WO2020237728A1 - 一种支持性能保障的操作模式虚拟机数量评估方法 - Google Patents

一种支持性能保障的操作模式虚拟机数量评估方法 Download PDF

Info

Publication number
WO2020237728A1
WO2020237728A1 PCT/CN2019/090867 CN2019090867W WO2020237728A1 WO 2020237728 A1 WO2020237728 A1 WO 2020237728A1 CN 2019090867 W CN2019090867 W CN 2019090867W WO 2020237728 A1 WO2020237728 A1 WO 2020237728A1
Authority
WO
WIPO (PCT)
Prior art keywords
virtual machines
adjust
residence time
operating mode
demand
Prior art date
Application number
PCT/CN2019/090867
Other languages
English (en)
French (fr)
Inventor
郭军
刘文凤
张斌
刘晨
侯帅
侯凯
李薇
柳波
王嘉怡
王馨悦
张瀚铎
张娅杰
Original Assignee
东北大学
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 东北大学 filed Critical 东北大学
Publication of WO2020237728A1 publication Critical patent/WO2020237728A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46Multiprogramming arrangements
    • G06F9/50Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5005Allocation of resources, e.g. of the central processing unit [CPU] to service a request
    • G06F9/5027Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46Multiprogramming arrangements
    • G06F9/50Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5061Partitioning or combining of resources
    • G06F9/5077Logical partitioning of resources; Management or configuration of virtualized resources
    • YGENERAL 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
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02DCLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00Energy efficient computing, e.g. low power processors, power management or thermal management

Definitions

  • the invention belongs to the field of cloud computing, and specifically relates to a method for evaluating the number of virtual machines in an operating mode supporting performance guarantee.
  • the present invention proposes a method for evaluating the number of virtual machines in an operating mode that supports performance guarantees, which specifically includes: the predicted value of concurrency X p , the current number of virtual machines in the system VM current , and the down-status residence time left, which needs to be adjusted
  • the number of virtual machines VM adjust .
  • the Random Forest concurrency can be obtained prediction value X P, X P obtained based on the demand for resources C demand to meet the system requirements of concurrency, and the demand for the amount of resources obtained, the minimum operation mode setting service number of virtual machines VM min
  • the number of working virtual machines in the system cannot be switched from continuously decreasing or maintaining a low value for a long time to a high value instantaneously, which will cause QoS degradation; setting ⁇ is the user's response time to the request Requirements, the larger the epsilon, the smaller the VM adjust, the smaller the total number of VMs in the system, the longer the response time of the corresponding request, and the increase in energy consumption.
  • a method for evaluating the number of virtual machines in an operating mode that supports performance guarantees which specifically includes the following steps:
  • Step 1 Initialize the parameters, including: the user's request response time epsilon, the minimum number of virtual machines VM min ;
  • Step 2 Calculate the available resource C available and the resource demand C demand , the formula is as follows:
  • C 0 is the demand for resources in a single request
  • X P is the predicted value of the amount of concurrency determined by the random forest
  • Step 3 Find the number of virtual machines VM adjust that needs to be adjusted, the formula is as follows:
  • reflects the user's request response time requirements: the larger the ⁇ , the smaller the VM adjust , and the fewer the total number of virtual machines in the system, and the corresponding request response time becomes longer, and vice versa, and q is the load.
  • Step 4 According to the size relationship between the current number of virtual machines VM current , the number of virtual machine adjustments VM adjust and the minimum number of virtual machines VM min , adjust the number of virtual machines in the operating mode to deal with the sudden concurrency of services;
  • Step 4.1 If the sum of the current number of virtual machines VM current and the number of virtual machine adjustments VM adjust is greater than the minimum number of virtual machines VM min , that is, VM current + VM adjust > VM min , then go to step 4.2, otherwise, go to step 4.3;
  • the present invention adopts a method for evaluating the number of operating mode virtual machines that supports performance guarantees.
  • resource adjustments are made in time to prevent excess or insufficient performance and affect service quality .
  • the classic concurrency prediction method is used to obtain the prediction result of concurrency.
  • the performance is insufficient, a certain number of virtual machines need to be transferred from the hot backup to the operation mode.
  • the transfer of virtual machines from hot backup reduces reliability.
  • FIG. 1 is a flowchart of a method for evaluating the number of virtual machines in an operation mode supporting performance guarantee according to an embodiment of the present invention.
  • a method for evaluating the number of virtual machines in an operating mode that supports performance guarantee specifically includes the following steps:
  • Step 1 Initialize the parameters, including: the user's request response time ⁇ is set to 1, and the minimum number of virtual machines VM min is 1;
  • Step 2 Calculate the available resource C available and the resource demand C demand , the formula is as follows:
  • C 0 is the demand for resources in a single request
  • X P is the predicted value of the amount of concurrency determined by the random forest
  • Step 3 Calculate the number of virtual machines that need to be adjusted VM adjust , the formula is as follows:
  • reflects the user's request response time requirements: the larger the ⁇ , the smaller the VM adjust , the fewer the total number of virtual machines in the system, and the corresponding request response time becomes longer, and vice versa, and q is the load.
  • Step 4 According to the size relationship between the current number of virtual machines VM current , the number of virtual machine adjustments VM adjust and the minimum number of virtual machines VM min , adjust the number of virtual machines in the operating mode to deal with the sudden concurrency of services;
  • Step 4.1 If the sum of the current number of virtual machines VM current and the number of virtual machine adjustments VM adjust is greater than the minimum number of virtual machines VM min , that is, VM current + VM adjust > VM min , then go to step 4.2, otherwise, go to step 4.3;

