WO2022110443A1 - 一种基于容器管制的云数据中心节能方法及系统 - Google Patents

一种基于容器管制的云数据中心节能方法及系统 Download PDF

Info

Publication number
WO2022110443A1
WO2022110443A1 PCT/CN2020/139676 CN2020139676W WO2022110443A1 WO 2022110443 A1 WO2022110443 A1 WO 2022110443A1 CN 2020139676 W CN2020139676 W CN 2020139676W WO 2022110443 A1 WO2022110443 A1 WO 2022110443A1
Authority
WO
WIPO (PCT)
Prior art keywords
server
container
containers
optional
utilization
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/CN2020/139676
Other languages
English (en)
French (fr)
Inventor
徐敏贤
宋承浩
须成忠
叶可江
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.)
Shenzhen Institute of Advanced Technology of CAS
Original Assignee
Shenzhen Institute of Advanced Technology of CAS
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 Shenzhen Institute of Advanced Technology of CAS filed Critical Shenzhen Institute of Advanced Technology of CAS
Publication of WO2022110443A1 publication Critical patent/WO2022110443A1/zh
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
    • 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
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
    • G06F9/45533—Hypervisors; Virtual machine monitors
    • G06F9/45558—Hypervisor-specific management and integration aspects
    • 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
    • G06F1/3206—Monitoring of events, devices or parameters that trigger a change in power modality
    • 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
    • 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
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
    • G06F9/45533—Hypervisors; Virtual machine monitors
    • G06F9/45558—Hypervisor-specific management and integration aspects
    • G06F2009/4557—Distribution of virtual machine instances; Migration and load balancing
    • 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
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
    • G06F9/45533—Hypervisors; Virtual machine monitors
    • G06F9/45558—Hypervisor-specific management and integration aspects
    • G06F2009/45575—Starting, stopping, suspending or resuming virtual machine instances
    • 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
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
    • G06F9/45533—Hypervisors; Virtual machine monitors
    • G06F9/45558—Hypervisor-specific management and integration aspects
    • G06F2009/45595—Network integration; Enabling network access in virtual machine instances
    • 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 invention relates to the field of cloud computing energy saving, in particular, to a method and system for energy saving of cloud data centers based on container control.
  • cloud computing technology As an important part of information technology, is widely used due to its pay-as-you-go pricing model, low operating costs, high scalability and easy accessibility. Applied in many areas of life.
  • the huge energy consumption and the carbon emissions generated by cloud data centers have attracted widespread attention from researchers.
  • the high energy consumption of cloud computing data centers has also become a constraint restricting the development of cloud computing, and existing energy-saving technologies such as virtual machine migration and dynamic voltage and frequency adjustment are difficult to produce effective results when the data center is in an overall overload state.
  • Virtual machine migration Refers to an energy-saving method that minimizes energy consumption by distributing tasks among fewer machines while shutting down unused machines. Cloud computing servers can use this strategy to migrate virtual machines working on underutilized servers to other servers, and these spare servers will enter a low-power mode or shut down.
  • DVFS Dynamic Voltage Frequency Adjustment
  • the embodiments of the present invention provide an energy-saving method and system for a cloud data center based on container control, which can relieve server pressure and save energy in the case of overload according to the running state, and reduce the total energy consumption of the cloud data center the goal of.
  • a method for saving energy in a cloud data center based on container regulation including the following steps:
  • Idle servers switch to low power mode to save power.
  • monitoring whether the server is in an overloaded state includes:
  • Preset overload threshold for judging whether the server is overloaded
  • the server is considered overloaded.
  • i is the server number
  • ui is the utilization rate of server i
  • the regulator calculation formula is:
  • t represents the current time
  • ⁇ t is the value of the regulator when the time is t
  • n is the number of servers
  • the expected utilization reduction value is calculated according to the utilization calculation formula
  • the optional container deactivation policy includes a container priority policy with the lowest utilization, and the container priority policy with the lowest utilization is based on the deactivation calculation formula by selecting a set of optional containers to deactivate, so as to reduce the utilization of the overloaded host, Keep the reduced utilization below the overload threshold;
  • u' i is the utilization of the server after the optional container is deactivated, and u' i is equal to Expressed as The minimum absolute value of , dcl i is a list of deactivated optional containers;
  • the least number of containers first strategy which selects the least number of containers to be deactivated to save power consumption while deactivating fewer containers in order to provide more optional functions;
  • the random container selection strategy based on the random minimum number of containers priority strategy, randomly selects multiple containers to be deactivated to achieve the purpose of reducing energy consumption.
  • a cloud data center energy-saving system based on container control including: a cloud service repository, an execution environment, and a control center; the cloud service repository includes optional containers, and the optional containers can be The running status of the server is disabled or enabled; the execution environment provides an environment for the operation of the cloud service repository, and the control center includes the Brownout controller, the system monitor and the scheduling policy manager;
  • System Monitor is used to monitor server health and collect the status of server workload
  • the scheduling policy manager is used to provide the Brownout controller with policies that control optional container deactivation or activation;
  • the Brownout controller controls the deactivation or activation of optional containers based on the server running state.
  • the Brownout controller includes a regulator, and the value of the regulator is calculated according to the number of overloaded servers; the regulator controls the deactivation or opening of the optional container of the overloaded server;
  • scheduling policy manager includes:
  • the lowest utilization container first policy is based on the deactivation calculation formula.
  • the calculation formula includes the deactivated optional containers into the deactivated list, and the deactivated list sorts the optional containers based on the utilization in ascending order;
