范围。
本申请还提供了一种计算机可读存储介质，其上存储有计算机程序，该计算机可读存储 介质被计算机执行时实现上述任一方法实施例的功能。
本申请还提供了一种计算机程序产品，该计算机程序产品被计算机执行时实现上述任一方法实施例的功能。
本申请实施例还提供一种计算资源供应系统，该计算资源供应系统可以包括图5所对应的实施例中的节能设备和机房。该节能设备和机房可以参照上述任一方法实施例的介绍。
在上述实施例中，可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件实现时，可以全部或部分地以计算机程序产品的形式实现。所述计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行所述计算机指令时，全部或部分地产生按照本申请实施例所述的流程或功能。所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。所述计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一个计算机可读存储介质传输，例如，所述计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、数字用户线(digital subscriber line，DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。所述计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包含一个或多个可用介质集成的服务器、数据中心等数据存储设备。所述可用介质可以是磁性介质(例如，软盘、硬盘、磁带)、光介质(例如，高密度数字视频光盘(digital video disc，DVD))、或者半导体介质(例如，固态硬盘(solid state disk，SSD))等。
本领域普通技术人员可以理解：本申请中涉及的第一、第二等各种数字编号仅为描述方便进行的区分，并不用来限制本申请实施例的范围，先后顺序。
本申请中各表所示的对应关系可以被配置，也可以是预定义的。各表中的信息的取值仅仅是举例，可以配置为其他值，本申请并不限定。在配置信息与各参数的对应关系时，并不一定要求必须配置各表中示意出的所有对应关系。例如，本申请中的表格中，某些行示出的对应关系也可以不配置。又例如，可以基于上述表格做适当的变形调整，例如，拆分，合并等等。上述各表中标题示出参数的名称也可以采用通信装置可理解的其他名称，其参数的取值或表示方式也可以通信装置可理解的其他取值或表示方式。上述各表在实现时，也可以采用其他的数据结构，例如可以采用数组、队列、容器、栈、线性表、指针、链表、树、图、结构体、类、堆、散列表或哈希表等。
本申请中的预定义可以理解为定义、预先定义、存储、预存储、预协商、预配置、固化、或预烧制。
本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统、装置和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应以所述权利要求的保护范围为准。

Claims (34)

