融合双速率和工作休眠的虚拟机调度策略及参数优化

doi:10.11959/j.issn.1000-436x.2017298

通信学报 ›› 2017, Vol. 38 ›› Issue (12): 10-20.doi: 10.11959/j.issn.1000-436x.2017298

融合双速率和工作休眠的虚拟机调度策略及参数优化

金顺福^1,^2,³,郝闪闪^1,^2,³,王宝帅^1,^2,³

¹ 燕山大学信息科学与工程学院，河北秦皇岛 066004
² 河北省计算机虚拟技术与系统集成重点实验室，河北秦皇岛 066004
³ 河北省软件工程重点实验室,河北秦皇岛 066004

修回日期:2017-11-23 出版日期:2017-12-01 发布日期:2018-01-19
作者简介:金顺福（1966-），女，朝鲜族，内蒙古满洲里人，燕山大学教授、博士生导师，主要研究方向为网络资源分配与优化、排队论研究与应用等。|郝闪闪（1993-），女，河北邢台人，燕山大学硕士生，主要研究方向为云数据中心的节能策略与性能优化。|王宝帅（1993-），男，河北保定人，燕山大学硕士生，主要研究方向为网络资源分配与优化。
基金资助:
国家自然科学基金资助项目(61472342);河北省自然科学基金资助项目(F2017203141)

Virtual machine scheduling strategy based on dual-speed and work vacation mode and its parameter optimization

Shun-fu JIN^1,^2,³,Shan-shan HAO^1,^2,³,Bao-shuai WANG^1,^2,³

¹ School of Information Science and Engineering,Yanshan University,Qinhuangdao 066004,China
² Key Laboratory for Computer Virtual Technology and System Integration of Hebei Province,Qinhuangdao 066004,China
³ Key Laboratory for Software Engineering of Hebei Province,Qinhuangdao 066004,China

Revised:2017-11-23 Online:2017-12-01 Published:2018-01-19
Supported by:
The National Natural Science Foundation of China(61472342);The Natural Science Foundation of Hebei Province(F2017203141)

摘要/Abstract

摘要：

受日趋严格的环境标准的影响，高能耗和高污染成为制约云数据中心发展的重要因素。在保证云数据中心服务质量的前提下，应用动态电压调整技术，引入同步多重工作休眠模式，提出一种新型的虚拟机调度策略。通过构建具有自适应服务率和多重工作休假的二维连续时间马尔可夫（Markov）随机模型，运用矩阵几何解方法，从系统节能水平和用户请求平均延迟等方面评估虚拟机调度策略的性能。综合理论分析结果和仿真统计结果，验证虚拟机调度策略的有效性。从经济学角度出发，建立系统效用函数，改进萤火虫智能优化算法，给出工作休眠参数的优化方案，实现系统响应性能和节能效果之间的合理平衡。

关键词: 云数据中心, 虚拟机调度策略, 双速率, 多重工作休假排队, 效用函数, 智能优化

Abstract:

Due to the increasingly strict environmental standards,high pollution and high energy consumption have become the significant factors restricting the development of cloud data centers (CDC).Under the premise of guaranteeing the quality of service (QoS) of CDC,dynamic power management (DPM) technology was applied,synchronous multiple working sleep mode was introduced,and a novel virtual machine (VM) scheduling strategy was proposed.By establishing a two-dimensional continuous-time Markov stochastic model with adaptive service rate and synchronous multiple work vacations,and using the method of a matrix geometric solution,the performance of the VM scheduling strategy was evaluated in terms of energy saving level and average delay of requests.Numerical results with analysis and simulation verify the energy saving effectiveness of the VM scheduling strategy.In order to achieve a reasonable balance between the response performance and the energy-saving effect,a system utility function was constructed from the perspective of economics and design a researching algorithm of the sleep parameter based on the firefly algorithm(FA).

Key words: cloud data centers, virtual machine scheduling strategy, dual-speed, multiple working vacations queue, utility function, intelligent optimization

中图分类号:

TP393

金顺福,郝闪闪,王宝帅. 融合双速率和工作休眠的虚拟机调度策略及参数优化[J]. 通信学报, 2017, 38(12): 10-20.