  • the least number of containers first strategy which selects the least number of containers to be deactivated to save power consumption while deactivating fewer containers in order to provide more optional functions;
  • the random container selection strategy based on the random minimum number of containers priority strategy, randomly selects multiple containers to be deactivated to achieve the purpose of reducing energy consumption.
  • the cloud service repository also includes necessary containers, which will always keep running when started and cannot be stopped.
  • the beneficial effect of the present invention is: by monitoring in real time whether the server in work is in a load state, if there is a server in an overload state, the expected reduction in the utilization rate of the server under the overload state is calculated, and the utilization rate value of the expected reduction is calculated according to the expected reduction. to select the strategy for deactivating optional containers, to deactivate the corresponding optional containers, and the deactivation of optional containers makes the overloaded server switch to a low-power mode to save power consumption; the server of this application can dynamically Shut down optional containers to relieve server pressure and save energy when overloaded, and reduce the total energy consumption of cloud data centers.
  • FIG. 1 is a flowchart of a method for energy saving in a cloud data center based on container control according to the present invention
  • Fig. 2 is the flow chart of monitoring whether the server is in overload state of the present invention
  • FIG. 3 is a schematic block diagram of a cloud data center energy-saving system based on container control according to the present invention.
  • a method for saving energy in a cloud data center based on container regulation includes the following steps:
  • S101 Monitor in real time whether a server is in an overloaded state.
  • S103 Based on the value of the expected utilization reduction, select an optional container deactivation policy to deactivate the corresponding optional container.
  • S104 The overloaded server switches to a low power consumption mode to save power consumption.
  • the server in operation by monitoring in real time whether the server in operation is in a load state, if there is a server in an overload state, calculate the expected reduction in the utilization rate of the server in the overload state, and select the desired reduction in utilization rate.
  • Optional container deactivation strategy deactivate the corresponding optional container, and the deactivation of the optional container makes the overloaded server switch to a low power consumption mode to save power consumption; the server of this application can dynamically close the optional container according to the running state
  • Containers can relieve server pressure and save energy in case of overload, and achieve the purpose of reducing the total energy consumption of cloud data centers.
  • the containers are divided into essential containers and optional containers.
  • the essential containers will always keep running when started and cannot be deactivated; for example, the containers related to the database will be marked as essential services.
  • Optional containers can temporarily deactivate these containers based on the state of the server. If there is mutual communication between containers, it means that there is a connection relationship between them. By definition, if an optional service is deactivated, other services connected to this optional service (optional container) will also be placed in deactivated state. If a service or content provided by a container provider is defined as optional by its creator, it can be identified as an optional service. For example, an online recommendation engine in an online store system and a spell checker in an online editor system can be set as optional services in case of limited or overloaded resources.
  • the server before monitoring whether the server is in an overloaded state, it further includes:
  • the Sliding windows algorithm is used to predict the future workload.
  • the sliding window algorithm is as follows:
  • the sliding window assigns more weight to the request rate of the nearest interval; let the sliding window size L w be a constant integer value.
  • num(k) is the actual number of requests at time interval k, and the number of requests in time interval t is taken as the average value of requests in the sliding window Lw , as shown in the following formula:
  • the time interval for requesting prediction should not be less than the time within the sliding window size Lw time interval, ie t ⁇ Lw .
  • the number of active servers can be dynamically adjusted.
  • monitoring whether the server is in an overloaded state includes:
  • S201 Preset an overload threshold for determining whether the server is overloaded.
  • the utilization rate of the server is measured by CPU (central processing unit); for example, if the overload threshold is set to 85%, when it is detected that the CPU utilization rate of the server is higher than 85%, the server will considered to be overloaded.
  • the number of servers in an overloaded state is calculated according to an overloaded server calculation formula, and the overloaded server calculation formula is:
  • i is the server number
  • ui is the utilization rate of the server
  • Tu is the overload threshold; if ui is not less than Tu , then is 1, otherwise it is 0; n 0 is the number of all servers in the server cluster that are in an overloaded state.
  • the calculation of the expected reduction in the utilization of the server in the overloaded state includes:
  • the regulator calculation formula is:
  • t represents the current time
  • ⁇ t is the value of the regulator when the time is t
  • n is the number of servers
  • the expected utilization reduction value is calculated according to the utilization calculation formula
  • the optional container deactivation strategy includes the lowest utilization container priority strategy (Lowest Utilization Container First, LUCF).
  • the lowest utilization container first strategy is to deactivate by selecting a set of optional containers based on the deactivation calculation formula , to reduce the utilization rate of overloaded hosts, so that the utilization rate after the reduction is lower than the overload threshold;
  • u' i is the utilization of the server after the optional container is deactivated, and u' i is equal to Expressed as The minimum absolute value of , dcl i is an optional container for deactivation.
  • the lowest-utilized container priority policy sorts the list in ascending order based on the calculated optional containers, so that the lowest-utilized optional containers are at the beginning of the list; the specific sorting process is as follows:
  • the size of the list of dcl i is dcl i size(); the algorithm checks worker servers one by one, if the utilization ratio of the first optional container c 0 in server i is greater than Better yet, put c 0 in the optional container deactivation list. Since the container connected to c 0 also needs to be considered, this strategy also marks all optional containers connected to c 0 as in the dataset ( for a set of data) to describe how it is connected to other optional containers. However, if the utilization of the first optional container c0 is less than the expected reduction in utilization, the LUCF policy finds a sublist of optional containers to deactivate more optional containers. This sublist is the list of other sums of utilizations closest to the desired utilization reduction. As before, put these optional containers into the deactivated optional container list and put their connection labels into the collection. Other connected optional containers can then be found algorithmically and placed in the deactivated optional container list.
  • the optional container deactivation policy further includes:
  • the least number of containers first strategy which selects the least number of containers to be deactivated to save power consumption while deactivating fewer containers in order to provide more optional functions;
  • the random container selection strategy based on the random minimum number of containers priority strategy, randomly selects multiple containers to be deactivated to achieve the purpose of reducing energy consumption.
  • MNCF Minimum Number Containers First
  • min(dcl i size()) represents the size of the smallest deactivated container list.
  • the random container selection strategy uses the uniform distribution function U(0, dcl i size()-1) to randomly select multiple optional variables container to achieve the purpose of reducing energy consumption;
  • a cloud data center energy-saving system based on container control includes: a cloud service repository, an execution environment, and a control center;
  • the cloud service repository includes optional containers, which can be The selection container can be deactivated or activated according to the running status of the server;
  • the execution environment provides an environment for the operation of the cloud service repository, and the control center includes the Brownout controller, the system monitor and the scheduling policy manager;
  • System Monitor is used to monitor server health and collect the status of server workload
  • the scheduling policy manager is used to provide the Brownout controller with policies that control optional container deactivation or activation;
  • the Brownout controller controls the deactivation or activation of optional containers based on the server running state.
  • the system monitor is used to monitor the running status of the server in real time, such as whether there is a workload on the server; if there is a server with a load, the scheduling policy manager provides the Brownout controller with a policy to control the deactivation of optional containers; Brownout controls
  • the server controls the deactivation of the optional container according to the above policy; the deactivation of the optional container makes the overloaded server switch to a low power consumption mode to save power consumption; the server of this application can dynamically close the optional container according to the running state, so that the overload Under the circumstance of relieving server pressure and saving energy, the goal of reducing the total energy consumption of cloud data center is achieved.
  • the system monitor is a component that monitors the server running status and collects the server resource consumption status. It uses third-party toolkits to support its features, such as the public APIs (Public Application Programming Interfaces) in Grid'5000, which provide metrics about infrastructure, including server health, CPU (CPU, central processing unit, central processing unit) ) real-time data such as utilization and power consumption.
  • the public APIs Public Application Programming Interfaces
  • Grid'5000 which provide metrics about infrastructure, including server health, CPU (CPU, central processing unit, central processing unit) ) real-time data such as utilization and power consumption.
  • the scheduling policy manager provides and manages policies for the Brownout controller to schedule optional containers.
  • different strategies need to be designed for different preferences. For example, if a cloud service provider wants to balance a trade-off between energy consumption and quality of service, a strategy that considers the trade-off is preferred.
  • the Brownout controller includes a regulator, and the value of the regulator is calculated according to the number of overloaded servers; the Brownout controller controls the deactivation or opening of the optional container of the overloaded server according to the value of the regulator.
  • the value of the regulator is calculated based on the regulator calculation formula, which is:
  • t represents the current time
  • ⁇ t is the value of the regulator when the time is t
  • n is the number of servers
  • the utilization calculation formula is:
  • an optional container deactivation policy is selected according to the value of the expected utilization reduction; the scheduling policy manager includes three optional containers: the container with the lowest utilization priority policy, the minimum number of containers priority policy and the random container selection policy Deactivate policy.
  • u' i is the utilization of the server after the optional container is deactivated, and u' i is equal to Expressed as The smallest absolute value of , dcl i is a list of deactivated optional containers.
  • the least utilized container first policy sorts the list in ascending order based on the computed optional containers so that the least utilized optional container is at the beginning of the list.
  • the third type random container selection strategy, based on the random minimum number of container priority strategy, randomly selects multiple containers to be deactivated to achieve the purpose of reducing energy consumption.
  • the cloud service repository further includes necessary containers, which will always keep running when started and cannot be stopped.
  • Optional containers can temporarily deactivate these containers based on server status. If there is mutual communication between containers, it means that there is a connection relationship between them. By definition, if an optional service is deactivated, other services connected to this optional service will also be placed in deactivated state. If a service or content provided by a container provider is defined as optional by its creator, it can be identified as an optional service. For example, an online recommendation engine in an online store system and a spell checker in an online editor system can be set as optional services in case of limited or overloaded resources.
  • control center further includes a model manager, and the model manager is a model for maintaining energy consumption and service quality in the server;
  • the execution environment provides a running environment for the cloud service repository; in general, the main execution environments of containers are Docker, Kubernetes and Mesos. In this application, Docker is chosen as the execution environment for optional containers.
  • the system is built on the GRID'5000 as the cloud infrastructure.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
  • Power Sources (AREA)