  1. 一种应用于机房的节能方法，所述机房包括多个机柜和至少一个温度调节设备，其中每一个机柜包括至少一个服务器，每一个服务器用于提供计算资源，每一个温度调节设备用于调节所述机房的温度，所述机房还配置有设施管理设备，所述设施管理设备用于管理所述机房的计算资源和温度调节设备，其特征在于，所述方法包括：
    获取所述机房的冷量传导关系信息和所述机房的计算资源状态信息；
    根据所述冷量传导关系信息和所述计算资源状态信息，触发所述设施管理设备执行计算资源管理指令集，以在满足业务需求的同时使所述机房的总功率最小化；
    其中，所述冷量传导关系信息用于确定所述机房中的每一个机柜对所述机房中的每一个温度调节设备的冷量吸收能力，所述计算资源状态信息用于确定所述机房包含的计算资源的运行位置和运行状态。
  2. 根据权利要求1所述的方法，其特征在于，所述计算资源管理指令集为计算资源调整指令集，所述计算资源调整指令集用于调整所述机房中至少一个计算资源的运行位置。
  3. 根据权利要求2所述的方法，其特征在于，所述根据所述冷量传导关系信息和所述计算资源状态信息，触发所述设施管理设备执行计算资源管理指令集，包括：
    基于所述计算资源状态信息在所述机房中确定待迁移计算资源；
    基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定目标服务器，所述待迁移计算资源被迁移至所述目标服务器后所述机房的总功率小于迁移前所述机房的总功率；
    生成用于将所述待迁移计算资源迁移到所述目标服务器的指令集作为所述计算资源调整指令集。
  4. 根据权利要求3所述的方法，其特征在于，所述基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定目标服务器，包括：
    基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定多个候选服务器及每一个候选服务器对应的候选总功率，所述候选总功率为假设将所述待迁移计算资源迁移到所述每一个候选服务器后所述机房的总功率；
    在所述多个候选服务器中选择相应候选总功率最低的候选服务器作为所述目标服务器。
  5. 根据权利要求4所述的方法，其特征在于，所述基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定多个候选服务器及每一个候选服务器对应的候选总功率，包括：
    根据所述冷量传导关系信息确定分区冷量吸收关系信息和冷量共享关系信息；
    根据所述分区冷量吸收关系信息和所述冷量共享关系信息确定待选分区组合集合；
    根据所述冷量传导关系信息、所述待选分区组合集合和所述计算资源状态信息，为所述待迁移计算资源确定多个候选服务器及每一个候选服务器对应的候选总功率，所述多个候选服务器属于所述待选分区组合集合；
    其中，所述分区冷量吸收关系信息用于指示所述机房中各个分区中机柜对本分区中的温度调节设备的冷量吸收能力，一个分区包括一个温度调节设备以及至少一个机柜；
    所述冷量共享关系信息用于指示所述机房中每两个分区之间，一个分区的机柜对另一个分区的温度调节设备的冷量共享能力；
    所述待选分区组合集合包括具有不同分区数量的多个待选分区组合，其中，一个待选分 区组合为具有相同分区数量的一个或多个分区组合中冷量吸收能力和冷量共享能力最强的分区组合。
  6. 根据权利要求5所述的方法，其特征在于，所述基于所述冷量传导关系信息、所述待选分区组合集合和所述计算资源状态信息，为所述待迁移计算资源确定多个候选服务器，包括：
    基于所述计算资源状态信息，为所述待迁移计算资源确定所述待选分区组合集合中每个待选分区组合中的每个机柜上的备选服务器；其中，第一功率变化量不大于第二功率变化量，所述第一功率变化量为假设将所述待迁移计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量，所述第二功率变化量为假设将所述待迁移计算资源迁移至同个机柜上除所述备选服务器外的其他服务器后，所述其他服务器的功率变化量；
    基于所述计算资源状态信息和所述冷量传导关系信息，确定所述每个机柜所对应的功率均衡度，所述功率均衡度为假设将所述待迁移计算资源迁移到所述每个机柜上的备选服务器后，所述每个机柜所属的待选分区组合的功率均衡度；
    将所述每个待选分区组合中，相应功率均衡度最小的机柜上的备选服务器，组成所述多个候选服务器。
  7. 根据权利要求3所述的方法，其特征在于，所述基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定目标服务器，包括：
    基于所述计算资源状态信息，为所述待迁移计算资源确定所述机房中的每个机柜上的备选服务器；
    其中，第一功率变化量不大于第二功率变化量，所述第一功率变化量为假设将所述待迁移计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量，所述第二功率变化量为假设将所述待迁移计算资源迁移至同个机柜上除所述备选服务器外的其他服务器后，所述其他服务器的功率变化量；