Landscapes

  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Supply And Distribution Of Alternating Current (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

本发明提出一种支持性能保障的操作模式虚拟机数量评估方法,包括:初始化参数;求出资源可用量和资源需求量;求出需要调整的虚拟机数量;据虚拟机当前数量、虚拟机调整数量的和与虚拟机的最少数量之间的大小关系,调整操作模式虚拟机数量来应对服务突发并发量。根据并发量预测结果调整操作模式虚拟机数量以满足用户需求,当性能不足时,需要从热备份转移一定数量的虚拟机到操作模式。当性能过高时,减少一部分操作模式虚拟机会使系统可靠性增加并减低能耗。确定操作模式转移数量不仅会对性能造成影响,还会影响可靠性和能耗,本发明制定相应操作模式虚拟机资源调整策略,建立支持性能保障的操作模式数量调整方法。

Description

一种支持性能保障的操作模式虚拟机数量评估方法 技术领域
本发明属于云计算领域,具体涉及一种支持性能保障的操作模式虚拟机数量评估方法。
背景技术
在传统的云资源优化调度过程中,性能往往作为首要的目标加以考虑,很多云系统在构建时都是按照满足最大服务性能所需的资源进行配置的,然而这种方案在实际应用过程中会带来资源利用率低、能耗浪费严重等诸多问题。随着互联网不断发展,云系统规模逐渐扩大,这些问题越发严重。而且性能和能耗指标之间存在着复杂的制约关系,这使得在诸多的场景下,性能和能耗往往不可兼得。为执行一个任务分配更多虚拟机可以有效地提升性能,但是同时也会带来更多的能耗。因此,我们需要在性能和能耗之间进行权衡进而找到一个合理的优化方案。
由于静态的混合备用机制无法满足动态变化的服务并发量和服务失效率的运行场景,因此,必须采用动态优化策略,根据并发业务的变化和虚拟机失效率趋势,调整虚拟机的数量在支持性能保障条件下评估操作模式虚拟机的数量并进行调整以尽量降低能耗。
发明内容
基于以上技术不足,本发明提出一种支持性能保障的操作模式虚拟机数量评估方法,具体包括:并发量预测值X p,系统当前的虚拟机数量VM current,状态下调驻留时间left,需要调整的虚拟机数量VM adjust。由随机森林可求得并发量预测值X P,基于X P得到满足并发量要求的系统资源需求量C demand,求得资源可用量及需求量后,设置服务最小的操作模式虚拟机数量VM min以防止由于并发量持续降低后突然激增,导致系统工作虚拟机数量无法从连续降低或长期保持较低值的状态瞬时切换到较高值,而引发QoS降级;设置ε即用户对请求响应时间的要求,ε越大,VM adjust越小,系统中总的VM数越少,其对应请求响应时间变长,能耗将会增加。
一种支持性能保障的操作模式虚拟机数量评估方法,具体包括如下步骤:
步骤1:初始化参数,包括:用户对请求响应时间的要求ε,虚拟机最少数量VM min
步骤2:求出资源可用量C available和资源需求量C demand,公式如下:
C demand=X p*C 0               (1)
其中,C 0为单个请求的资源需求量,X P是由随机森林求得的并发量预测值;
步骤3:求出需要调整的虚拟机数量VM adjust,公式如下:
Figure PCTCN2019090867-appb-000001
其中,ε反映的则是用户对请求响应时间的要求:ε越大,VM adjust越小,系统中总的虚拟机数越少,其对应请求响应时间变长,反之亦然,q为负载。
步骤4:根据虚拟机当前数量VM current、虚拟机调整数量VM adjust的和与虚拟机的最少数量VM min之间的大小关系,调整操作模式虚拟机数量来应对服务突发并发量;
步骤4.1:若虚拟机当前的数量VM current与虚拟机调整数量VM adjust的和大于虚拟机的最少数量VM min,即VM current+VM adjust>VM min,则转到步骤4.2,否则,转到步骤4.3;