Abstract

本发明涉及云计算节能领域，具体涉及一种基于容器管制的云数据中心节能方法及系统，包括：实时监测是否存在服务器处于过载状态；如果存在处于过载状态下的服务器，则计算期望在处于过载状态下的服务器的利用率减少的数值；基于期望利用率减少的数值，选择可选容器停用策略，以停用相应的可选容器；空闲的服务器切换到低功耗模式以节省功耗。本发明根据期望减少的利用率数值来选择可选容器停用的策略，使相应的可选容器停用，可选容器的停用使空闲的服务器切换到低功耗模式以节省功耗；本申请服务器能根据运行状态，动态地关闭可选容器，使在过载的情况下缓解服务器压力以及起到节能的作用，并达到降低云数据中心总能耗的目的。

Description

一种基于容器管制的云数据中心节能方法及系统 技术领域
本发明涉及云计算节能领域，具体而言，涉及一种基于容器管制的云数据中心节能方法及系统。
背景技术
随着信息技术行业的发展，云计算技术作为信息技术的重要组成部分，以其现收现付的定价模式、较低的运营成本、高扩展性和易访问性等特点，使得它被广泛地应用在生活中的诸多领域。但是，巨大的能源消耗以及云数据中心产生的碳排放量的已经引起了研究人员的广泛关注。云计算数据中心的高能耗也成了制约云计算发展的一个约束，而现有的节能技术例如虚拟机迁移与动态电压频率调整在数据中心处于整体过载状态时，难以产生有效的效果。
找出一个有效的降低云计算中心服务器所消耗能量的方法已经成为了目前云计算节能领域研究的主要目标。降低云服务器的能耗不仅能降低服务器的运行成本，还能提高整体系统的可靠性。目前来说，现有的云数据中心降低能耗的主流方法有虚拟机迁移(VM consolidation)和动态电压频率调整(DVFS)。
(1)虚拟机迁移(VM consolidation)：指通过在较少的机器之间分配任务，同时关闭未使用的机器，来达到最大程度地减少能耗的节能方法。云计算服务器可以通过使用这种策略，使得那些在未充分利用的服务器上工作的虚拟机迁移到其他服务器上，而这些空余下来的服务器将会进入低能耗模式或者关闭。
(2)动态电压频率调整(DVFS)：指根据当前状态负载的大小，在计算性能和服务器能耗之间进行权衡的节能方法。举例而言，DVFS技术在处理器轻度负载时会降低频率和电压，以达到降低能耗的目的。而在机器重度负载时则会增加频率和电压。
现有的方法主要针对于在以虚拟机为粒度的云服务器能耗的优化问题上；而且当整个云服务器处于过载状态下时，现有的方法如虚拟机迁移(VM  consolidation)与动态电压频率调整(DVFS)不能有效地降低云服务器的能耗。
发明内容
本发明实施例提供了一种基于容器管制的云数据中心节能方法及系统，能根据运行状态，使在过载的情况下缓解服务器压力以及起到节能的作用，并达到降低云数据中心总能耗的目的。
根据本发明的一实施例，提供了一种基于容器管制的云数据中心节能方法，包括以下步骤：
实时监测是否存在服务器处于过载状态；
如果存在处于过载状态下的服务器，则计算期望在处于过载状态下的服务器的利用率减少的数值；
基于期望利用率减少的数值，选择可选容器停用策略，以停用相应的可选容器；
空闲的服务器切换到低功耗模式以节省功耗。
进一步地，在监测服务器是否处于过载状态前还包括：
基于历史数据对服务器的工作负载进行预测；
依据预测得出的工作负载，调整进行工作的服务器的数量。
进一步地，在监测服务器是否处于过载状态中包括：
预设用于判断服务器是否过载的过载阈值；
检测服务器的利用率；
将利用率与过载阈值进行对比；
若利用率高于过载阈值，则服务器被视为处于过载状态。
进一步地，根据过载服务器计算公式计算处于过载状态的服务器数量：
过载服务器计算公式为：
Figure PCTCN2020139676-appb-000001
Figure PCTCN2020139676-appb-000002
其中，
Figure PCTCN2020139676-appb-000003
用来表示服务器是否处于过载状态，i为服务器编号，u i为服务器i的利用率，T u为过载阈值；如果u i不小于T u，则
Figure PCTCN2020139676-appb-000004
为1，否则为0；n 0为服务器集群中处于过载状态的服务器的全部数量。
进一步地，在计算期望在处于过载状态下的服务器的利用率减少的数值中包括：
首先基于调节器计算公式计算过载状态下服务器的调节器的值，调节器计算公式为：
Figure PCTCN2020139676-appb-000005
其中，t表示当前时间，θ t为当时间为t时调节器的值，n为服务器数量；
然后基于调节器的值根据利用率计算公式计算期望利用率减少的数值；
利用率计算公式为：
Figure PCTCN2020139676-appb-000006
其中，
Figure PCTCN2020139676-appb-000007
是期望利用率减少的数值。
进一步地，可选容器停用策略中包括利用率最低的容器优先策略，利用率最低的容器优先策略是基于停用计算公式通过选择一组可选容器停用，以减少过载主机的利用率，使减少后的利用率低于过载阈值；
停用计算公式为：
Figure PCTCN2020139676-appb-000008
其中，u' i为可选容器停用后服务器的利用率，u' i等于
Figure PCTCN2020139676-appb-000009
表 示为
Figure PCTCN2020139676-appb-000010
的最小绝对值，dcl i为停用的可选容器形成的列表；
可选容器停用策略中还包括：
最少数量的容器优先策略，该策略选择停用最少的容器，以在达到节省功耗的同时，能停用更少的容器，以便提供更多可选功能；
随机容器选择策略，基于随机的最小数量的容器优先策略，随机选择多个容器停用来达到减少能源消耗的目的。
根据本发明的另一实施例，提供了一种基于容器管制的云数据中心节能系统，包括：云服务存储库、执行环境、调控中心；云服务存储库包括可选容器，可选容器能根据服务器的运行状态停用或开启；执行环境为云服务存储库的运行提供环境，调控中心包括Brownout控制器、系统监视器及调度策略管理器；
系统监视器用于监视服务器运行状况并收集服务器工作负载的状态；
调度策略管理器用于为Brownout控制器提供控制可选容器停用或开启的策略；
Brownout控制器根据服务器运行状态控制可选容器的停用或开启。
进一步地，Brownout控制器中包含调节器，调节器的值根据过载的服务器数量计算得出；调节器控制过载服务器的可选容器的停用或开启；
进一步地，调度策略管理器包括：
利用率最低的容器优先策略，利用率最低的容器优先策略基于停用计算公式通过选择一组容器并停用以减少过载主机的利用率，使减少后的利用率低于过载阈值；通过停用计算公式将停用的可选容器纳入停用列表，停用列表基于利用率升序对可选容器进行排序；
最少数量的容器优先策略，该策略选择停用最少的容器，以在达到节省功耗的同时，能停用更少的容器，以便提供更多可选功能；
随机容器选择策略，基于随机的最小数量的容器优先策略，随机选择多个容器停用来达到减少能源消耗的目的。
进一步地，云服务存储库还包括必要容器，必要容器启动时将始终保持运行状态，无法被停止。