Shun-fu JIN,Shan-shan HAO,Bao-shuai WANG. Virtual machine scheduling strategy based on dual-speed and work vacation mode and its parameter optimization[J]. Journal on Communications, 2017, 38(12): 10-20.

图/表 8

图1

表1

图2

表2

图3

图4

图5

表3

参考文献 16

[1]	JIN X , ZHANG F , VASILAKOS A V ,et al. Green data centers:a survey,perspectives,and future directions[J]. arXiv:1608.00687.
[2]	DUAN L , ZHAN D , HOHNERLEIN J . Optimizing cloud data center energy efficiency via dynamic prediction of CPU idle intervals[C]// IEEE International Conference on Cloud Computing. 2015: 985-988.
[3]	LI K . Improving multicore server performance and reducing energy consumption by workload dependent dynamic power management[J]. IEEE Transactions on Cloud Computing, 2016,4(2): 122-137.
[4]	WANG Y , XIE Q , AMMARI A ,et al. Deriving a near-optimal power management policy using model-free reinforcement learning and Bayesian classification[C]// IEEE Design Automation Conference. 2011: 41-46.
[5]	GAO Y , GUAN H , QI Z ,et al. Quality of service aware power management for virtualized data centers[J]. Journal of Systems Architecture, 2013,59(5): 245-259.
[6]	LIAO D , LI K , SUN G ,et al. Energy and performance management in large data centers:a queuing theory perspective[C]// International Conference on Computing,Networking and Communications. 2015: 287-291.
[7]	SHEN D , LUO J , DONG F ,et al. Stochastic modeling of dynamic right-sizing for energy-efficiency in cloud data centers[J]. Future Generation Computer System, 2015,48(C): 82-95.
[8]	马晓彤, 金顺福, 刘建平 ,等. 认识无线电网络中的基站节能策略及纳什均衡研究[J]. 通信学报, 2017,37(7): 172-181.
	MA X T , JIN S F , LIU J P ,et al. Study on energy saving strategy and Nash equilibrium of base station in cognitive radio network[J]. Journal on Communications, 2017,37(7): 172-181.
[9]	CHOU C H , WONG D , BHUYAN L N . DynSleep:fine-grained power management for a latency-critical data center application[C]// The 2016 International Symposium on Low Power Electronics and Design. 2016: 212-217.
[10]	CHEN Y L , CHANG M F , LIANG W Y ,et al. Performance and energy efficient dynamic voltage and frequency scaling scheme for multicore embedded system[C]// IEEE International Conference on Consumer Electronics. 2016: 58-59.
[11]	TIAN N , ZHANG Z G , Vacation queueing models theory and applications[M]. Springer US, 2006.
[12]	LATOUCHE G , RAMASWAMI V . Introduction to matrix analytic methods in stochastic modeling[M]// Society for Industrial and Applied Mathematics. 2001.
[13]	GREENBAUM A , . Iterative methods for solving linear systems[M]// Society for Industrial and Applied Mathematics. 2001.
[14]	陈赓, 夏玮玮, 沈连丰 . 基于传输速率自适应的动态带宽分配算法[J]. 通信学报, 2014,35(5): 25-32.
	CHEN G , XIA W W , SHEN L F . Dynamic bandwidth allocation algorithm based on transmission rate adaptation[J]. Journal on Communications, 2014,35(5): 25-32.
[15]	GANDOMI A H , YANG X S , ALAVI A H . Mixed variable structural optimization using firefly algorithm[J]. Computers＆Structures, 2011,89(24): 2325-2336.
[16]	YU S , ZHU S , MA Y ,et al. Enhancing firefly algorithm using generalized opposition-based learning[J]. Computing, 2015,97(7): 741-754.