步骤4.2:判断虚拟机调整的数量VM adjust是否小于0,若虚拟机调整的数量VM adjust小于0,则转到步骤4.4;若虚拟机调整的数量VM adjust大于等于0,则VM demand=VM adjust+VM current,状态驻留时间left为最大状态驻留时间maxtime;
步骤4.3:判断状态驻留时间left是否等于非零值,若状态驻留时间left等于非零值,则VM demand=+VM current,状态驻留时间left=left-1;若状态驻留时间left等于零,则VM demand=VM min,状态驻留时间left为最大状态驻留时间maxtime;
步骤4.4:判断状态驻留时间left是否等于零,若状态驻留时间left为零,则VM demand=VM adjust+VM current;若状态驻留时间left不为零,则VM demand=VM current,状态驻留时间left为left=left-1;
有益技术效果:
本发明采用支持性能保障的操作模式虚拟机数量评估方法,为了解决高并发等因素造成系统对操作模式数量需求产生动态的变化的影响,及时作出资源调整,防止出现性能过剩或不足,影响服务质量。由于并发量预测方面的研究比较成熟,引用经典的并发量预测方法得到并发量预测结果。然后根据并发量预测结果调整操作模式虚拟机数量以满足用户需求,当性能不足时,需要从热备份转移一定数量的虚拟机到操作模式。但从热备份转移虚拟机会使可靠性减低。当性能过高时,减少一部分操作模式虚拟机会使系统可靠性增加并减低能耗。确定操作模式转移数量不仅会对性能造成影响还会影响可靠性和能耗,所以对操作模式虚拟 机数量动态调整很重要。制定相应操作模式虚拟机资源调整策略,建立支持性能保障的操作模式数量调整方法。
附图说明
图1为本发明实施例的一种支持性能保障的操作模式虚拟机数量评估方法流程图。
具体实施方式
下面结合附图和具体实施实例对发明做进一步说明,一种支持性能保障的操作模式虚拟机数量评估方法,;
一种支持性能保障的操作模式虚拟机数量评估方法,如图1所示,具体包括如下步骤:
步骤1:初始化参数,包括:用户对请求响应时间的要求ε设置为1,虚拟机最少数量VM min为1;
步骤2:求出资源可用量C available和资源需求量C demand,公式如下:
C demand=X p*C 0              (1)
其中,C 0为单个请求的资源需求量,X P是由随机森林求得的并发量预测值;
步骤3:求出需要调整的虚拟机数量VM adjust,公式如下:
Figure PCTCN2019090867-appb-000002
其中,ε反映的则是用户对请求响应时间的要求:ε越大,VM adjust越小,系统中总的虚拟机数越少,其对应请求响应时间变长,反之亦然,q为负载。
步骤4:根据虚拟机当前数量VM current、虚拟机调整数量VM adjust的和与虚拟机的最少数量VM min之间的大小关系,调整操作模式虚拟机数量来应对服务突发并发量;
步骤4.1:若虚拟机当前的数量VM current与虚拟机调整数量VM adjust的和大于虚拟机的最少数量VM min,即VM current+VM adjust>VM min,则转到步骤4.2,否则,转到步骤4.3;
步骤4.2:判断虚拟机调整的数量VM adjust是否小于0,若虚拟机调整的数量VM adjust小于0,则转到步骤4.4;若虚拟机调整的数量VM adjust大于等于0,则VM demand=VM adjust+VM current,状态驻留时间left为最大状态驻留时间maxtime;
步骤4.3:判断状态驻留时间left是否等于非零值,若状态驻留时间left等于非零值,则VM demand=+VM current,状态驻留时间left=left-1;若状态驻留时间left等于零,则VM demand=VM min, 状态驻留时间left为最大状态驻留时间maxtime;
步骤4.4:判断状态驻留时间left是否等于零,若状态驻留时间left为零,则VM demand=VM adjust+VM current;若状态驻留时间left不为零,则VM demand=VM current,状态驻留时间left为left=left-1。