本发明的有益效果在于：通过实时监测工作中的服务器是否处于负载状况态，如果存在处于过载状态下的服务器，则计算过载状态下的服务器利用率期望减少的数量，根据期望减少的利用率数值来选择可选容器停用的策略，使相应的可选容器停用，可选容器的停用使过载的服务器切换到低功耗模式以节省功耗；本申请服务器能根据运行状态，动态地关闭可选容器，使在过载的情况下缓解服务器压力以及起到节能的作用，并达到降低云数据中心总能耗的目的。
附图说明
此处所说明的附图用来提供对本发明的进一步理解，构成本申请的一部分，本发明的示意性实施例及其说明用于解释本发明，并不构成对本发明的不当限定。在附图中：
图1为本发明基于容器管制的云数据中心节能方法的流程图；
图2为本发明监测服务器是否处于过载状态中的流程图；
图3为本发明基于容器管制的云数据中心节能系统的原理框图。
具体实施方式
为了使本技术领域的人员更好地理解本发明方案，下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分的实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都应当属于本发明保护的范围。
需要说明的是，本发明的说明书和权利要求书及上述附图中的术语“第一”、“第二”等是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。应该理解这样使用的数据在适当情况下可以互换，以便这里描述的本发明的实施例能够以除了在这里图示或描述的那些以外的顺序实施。此外，术语“包括”和“具有” 以及他们的任何变形，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、产品或设备不必限于清楚地列出的那些步骤或单元，而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备固有的其它步骤或单元。
实施例1
根据本发明一实施例，提供了一种基于容器管制的云数据中心节能方法，参见图1和图2，包括以下步骤：
S101：实时监测是否存在服务器处于过载状态。
S102：如果存在处于过载状态下的服务器，则计算期望在处于过载状态下的服务器的利用率减少的数值。
S103：基于期望利用率减少的数值，选择可选容器停用策略，以停用相应的可选容器。
S104：过载的服务器切换到低功耗模式以节省功耗。
本实施例中，通过实时监测工作中的服务器是否处于负载状况态，如果存在处于过载状态下的服务器，则计算过载状态下的服务器利用率期望减少的数量，根据期望减少的利用率数值来选择可选容器停用的策略，使相应的可选容器停用，可选容器的停用使过载服务器切换到低功耗模式以节省功耗；本申请服务器能根据运行状态，动态地关闭可选容器，使在过载的情况下缓解服务器压力以及起到节能的作用，并达到降低云数据中心总能耗的目的。
本实施例中，容器分为必要容器和可选容器，必要容器在启动时将始终保持运行状态，无法被停用；例如与数据库相关的容器将被标记为必要服务。
可选容器可以根据服务器的状态暂时停用这些容器。如果容器之间存在相互通信，则表示它们之间存在连接关系。通过定义，如果一个可选服务被停用，那么与这个可选服务(可选容器)相连接的其它服务也将被置于停用状态。如果容器提供者提供的服务或内容被其创建者定义为可选，则可以将其标识为可选服务。例如，在资源受限或过载的情况下，可以将在线商店系统中的在线推荐引擎和在 线编辑器系统中的拼写检查器设置为可选服务。
作为优选的技术方案中，在监测服务器是否处于过载状态前还包括：
基于历史数据对服务器的工作负载进行预测。
依据预测得出的工作负载，调整进行工作的服务器的数量。
本实施例中，依据现有的工作负载，使用Sliding windows(滑动窗口)算法来预测未来的工作负载。滑动窗口算法如下所示：
设定滑动窗口的大小L w，在时间间隔为L w内的请求数量，t为待预测的时间间隔(t≥L w)；
Figure PCTCN2020139676-appb-000011
为在时间间隔t内请求数量的算法输出的预测值；k表示时间，当k处于t-L w与t-1之间时
Figure PCTCN2020139676-appb-000012
直至
Figure PCTCN2020139676-appb-000013
时返回
Figure PCTCN2020139676-appb-000014
的值。
滑动窗口将更多的权重赋给相隔最近的时间间隔的请求率；令滑动窗口大小L w为恒定的整数值。num(k)是在时间间隔k处的实际请求数量，以时间间隔t内的请求数量作为滑动窗口L w中请求的平均值，如下公式所示：
Figure PCTCN2020139676-appb-000015
为了确保有足够的历史数据可用于预测，请求预测的时间间隔不应小于滑动窗口大小L w时间间隔内的时间，即t≥L w。依据预测得出的工作负载，处于工作的服务器数量可以被动态地调整。
本实施例中，在监测服务器是否处于过载状态中包括：
S201：预设用于判断服务器是否过载的过载阈值。
S202：检测服务器的利用率。
S203：将利用率与过载阈值进行对比。
S204：若利用率高于过载阈值，则服务器被视为处于过载状态。
本实施例中，服务器的利用率采用CPU(中央处理器central processing unit) 来衡量；例如，将过载阈值设定为85％，当检测到服务器的CPU利用率高于85％时，则服务器将被视为处于过载状态。
本实施例中，根据过载服务器计算公式计算处于过载状态的服务器数量，过载服务器计算公式为：
Figure PCTCN2020139676-appb-000016
Figure PCTCN2020139676-appb-000017
其中，
Figure PCTCN2020139676-appb-000018
用来表示服务器是否处于过载状态，i为服务器编号，u i为服务器的利用率，T u为过载阈值；如果u i不小于T u，则
Figure PCTCN2020139676-appb-000019
为1，否则为0；n 0为服务器集群中处于过载状态的全部服务器的数量。
本实施例中，在计算期望在处于过载状态下的服务器的利用率减少利用率的数值中包括：
首先基于调节器计算公式计算过载状态下服务器的调节器的值，调节器计算公式为：
Figure PCTCN2020139676-appb-000020
其中，t表示当前时间，θ t为当时间为t时调节器的值，n为服务器数量；
然后基于调节器的值根据利用率计算公式计算期望利用率减少的数值；
利用率计算公式为：
Figure PCTCN2020139676-appb-000021
其中，
Figure PCTCN2020139676-appb-000022
是期望利用率减少的数值。
本实施例中，可选容器停用策略包括利用率最低的容器优先策略(Lowest Utilization Container First，LUCF)，利用率最低的容器优先策略是基于停用计算公式通过选择一组可选容器停用，以减少过载主机的利用率，使减少后的利用率 低于过载阈值；
停用计算公式为：
Figure PCTCN2020139676-appb-000023
其中，u' i为可选容器停用后服务器的利用率，u' i等于
Figure PCTCN2020139676-appb-000024
表示为
Figure PCTCN2020139676-appb-000025