    基于所述计算资源状态信息和所述冷量传导关系信息，确定所述每个机柜所对应的功率均衡度，所述功率均衡度为假设将所述待迁移计算资源迁移到所述每个机柜上的备选服务器后，所述机房的功率均衡度；
    在所述多个机柜中选择相应功率均衡度最小的机柜上的备选服务器作为所述目标服务器。
  8. 根据权利要求2-7任一项所述的方法，其特征在于，所述方法还包括：
    计算假设执行所述计算资源调整指令集后，所述多个机柜中每个机柜的功率；
    根据所述多个机柜中每个机柜的功率确定所述至少一个温度调节设备中每个温度调节设备的温度设定参数，所述温度设定参数用于满足执行所述计算资源调整指令集之后，所述多个机柜的温度需求；
    触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。
  9. 根据权利要求5或7所述的方法，其特征在于，所述方法还包括：
    计算假设执行所述计算资源调整指令集后，所述多个机柜中每个机柜的功率；
    根据所述每个温度调节设备所在的分区所包含的机柜的功率确定每个温度调节设备的温度设定参数；
    触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。
  10. 根据权利要求1所述的方法，其特征在于，所述计算资源管理指令集为计算资源分 配指令集，所述计算资源分配指令集用于在所述机房中分配至少一个计算资源。
  11. 根据权利要求10所述的方法，其特征在于，所述根据所述冷量传导关系信息和所述计算资源状态信息，触发所述设施管理设备执行计算资源管理指令集，包括：
    接收业务需求，所述业务需求用于请求运行所述机房中的计算资源；
    基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息确定，满足所述业务需求的目标计算资源及提供所述目标计算资源的目标服务器，所述目标服务器提供所述目标计算资源前后所述机房的总功率的增量最小化；
    生成用于将在所述目标服务器分配所述目标计算资源的指令集作为所述计算资源分配指令集。
  12. 根据权利要求11所述的方法，基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息确定满足所述业务需求的目标计算资源及提供所述目标计算资源的目标服务器，包括：
    基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息为所述目标计算资源确定多个候选服务器及每一个候选服务器对应的候选增量，所述候选增量为假设将所述目标计算资源迁移到所述每一个候选服务器后所述机房的总功率的增量；
    在所述多个候选服务器集合中选择相应候选增量最低的候选服务器作为所述目标服务器。
  13. 根据权利要求12所述的方法，其特征在于，基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息为所述目标计算资源确定多个候选服务器及每一个候选服务器对应的候选增量，包括：
    根据所述冷量传导关系信息确定分区冷量吸收关系信息和冷量共享关系信息；
    根据所述分区冷量吸收关系信息和所述冷量共享关系信息确定待选分区组合集合；
    根据所述业务需求、所述冷量传导关系信息、所述待选分区组合集合和所述计算资源状态信息，为所述目标计算资源确定多个候选服务器及每一个候选服务器对应的候选增量，所述多个候选服务器属于所述待选分区组合集合；
    其中，所述分区冷量吸收关系信息用于指示所述机房中各个分区中机柜对本分区中的温度调节设备的冷量吸收能力，一个分区包括一个温度调节设备以及至少一个机柜；
    所述冷量共享关系信息用于指示所述机房中每两个分区之间，一个分区的机柜对另一个分区的温度调节设备的冷量共享能力；
    所述待选分区组合集合包括具有不同分区数量的多个待选分区组合，其中，一个待选分区组合为具有相同分区数量的一个或多个分区组合中冷量吸收能力和冷量共享能力最强的分区组合。
  14. 根据权利要求13所述的方法，其特征在于，所述根据所述业务需求、所述冷量传导关系信息、所述待选分区组合集合和所述计算资源状态信息，为所述目标计算资源确定多个候选服务器，包括：
    基于所述业务需求、计算资源状态信息，为所述目标计算资源确定所述待选分区组合集合中每个待选分区组合中的每个机柜上的备选服务器；其中，所述备选服务器具备提供所述目标计算资源的能力，且第三功率变化量不大于第四功率变化量，所述第三功率变化量为假设将所述目标计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量，所述第四功率变化量为假设将所述目标计算资源迁移至同个机柜上除所述备选服务器外的其他服务器的功率变化量；
    基于所述计算资源状态信息和所述冷量传导关系信息，确定所述每个机柜所对应的功率 均衡度，所述功率均衡度为假设将所述目标计算资源迁移到所述每个机柜上的备选服务器后，所述每个机柜所属的待选分区组合的功率均衡度；
    将所述每个待选分区组合中，相应功率均衡度最小的机柜上的备选服务器，组成所述多个候选服务器。