Claims (2)

  1. 一种支持性能保障的操作模式虚拟机数量评估方法,其特征在于,具体步骤如下:
    步骤1:初始化参数,包括:用户对请求响应时间的要求ε,虚拟机最少数量VM min
    步骤2:求出资源可用量C available和资源需求量C demand,公式如下:
    C demand=X p*C 0       (1)
    其中,C 0为单个请求的资源需求量,X P是由随机森林求得的并发量预测值;
    步骤3:求出需要调整的虚拟机数量VM adjust,公式如下:
    Figure PCTCN2019090867-appb-100001
    其中,ε反映的则是用户对请求响应时间的要求;
    步骤4:根据虚拟机当前数量VM current、虚拟机调整数量VM adjust的和与虚拟机的最少数量VM min之间的大小关系,调整操作模式虚拟机数量来应对服务突发并发量。
  2. 根据权利要求1所述支持性能保障的操作模式虚拟机数量评估方法,其特征在于,所述步骤4具体包括如下步骤:
    步骤4.1:若虚拟机当前的数量VM current与虚拟机调整数量VM adjust的和大于虚拟机的最少数量VM min,即VM current+VM adjust>VM min,则转到步骤4.2,否则,转到步骤4.3;
    步骤4.2:判断虚拟机调整的数量VM adjust是否小于0,若虚拟机调整的数量VM adjust小于0,则转到步骤4.4;若虚拟机调整的数量VM adjust大于等于0,则VM demand=VM adjust+VM current,状态驻留时间left为最大状态驻留时间maxtime;
    步骤4.3:判断状态驻留时间left是否等于非零值,若状态驻留时间left等于非零值,则VM demand=+VM current,状态驻留时间left=left-1;若状态驻留时间left等于零,则VM demand=VM min,状态驻留时间left为最大状态驻留时间maxtime;
    步骤4.4:判断状态驻留时间left是否等于零,若状态驻留时间left为零,则VM demand=VM adjust+VM current;若状态驻留时间left不为零,则VM demand=VM current,状态驻留时间left为left=left-1。
PCT/CN2019/090867 2019-05-31 2019-06-12 一种支持性能保障的操作模式虚拟机数量评估方法 WO2020237728A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910466883.8A CN110162406B (zh) 2019-05-31 2019-05-31 一种支持性能保障的操作模式虚拟机数量评估方法
CN201910466883.8 2019-05-31

Publications (1)

Publication Number Publication Date
WO2020237728A1 true WO2020237728A1 (zh) 2020-12-03

Family

ID=67630389

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/090867 WO2020237728A1 (zh) 2019-05-31 2019-06-12 一种支持性能保障的操作模式虚拟机数量评估方法

Country Status (2)

Country Link
CN (1) CN110162406B (zh)
WO (1) WO2020237728A1 (zh)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104407688A (zh) * 2014-10-29 2015-03-11 哈尔滨工业大学深圳研究生院 基于树回归的虚拟化云平台能耗测量方法及系统
US9971621B1 (en) * 2015-02-02 2018-05-15 Amazon Technologies, Inc. Hotpooling virtual machines
CN108182115A (zh) * 2017-12-28 2018-06-19 福州大学 一种云环境下的虚拟机负载均衡方法
CN108984269A (zh) * 2018-07-16 2018-12-11 中山大学 基于随机回归森林模型的容器资源供给方法及系统
CN109088747A (zh) * 2018-07-10 2018-12-25 郑州云海信息技术有限公司 云计算系统中资源的管理方法和装置
CN109791504A (zh) * 2016-09-21 2019-05-21 埃森哲环球解决方案有限公司 针对应用容器的动态资源配置