的最小绝对值，dcl i为停用的可选容器。
在可选容器停用之后，利用率最低的容器优先策略根据将计算出的可选容器按照升序进行排序列表，以使利用率最低的可选容器位于列表的开头；具体排序过程如下：
dcl i的列表的大小是dcl isize()；该算法逐一检查工作服务器，如果在服务器i中的第一个可选容器c 0的利用率比
Figure PCTCN2020139676-appb-000026
要好，则将c 0置入可选容器停用列表中。由于还需要考虑到与c 0相连接的容器，所以该策略也将所有与c 0相连接的可选容器全部标记为数据集中的
Figure PCTCN2020139676-appb-000027
(
Figure PCTCN2020139676-appb-000028
为一组数据)来描述它是如何与其它可选容器相连的。但是，如果第一个可选容器c 0的利用率小于期望的利用率减少，则LUCF策略会找到一个可选容器子列表来停用更多可选容器。该子列表是其它最接近期望利用率减少的利用率之和的列表。与前述的操作相同，将这些可选容器放入已停用的可选容器列表中，并将它们的连接标签放入集合中。之后就能依据算法找到其它已连接的可选容器并将其放入停用的可选容器列表中。
本实施例中，可选容器停用策略中还包括：
最少数量的容器优先策略，该策略选择停用最少的容器，以在达到节省功耗的同时，能停用更少的容器，以便提供更多可选功能；
随机容器选择策略，基于随机的最小数量的容器优先策略，随机选择多个容器停用来达到减少能源消耗的目的。
具体的，为了停用更少的容器，以便提供更多可选功能，实施最少数量的容器优先策略(Minimum Number Containers First，MNCF)，该策略选择最小数量 的容器，同时节省了功耗。与LUCF非常相似；
如公式所示：
Figure PCTCN2020139676-appb-000029
其中min(dcl isize())表示最小的已停用容器列表的大小。
具体的，随机选择的dcl i子集的均匀分布离散随机变量，随机容器选择策略(Random container selection，RCS)使用均匀分布函数U(0,dcl isize()-1)随机选择多个可选容器来达到减少能源消耗的目的；
如公式所示：
Figure PCTCN2020139676-appb-000030
实施例2
根据本发明的另一实施例，提供了一种基于容器管制的云数据中心节能系统，参见图3，包括：云服务存储库、执行环境、调控中心；云服务存储库包括可选容器，可选容器能根据服务器的运行状态停用或开启；执行环境为云服务存储库的运行提供环境，调控中心包括Brownout控制器、系统监视器及调度策略管理器；
系统监视器用于监视服务器运行状况并收集服务器工作负载的状态；
调度策略管理器用于为Brownout控制器提供控制可选容器停用或开启的策略；
Brownout控制器根据服务器运行状态控制可选容器的停用或开启。
本实施例中，通过系统监视器实时监测服务器的运行状况，如服务器是否存在工作负载；如果存在负载的服务器，则调度策略管理器为Brownout控制器提 供控制可选容器停用的策略；Brownout控制器根据上述策略控制可选容器停用；可选容器的停用使过载的服务器切换到低功耗模式以节省功耗；本申请服务器能根据运行状态，动态地关闭可选容器，使在过载的情况下缓解服务器压力以及起到节能的作用，并达到降低云数据中心总能耗的目的。
本实施例中，系统监视器是监视服务器运行状况并收集服务器资源消耗状态的组件。它使用第三方工具包来支持其功能，例如Grid’5000中的public API(公共应用编程接口)，这些API提供有关基础结构指标，包括服务器运行状况，CPU(CPU，central processing unit，中央处理器)利用率和功耗等实时数据。
本实施例中，调度策略管理器为Brownout控制器提供和管理策略以调度可选容器。为了确保能源预算和服务质量约束，需要针对不同偏好设计不同的策略。例如，如果云服务提供商想要平衡能量消耗和服务质量之间的折衷，则首选考虑折衷的策略。
本实施例中，Brownout控制器中包含调节器，调节器的值根据过载的服务器数量计算得出；Brownout控制器依据调节器的值控制过载服务器的可选容器的停用或开启。
调节器的值基于调节器计算公式计算，调节器计算公式为：
Figure PCTCN2020139676-appb-000031
其中，t表示当前时间，θ t为当时间为t时调节器的值，n为服务器数量；
基于调节器的值根据利用率计算公式计算期望利用率减少的数值；利用率计算公式为：
Figure PCTCN2020139676-appb-000032
其中，
Figure PCTCN2020139676-appb-000033
是期望利用率减少的数值。
本实施例中，根据期望利用率减少的数值选择可选容器停用策略；调度策略管理器包括：利用率最低的容器优先策略、最少数量的容器优先策略及随机容器选择策略三种可选容器停用策略。
第一种：利用率最低的容器优先策略，基于停用计算公式通过选择一组容器并停用以减少过载主机的利用率，使减少后的利用率低于过载阈值；通过停用计算公式计算并将停用的可选容器归纳为停用列表，停用列表基于利用率升序对可选容器进行排序；
停用计算公式为：
Figure PCTCN2020139676-appb-000034
其中，u' i为可选容器停用后服务器的利用率，u' i等于
Figure PCTCN2020139676-appb-000035
表示为
Figure PCTCN2020139676-appb-000036
的最小绝对值，dcl i为停用的可选容器形成的列表。
在可选容器列停用之后，利用率最低的容器优先策略根据将计算出的可选容器按照升序进行排序列表，以使利用率最低的可选容器位于列表的开头。
第二种：最少数量的容器优先策略，该策略选择停用最少的容器，以在达到节省功耗的同时，能停用更少的容器，以便提供更多可选功能；
第三种：随机容器选择策略，基于随机的最小数量的容器优先策略，随机选择多个容器停用来达到减少能源消耗的目的。
本实施例中，云服务存储库还包括必要容器，必要容器启动时将始终保持运行状态，无法被停止。
可选容器可以根据服务器状态暂时停用这些容器。如果容器之间存在相互通信，则表示它们之间存在连接关系。通过定义，如果一个可选服务被停用，那么与这个可选服务相连接的其他服务也将被置于停用状态。如果容器提供者提供的服务或内容被其创建者定义为可选，则可以将其标识为可选服务。例如，在资源受限或过载的情况下，可以将在线商店系统中的在线推荐引擎和在线编辑器系统中的拼写检查器设置为可选服务。
本实施例中，调控中心还包括模型管理器，模型管理器为用于维护服务器中的能量消耗和服务质量的模型；
本实施例中，执行环境为云服务存储库提供运行的环境；通常情况容器的主要执行环境是Docker，Kubernetes和Mesos。在本申请中，选择使用Docker作为可选容器的执行环境。
本实施例中，系统建立在以GRID’5000作为云基础架构之上。
以上所述仅是本发明的优选实施方式，应当指出，对于本技术领域的普通技术人员来说，在不脱离本发明原理的前提下，还可以做出若干改进和润饰，这些改进和润饰也应视为本发明的保护范围。