  15. 根据权利要求11所述的方法，其特征在于，所述根据所述冷量传导关系信息和所述计算资源状态信息，触发所述设施管理设备执行计算资源管理指令集，包括：
    基于所述计算资源状态信息，为所述目标计算资源确定所述机房中的每个机柜上的备选服务器；其中，所述备选服务器具备提供所述目标计算资源的能力，且第三功率变化量不大于第四功率变化量，所述第三功率变化量为假设将所述目标计算资源迁移至所述备选服务器后，所述备选服务器的功率变化量，所述第四功率变化量为假设将所述目标计算资源迁移至同个机柜上除所述备选服务器外的其他服务器后，所述其他服务器的功率变化量；
    基于所述计算资源状态信息和所述冷量传导关系信息，确定所述每个机柜所对应的功率均衡度，所述功率均衡度为假设将所述目标计算资源迁移到所述每个机柜上的备选服务器后，所述机房的功率均衡度；
    在所述多个机柜中选择相应功率均衡度最小的机柜上的备选服务器作为所述目标服务器。
  16. 根据权利要求11-15任一项所述的方法，其特征在于，所述方法还包括：
    计算假设执行所述计算资源分配指令集后，所述多个机柜中每个机柜的功率；
    根据所述多个机柜中每个机柜的功率确定所述至少一个温度调节设备中每个温度调节设备的温度设定参数，所述温度设定参数用于满足执行所述计算资源分配指令集之后，所述多个机柜的温度需求；
    触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。
  17. 根据权利要求14或15所述的方法，其特征在于，所述方法还包括：
    计算假设执行所述计算资源分配指令集后，所述多个机柜中每个机柜的功率；
    根据所述每个温度调节设备所在的分区所包含的机柜的功率确定每个温度调节设备的温度设定参数；
    触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。
  18. 根据权利要求3或11所述的方法，其特征在于，所述机房的总功率，包括所述多个机柜的计算功率和所述至少一个温度调节设备的温控功率。
  19. 根据权利要求1-18任一项所述的方法，其特征在于，第一机柜包含于所述多个机柜，第一温度调节设备包含于所述多个温度调节设备，所述方法还包括：
    基于第一温度、第二温度、第一进风温度和第二进风温度确定第一机柜与第一温度调节设备的冷量传导关系信息；
    其中，所述第一进风温度为基于第一设定模型模拟所述机房运行而得出的所述第一机柜的进风温度，所述第一设定模型基于所述机房构建，所述第一设定模型包括：第一功率的所述多个机柜、所述第一温度的所述至少一个温度调节设备；
    所述第二进风温度为基于第二设定模型模拟所述机房运行而得出的所述第一机柜的进风温度，所述第二设定模型基于所述机房构建，所述第二设定模型包括：所述第一功率的所述多个机柜、所述第二温度的所述第一温度调节设备，以及所述第一温度的所述至少一个温度调节设备中除所述第一温度调节设备的其他温度调节设备。
  20. 根据权利要求5或14所述的方法，其特征在于，所述机房包括第一分区，所述第一分区包括第二温度调节设备和至少一个第二机柜；所述方法还包括：
    基于所述至少一个第二机柜中的每一个机柜对所述第二温度调节设备的冷量吸收能力确定，所述第一分区的分区冷量吸收关系信息。
  21. 根据权利要求5或14所述的方法，其特征在于，所述机房包括第一分区和第二分区，所述第一分区包括第二温度调节设备和至少一个第二机柜，所述第二分区包括第三温度调节设备和至少一个第三机柜；所述方法还包括：
    基于所述至少一个第二机柜中的每一个机柜对所述第三温度调节设备的冷量吸收能力，以及所述至少一个第三机柜中的每一个机柜对所述第二温度调节设备的冷量吸收能力确定，所述第一分区与所述第二分区的冷量共享关系信息。
  22. 根据权利要求5或14所述的方法，其特征在于，所述机房包括第一分区，所述第一分区包括第二温度调节设备和至少一个第二机柜；
    其中，每一个第二机柜对所述第二温度调节设备的冷量吸收能力，强于所述每一个第二机柜对所述多个温度调节设备中除所述第二温度调节设备外的其他温度调节设备的冷量吸收能力。
  23. 一种节能设备，所述节能设备用于节省机房的能耗，所述机房包括多个机柜和至少一个温度调节设备，其中每一个机柜包括至少一个服务器，每一个服务器用于提供计算资源，每一个温度调节设备用于调节所述机房的温度，所述机房还配置有设施管理设备，所述设施管理设备用于管理所述机房的计算资源和温度调节设备，其特征在于，所述节能设备包括获取单元和处理单元：
    所述获取单元，用于获取所述机房的冷量传导关系信息和所述机房的计算资源状态信息；
    所述处理单元，用于根据所述冷量传导关系信息和所述计算资源状态信息，触发设施管理设备执行计算资源管理指令集，以在满足业务需求的同时使所述机房的总功率最小化；
    其中，所述冷量传导关系信息用于确定所述机房中的每一个机柜对所述机房中的每一个温度调节设备的冷量吸收能力，所述计算资源状态信息用于确定所述机房包含的计算资源的运行位置和运行状态。
  24. 根据权利要求23所述的节能设备，其特征在于，所述计算资源管理指令集为计算资源调整指令集，所述计算资源调整指令集用于调整所述机房中至少一个计算资源的运行位置。