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102103518B (zh) * 2011-02-23 2013-11-13 运软网络科技(上海)有限公司 一种在虚拟化环境中管理资源的系统及其实现方法
US8732291B2 (en) * 2012-01-13 2014-05-20 Accenture Global Services Limited Performance interference model for managing consolidated workloads in QOS-aware clouds
CN104283946B (zh) * 2014-09-26 2017-08-25 东北大学 一种单物理机下多虚拟机的资源自适应调整系统及方法
CN107404523A (zh) * 2017-07-21 2017-11-28 中国石油大学(华东) 云平台自适应资源调度系统和方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104407688A (zh) * 2014-10-29 2015-03-11 哈尔滨工业大学深圳研究生院 基于树回归的虚拟化云平台能耗测量方法及系统
US9971621B1 (en) * 2015-02-02 2018-05-15 Amazon Technologies, Inc. Hotpooling virtual machines
CN109791504A (zh) * 2016-09-21 2019-05-21 埃森哲环球解决方案有限公司 针对应用容器的动态资源配置
CN108182115A (zh) * 2017-12-28 2018-06-19 福州大学 一种云环境下的虚拟机负载均衡方法
CN109088747A (zh) * 2018-07-10 2018-12-25 郑州云海信息技术有限公司 云计算系统中资源的管理方法和装置
CN108984269A (zh) * 2018-07-16 2018-12-11 中山大学 基于随机回归森林模型的容器资源供给方法及系统

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
孟煜 等 (MENG, YU ET AL.): "云计算环境下云服务用户并发量的区间预测模型 (Prediction Interval Estimation Model of User Concurrent Requests for Cloud Service in Cloud Environment)", 计算机学报 (CHINESE JOURNAL OF COMPUTERS), vol. 40, no. 2, 28 February 2017 (2017-02-28), XP055750683, ISSN: 0254-4164, DOI: 20200224150233A *

Also Published As

Publication number Publication date
CN110162406B (zh) 2023-01-17
CN110162406A (zh) 2019-08-23

Similar Documents

Publication Publication Date Title
US9218046B2 (en) Methods and apparatuses for controlling thread contention
US8949431B2 (en) Method for live migration of virtual machines
CN102929720B (zh) 一种节能作业调度系统
CN108664116B (zh) 网络功能虚拟化的自适应省电方法、装置及cpu控制器
US20130080809A1 (en) Server system and power managing method thereof
CN108983946B (zh) 一种服务器功耗控制方法、系统及设备
US9652027B2 (en) Thread scheduling based on performance state and idle state of processing units
TW201821981A (zh) 動態調整多核心處理器的任務負載配置的裝置以及方法
CN112799828A (zh) 一种基于强化学习的云平台资源调度策略
WO2017133192A1 (zh) 一种业务控制方法以及业务控制装置
CN108228356B (zh) 一种流数据的分布式动态处理方法
WO2014110743A1 (zh) 一种数据中心间资源调度方法和设备
CN115878260A (zh) 一种低碳自适应云主机任务调度系统
CN109951317B (zh) 一种基于用户驱动的流行度感知模型的缓存替换方法
WO2020237728A1 (zh) 一种支持性能保障的操作模式虚拟机数量评估方法
CN112306696B (zh) 一种节能高效的边缘计算任务卸载方法和系统
WO2021057023A1 (zh) 一种基于部件温度自动分配计算资源的降功耗方法和系统
CN109144664B (zh) 一种基于用户服务质量需求差异的虚拟机动态迁移方法
CN105430083B (zh) 一种云平台部署方法
CN105700951B (zh) 一种实现cpu业务迁移的方法及装置
CN108874517B (zh) 固定优先级待机备用系统利用率划分能耗优化方法
Fan et al. GreenSleep: a multi-sleep modes based scheduling of servers for cloud data center
CN105933702A (zh) 一种基于任务敏感的功耗控制方法
CN112214295B (zh) 一种多cpu/gpu异构服务器集群的低能耗作业调度方法
WO2020237727A1 (zh) 一种支持可靠性保障的冷热操模式虚拟机数量评估方法

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: 19931005

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: 19931005

Country of ref document: EP

Kind code of ref document: A1