Claims (10)

  1. 一种基于容器管制的云数据中心节能方法，其特征在于，包括以下步骤：
    实时监测是否存在服务器处于过载状态；
    如果存在处于过载状态下的服务器，则计算期望在处于过载状态下的所述服务器的利用率减少的数值；
    基于期望所述利用率减少的数值，选择可选容器停用策略，以停用相应的可选容器；
    过载的所述服务器切换到低功耗模式以节省功耗。
  2. 根据权利要求1所述的基于容器管制的云数据中心节能方法，其特征在于，在所述实时监测是否存在服务器处于过载状态前还包括：
    基于历史数据对所述服务器的工作负载进行预测；
    依据预测得出的所述工作负载，调整进行工作的所述服务器的数量。
  3. 根据权利要求1所述的基于容器管制的云数据中心节能方法，其特征在于，在所述监测服务器是否处于过载状态中包括：
    预设用于判断所述服务器是否过载的过载阈值；
    检测所述服务器的利用率；
    将所述利用率与所述过载阈值进行对比；
    若所述利用率高于所述过载阈值，则所述服务器被视为处于过载状态。
  4. 根据权利要求3所述的基于容器管制的云数据中心节能方法，其特征在于，根据过载服务器计算公式计算处于过载状态的所述服务器数量：
    所述过载服务器计算公式为：
    Figure PCTCN2020139676-appb-100001
    Figure PCTCN2020139676-appb-100002
    其中，
    Figure PCTCN2020139676-appb-100003
    用来表示所述服务器是否处于过载状态，i为服务器编号，u i为服务器i的利用率，T u为所述过载阈值；如果u i不小于T u，则
    Figure PCTCN2020139676-appb-100004
    为1，否则为0；n 0为所述服务器集群中处于过载状态的全部数量。
  5. 根据权利要求4所述的基于容器管制的云数据中心节能方法，其特征在于，在所述计算期望在处于过载状态下的所述服务器的利用率减少的数值中包括：
    首先基于调节器计算公式计算过载状态下所述服务器的调节器的值，所述调节器计算公式为：
    Figure PCTCN2020139676-appb-100005
    其中，t表示当前时间，θ t为当时间为t时调节器的值，n为服务器数量；
    然后基于所述调节器的值根据利用率计算公式计算期望所述利用率减少的数值；
    所述利用率计算公式为：
    Figure PCTCN2020139676-appb-100006
    其中，
    Figure PCTCN2020139676-appb-100007
    是期望利用率减少的数值。
  6. 根据权利要求5所述的基于容器管制的云数据中心节能方法，其特征在于，所述可选容器停用策略中包括利用率最低的容器优先策略，所述利用率最低的容器优先策略是基于停用计算公式通过选择一组所述可选容器停用，以减少过载主机的利用率，使减少后的所述利用率低于所述过载阈值；
    所述停用计算公式为：
    Figure PCTCN2020139676-appb-100008
    其中，u i'为可选容器停用后服务器利用率，u i'等于
    Figure PCTCN2020139676-appb-100009
    表示 为
    Figure PCTCN2020139676-appb-100010
    的最小绝对值，dcl i为停用的可选容器形成的列表；
    所述可选容器停用策略中还包括：
    所述最少数量的容器优先策略，该策略选择停用最少的容器，以在达到节省功耗的同时，能停用更少的容器，以便提供更多可选功能；
    随机容器选择策略，基于随机的最小数量的容器优先策略，随机选择多个容器停用来达到减少能源消耗的目的。
  7. 一种基于容器管制的云数据中心节能系统，其特征在于，包括：云服务存储库、执行环境、调控中心；所述云服务存储库包括可选容器，所述可选容器能根据服务器的运行状态停用或开启；所述执行环境为云服务存储库的运行提供环境，所述调控中心包括Brownout控制器、系统监视器及调度策略管理器；
    所述系统监视器用于监视所述服务器运行状况并收集所述服务器工作负载的状态；
    所述调度策略管理器用于为所述Brownout控制器提供控制可选容器停用或开启的策略；
    所述Brownout控制器根据所述服务器运行状态控制可选容器的停用或开启。
  8. 根据权利要求7所述的基于容器管制的云数据中心节能系统，其特征在于，所述Brownout控制器中包含调节器，所述调节器的值根据过载的所述服务器数量计算得出；所述调节器控制所述过载服务器的所述可选容器的停用或开启。
  9. 根据权利要求8所述的基于容器管制的云数据中心节能系统，其特征在于，所述调度策略管理器包括：
    利用率最低的容器优先策略，所述利用率最低的容器优先策略基于停用计算公式通过选择一组容器并停用以减少过载主机的利用率，使减少后的所述利用率低于所述过载阈值；通过停用计算公式将停用的所述可选容器纳入停用列表，所述停用列表基于所述利用率升序对所述可选容器进行排序；
    所述最少数量的容器优先策略，该策略选择停用最少的容器，以在达到节省 功耗的同时，能停用更少的容器，以便提供更多可选功能；
    所述随机容器选择策略，基于随机的最小数量的容器优先策略，随机选择多个容器停用来达到减少能源消耗的目的。
  10. 根据权利要求7所述的基于容器管制的云数据中心节能系统，其特征在于，所述云服务存储库还包括必要容器，所述必要容器启动时将始终保持运行状态，无法被停止。
PCT/CN2020/139676 2020-11-24 2020-12-25 一种基于容器管制的云数据中心节能方法及系统 Ceased WO2022110443A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011332616.0 2020-11-24
CN202011332616.0A CN112433815A (zh) 2020-11-24 2020-11-24 一种基于容器管制的云数据中心节能方法及系统