  25. 根据权利要求24所述的节能设备，其特征在于，所述处理单元具体用于：
    基于所述计算资源状态信息在所述机房中确定待迁移计算资源；
    基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定目标服务器，所述待迁移计算资源被迁移至所述目标服务器后所述机房的总功率小于迁移前所述机房的总功率；
    生成用于将所述待迁移计算资源迁移到所述目标服务器的指令集作为所述计算资源调整指令集。
  26. 根据权利要求25所述的节能设备，所述处理单元具体用于：
    基于所述冷量传导关系信息和所述计算资源状态信息为所述待迁移计算资源确定多个候选服务器及每一个候选服务器对应的候选总功率，所述候选总功率为假设将所述待迁移计算资源迁移到所述每一个候选服务器后所述机房的总功率；
    在所述多个候选服务器中选择相应候选总功率最低的候选服务器作为所述目标服务器。
  27. 根据权利要求24-26任一项所述的节能设备，其特征在于，所述处理单元还用于：
    计算假设执行所述计算资源调整指令集后，所述多个机柜中每个机柜的功率；
    根据所述多个机柜中每个机柜的功率确定所述至少一个温度调节设备中每个温度调节设备的温度设定参数，所述温度设定参数用于满足执行所述计算资源调整指令集之后，所述多个机柜的温度需求；
    触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。
  28. 根据权利要求23所述的节能设备，其特征在于，所述计算资源管理指令集为计算资源分配指令集，所述计算资源分配指令集用于在所述机房中分配至少一个计算资源。
  29. 根据权利要求28所述的节能设备，其特征在于，所述处理单元具体用于：
    接收业务需求，所述业务需求用于请求运行所述机房中的计算资源；
    基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息确定，满足所述业务需求的目标计算资源及提供所述目标计算资源的目标服务器，所述目标服务器提供所述目标计算资源前后所述机房的总功率的增量最小化；
    生成用于将在所述目标服务器分配所述目标计算资源的指令集作为所述计算资源分配指令集。
  30. 根据权利要求29所述的节能设备，其特征在于，所述处理单元具体用于：
    基于所述业务需求、所述冷量传导关系信息和所述计算资源状态信息为所述目标计算资源确定多个候选服务器及每一个候选服务器对应的候选增量，所述候选增量为假设将所述目标计算资源迁移到所述每一个候选服务器后所述机房的总功率的增量；
    在所述多个候选服务器集合中选择相应候选增量最低的候选服务器作为所述目标服务器。
  31. 根据权利要求28-30任一项所述的节能设备，其特征在于，所述处理单元还用于：
    获取假设执行所述计算资源分配指令集后，所述多个机柜中每个机柜的功率；
    根据所述多个机柜中每个机柜的功率确定所述至少一个温度调节设备中每个温度调节设备的温度设定参数，所述温度设定参数用于满足执行所述计算资源分配指令集之后，所述多个机柜的温度需求；
    触发所述设施管理设备执行温度调节指令集，所述温度调节指令集是根据所述温度设定参数生成的。
  32. 一种节能设备，所述节能设备用于节省机房的能耗，所述机房包括多个机柜和至少一个温度调节设备，其中每一个机柜包括至少一个服务器，每一个服务器用于提供计算资源，每一个温度调节设备用于调节所述机房的温度，所述机房还配置有设施管理设备，所述设施管理设备用于管理所述机房的计算资源和温度调节设备，其特征在于，所述节能设备包括处理器和存储器，所述存储器用于存储程序指令，所述处理器被配置用于调用所述程序指令来执行如权利要求1-22任一所述的方法。
  33. 一种计算资源供应系统，所述计算资源供应系统包括机房，所述机房包括多个机柜和至少一个温度调节设备，其中每一个机柜包括至少一个服务器，每一个服务器用于提供计算资源，每一个温度调节设备用于调节所述机房的温度，所述机房还配置有设施管理设备，所述设施管理设备用于管理所述机房的计算资源和温度调节设备，其特征在于，所述计算资源供应系统还包括如权利要求23-32任一所述的节能设备。
  34. 一种计算机可读存储介质，其特征在于，所述计算机可读存储介质用于存储指令，当所述指令被执行时，使得如权利要求1-22任一项所述的方法被实现。
PCT/CN2022/082480 2021-03-25 2022-03-23 一种应用于机房的节能方法、装置及系统 Ceased WO2022199608A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP22774264.0A EP4300262A4 (en) 2021-03-25 2022-03-23 ENERGY SAVING METHOD, APPARATUS AND SYSTEM APPLIED TO A MACHINE ROOM
US18/472,789 US20240023296A1 (en) 2021-03-25 2023-09-22 Energy-saving method applied to machine room, apparatus, and system

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110321118.4 2021-03-25
CN202110321118.4A CN115129141A (zh) 2021-03-25 2021-03-25 一种应用于机房的节能方法、装置及系统

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US18/472,789 Continuation US20240023296A1 (en) 2021-03-25 2023-09-22 Energy-saving method applied to machine room, apparatus, and system

Publications (1)

Publication Number Publication Date
WO2022199608A1 true WO2022199608A1 (zh) 2022-09-29

Family

ID=83373748

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/082480 Ceased WO2022199608A1 (zh) 2021-03-25 2022-03-23 一种应用于机房的节能方法、装置及系统

Country Status (4)

Country Link
US (1) US20240023296A1 (zh)
EP (1) EP4300262A4 (zh)
CN (1) CN115129141A (zh)
WO (1) WO2022199608A1 (zh)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116266904A (zh) * 2021-12-17 2023-06-20 华为技术有限公司 确定应用服务器的方法和装置
CN116319306A (zh) * 2023-03-28 2023-06-23 中国工商银行股份有限公司 设备部署方法、设备部署装置、电子设备及存储介质
CN118502310B (zh) * 2024-05-30 2024-11-15 中电(沈阳)能源投资有限公司 一种机房集成监控管理方法与系统
CN118984562B (zh) * 2024-08-05 2025-02-18 奇墨科技(广州)有限公司 基于使用数据的idc机房设施省电方案生成系统及方法
CN119668388A (zh) * 2025-02-24 2025-03-21 四川航启科技发展有限公司 水冷供电机柜功率匹配方法、装置、设备及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109600956A (zh) * 2018-12-10 2019-04-09 国家电网有限公司信息通信分公司 一种数据中心服务器放置位置优化算法和系统
CN109800066A (zh) * 2018-12-13 2019-05-24 中国科学院信息工程研究所 一种数据中心节能调度方法及系统
US20200034171A1 (en) * 2018-07-25 2020-01-30 Vmware, Inc. Methods and apparatus to adjust energy requirements in a data center
CN111867330A (zh) * 2020-07-24 2020-10-30 广州亿效能科技有限公司 一种基于it负载变化的通讯机楼机房调节方法和系统