Publications (1)

Publication Number Publication Date
WO2022110443A1 true WO2022110443A1 (zh) 2022-06-02

Family

ID=74694043

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/139676 Ceased WO2022110443A1 (zh) 2020-11-24 2020-12-25 一种基于容器管制的云数据中心节能方法及系统

Country Status (2)

Country Link
CN (1) CN112433815A (zh)
WO (1) WO2022110443A1 (zh)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103605418A (zh) * 2013-10-23 2014-02-26 曙光信息产业(北京)有限公司 集群服务器的功耗调节方法和装置
CN109981396A (zh) * 2019-01-22 2019-07-05 平安普惠企业管理有限公司 docker服务容器集群的监控方法及装置、介质及电子设备
CN110347498A (zh) * 2019-06-10 2019-10-18 华南理工大学 一种容器和虚拟机混合云环境下的负载动态迁移方法

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5378946B2 (ja) * 2009-10-26 2013-12-25 株式会社日立製作所 サーバ管理装置およびサーバ管理方法
CN108170522B (zh) * 2017-12-06 2021-06-01 南京邮电大学 一种基于动态阈值的云计算虚拟机迁移控制方法
CN109901932B (zh) * 2019-03-12 2023-04-07 东北大学 一种基于虚拟机的服务器整合方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103605418A (zh) * 2013-10-23 2014-02-26 曙光信息产业(北京)有限公司 集群服务器的功耗调节方法和装置
CN109981396A (zh) * 2019-01-22 2019-07-05 平安普惠企业管理有限公司 docker服务容器集群的监控方法及装置、介质及电子设备
CN110347498A (zh) * 2019-06-10 2019-10-18 华南理工大学 一种容器和虚拟机混合云环境下的负载动态迁移方法

Also Published As

Publication number Publication date
CN112433815A (zh) 2021-03-02

Similar Documents

Publication Publication Date Title
Arshad et al. Utilizing power consumption and SLA violations using dynamic VM consolidation in cloud data centers
Zhou et al. Minimizing SLA violation and power consumption in Cloud data centers using adaptive energy-aware algorithms
CN105159751B (zh) 一种云数据中心中能量高效的虚拟机迁移方法
Wang et al. Energy-aware dynamic virtual machine consolidation for cloud datacenters
CN113672383B (zh) 一种云计算资源调度方法、系统、终端以及存储介质
US8286165B2 (en) Server management apparatus and server management method
CN113535409B (zh) 一种面向能耗优化的无服务器计算资源分配系统
Abdelsamea et al. Virtual machine consolidation challenges: A review
CN102662750A (zh) 基于弹性虚拟机池的虚拟机资源优化控制方法及其系统
US20090254909A1 (en) Methods and Apparatus for Power-aware Workload Allocation in Performance-managed Computing Environments
US8583952B2 (en) Power management of computers based on user inactivity and power state requirements for active processes
CN103970256B (zh) 一种基于内存压缩和cpu动态调频的节能方法及系统
Zhou et al. Energy-efficient virtual machine consolidation algorithm in cloud data centers
CN115665158A (zh) 一种容器集群服务动态管理方法及系统
Garg et al. Energy efficient virtual machine migration approach with SLA conservation in cloud computing
Jiang et al. Adaptive thresholds determination for saving cloud energy using three-way decisions
Melhem et al. Selection process approaches in live migration: A comparative study
Zhang et al. An Energy and SLA‐Aware Resource Management Strategy in Cloud Data Centers
WO2022104500A1 (zh) 一种负载控制方法、装置、计算机设备及存储介质
Jing et al. Hlem-vmp: An effective virtual machine placement algorithm for minimizing sla violations in cloud data centers
Mandal et al. PbV mSp: A priority-based VM selection policy for VM consolidation in green cloud computing
Tan et al. A cost-aware operator migration approach for distributed stream processing system
WO2022110443A1 (zh) 一种基于容器管制的云数据中心节能方法及系统
CN120179054A (zh) 一种计算机节能状态控制方法及系统
Vaneet Kumar et al. Load aware hotspot selection for SLA improvement in cloud computing and protect environment by reduction In CO2 Emissions

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: 20963331

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20963331

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 20963331

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 27.05.2024)

122 Ep: pct application non-entry in european phase

Ref document number: 20963331

Country of ref document: EP

Kind code of ref document: A1