Family Cites Families (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7620613B1 (en) * 2006-07-28 2009-11-17 Hewlett-Packard Development Company, L.P. Thermal management of data centers
US7676280B1 (en) * 2007-01-29 2010-03-09 Hewlett-Packard Development Company, L.P. Dynamic environmental management
US8260928B2 (en) * 2008-05-05 2012-09-04 Siemens Industry, Inc. Methods to optimally allocating the computer server load based on the suitability of environmental conditions
US8782234B2 (en) * 2008-05-05 2014-07-15 Siemens Industry, Inc. Arrangement for managing data center operations to increase cooling efficiency
CN102099791B (zh) * 2008-09-17 2012-11-07 株式会社日立制作所 信息处理系统的运用管理方法
US8560677B2 (en) * 2009-02-13 2013-10-15 Schneider Electric It Corporation Data center control
US8655610B2 (en) * 2010-03-24 2014-02-18 International Business Machines Corporation Virtual machine placement for minimizing total energy cost in a datacenter
US20120030356A1 (en) * 2010-07-30 2012-02-02 International Business Machines Corporation Maximizing efficiency in a cloud computing environment
US8467906B2 (en) * 2010-08-10 2013-06-18 Facebook, Inc. Load balancing tasks in a data center based on pressure differential needed for cooling servers
JP5891680B2 (ja) * 2011-09-27 2016-03-23 富士通株式会社 電力制御装置、電力制御方法、および電力制御プログラム
WO2016130453A1 (en) * 2015-02-09 2016-08-18 Schneider Electric It Corporation System and methods for simulation-based optimization of data center cooling equipment

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200034171A1 (en) * 2018-07-25 2020-01-30 Vmware, Inc. Methods and apparatus to adjust energy requirements in a data center
CN109600956A (zh) * 2018-12-10 2019-04-09 国家电网有限公司信息通信分公司 一种数据中心服务器放置位置优化算法和系统
CN109800066A (zh) * 2018-12-13 2019-05-24 中国科学院信息工程研究所 一种数据中心节能调度方法及系统
CN111867330A (zh) * 2020-07-24 2020-10-30 广州亿效能科技有限公司 一种基于it负载变化的通讯机楼机房调节方法和系统

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP4300262A4

Also Published As

Publication number Publication date
CN115129141A (zh) 2022-09-30
EP4300262A4 (en) 2024-09-04
US20240023296A1 (en) 2024-01-18
EP4300262A1 (en) 2024-01-03

Similar Documents

Publication Publication Date Title
WO2022199608A1 (zh) 一种应用于机房的节能方法、装置及系统
US9176508B2 (en) Managing workload distribution among computing systems to optimize heat dissipation by computing systems
CN101556493B (zh) 信息处理系统的运用管理方法及运用管理装置
US10925191B2 (en) Methods and apparatus to manage power delivery for a data center based on predicted power consumption
US10776149B2 (en) Methods and apparatus to adjust energy requirements in a data center
CN103970607B (zh) 使用等价集合来计算优化虚拟机分配的方法和装置
CN106233276B (zh) 网络可访问块存储装置的协调准入控制
US10187452B2 (en) Hierarchical dynamic scheduling
CN103677958B (zh) 一种虚拟化集群的资源调度方法及装置
Liang et al. The placement method of resources and applications based on request prediction in cloud data center
CN109375994A (zh) 基于rbf神经网络的数据中心任务温度预测及调度方法
CN106790726A (zh) 一种基于Docker云平台的优先级队列动态反馈负载均衡资源调度方法
CN102156665A (zh) 一种虚拟化系统竞争资源差异化服务方法
CN105426241A (zh) 一种基于云计算数据中心的统一资源调度节能方法
CN108762687B (zh) Io服务质量控制方法、装置、设备及存储介质
US11579942B2 (en) VGPU scheduling policy-aware migration
Sun et al. Research and simulation of task scheduling algorithm in cloud computing
Singh et al. Crow–penguin optimizer for multiobjective task scheduling strategy in cloud computing
CN115220900A (zh) 一种基于作业功耗预测的节能调度方法及系统
Li et al. An energy-aware virtual machine scheduling approach for cloud data centers
Li et al. Towards energy-efficient and thermal-aware data placement for storage clusters
Thawari et al. An efficient data locality driven task scheduling algorithm for cloud computing
Tai et al. SLA-aware data migration in a shared hybrid storage cluster
Toutova et al. Principals of Optimal Cloud Resource Management
CN111767225B (zh) 一种云计算环境下第三方软件测试系统及方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22774264

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 2022774264

Country of ref document: EP

ENP Entry into the national phase

Ref document number: 2022774264

Country of ref document: EP

Effective date: 20230928

NENP Non-entry into the national phase

Ref country code: DE