系统水平	初始的虚拟机状态	一步转移后的虚拟机状态
[0,θ₂-1]	状态Ⅰ	状态Ⅰ
θ₂	状态Ⅰ	状态Ⅰ
θ₂	状态Ⅱ	状态Ⅰ或状态Ⅱ
[θ₂+1，θ₁-1]	状态Ⅰ	状态Ⅰ
[θ₂+1，θ₁-1]	状态Ⅱ	状态Ⅱ
θ₁	状态Ⅰ	状态Ⅰ或状态Ⅱ
θ₁	状态Ⅱ	状态Ⅱ
[θ₁+1,n]	状态Ⅱ	状态Ⅱ
n+1	状态Ⅱ	状态Ⅱ
n+1	状态Ⅲ	状态Ⅱ或状态Ⅲ
[n+2,n+m+θ₃]	状态Ⅱ	状态Ⅱ
[n+2,n+m+θ₃]	状态Ⅲ	状态Ⅲ
(n+m+θ₃+1,∞）	状态Ⅱ	状态Ⅱ或状态Ⅲ
(n+m+θ₃+1,∞）	状态Ⅲ	状态Ⅲ

参数	值
主模块中虚拟机的数量n	30
备用模块中虚拟机的数量m	20
双阈值θ₁和θ₂	20,10
用户请求的到达率λ	15
低速运行服务率μ_l	0.2
高速运行服务率μ_h	0.4
能耗因子k	10

μ_s	θ₃	δ^*	F(δ^*)
	20	0.022 7	41.988 3
0.10	30	0.027 2	42.356 8
0.10	40	0.033 3	42.671 5
	50	0.042 2	42.931 8
	20	0.019 5	42.248 3
0.11	30	0.023 4	42.616 4
0.11	40	0.028 9	42.934 2
	50	0.037 0	43.198 8
	20	0.016 1	42.547 4
0.12	30	0.019 4	42.908 2
	40	0.024 1	43.224 6
	50	0.031 2	43.491 4

融合双速率和工作休眠的虚拟机调度策略及参数优化

Virtual machine scheduling strategy based on dual-speed and work vacation mode and its parameter optimization

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 16

相关文章 15

Metrics

推荐阅读 0

[1]	刘雷, 陈晨, 冯杰, 裴庆祺, 何辞, 窦志斌. 车载边缘计算中任务卸载和服务缓存的联合智能优化[J]. 通信学报, 2021, 42(1): 18-26.
[2]	王耀民,王霞,董易,高莲,张松海,施心陵. 面向云数据中心的多业务差异化流量管理优化策略[J]. 通信学报, 2019, 40(11): 45-56.
[3]	吴昊,范九伦,赖成喆,刘建华. 基于攻防博弈和蒙特卡洛模拟的网站防御策略选取方法[J]. 通信学报, 2018, 39(8): 48-55.
[4]	陈雷,甘士忠,张立毅,王光艳. 基于样条插值与人工蜂群优化的非线性盲源分离算法[J]. 通信学报, 2017, 38(7): 36-46.
[5]	马晓彤,金顺福,刘建平,霍占强. 认知无线电网络中的基站节能策略及纳什均衡研究[J]. 通信学报, 2016, 37(7): 172-181.
[6]	王大鸣,陈松,崔维嘉,王强. 多用户MIMO-OFDM系统基于QoE效用函数的跨层资源分配[J]. 通信学报, 2014, 35(9): 175-183.
[7]	陈赓,夏玮玮,沈连丰. 基于传输速率自适应的动态带宽分配算法[J]. 通信学报, 2014, 35(5): 25-32.
[8]	陈赓，夏玮玮，沈连丰. 基于传输速率自适应的动态带宽分配算法[J]. 通信学报, 2014, 35(5): 4-32.
[9]	李为，熊春林，王德刚，张晓瀛，魏急波. 基于多维离散粒子群优化的协同OFDMA系统跨层资源分配[J]. 通信学报, 2014, 35(4): 3-24.
[10]	李为,熊春林,王德刚,张晓瀛,魏急波. 基于多维离散粒子群优化的协同OFDMA系统跨层资源分配[J]. 通信学报, 2014, 35(4): 17-24.
[11]	邓寒冰1，张霞1,2，刘积仁1. 效用驱动的多agent合作机制[J]. 通信学报, 2013, 34(7): 14-133.
[12]	邓寒冰,张霞,刘积仁. 效用驱动的多agent合作机制[J]. 通信学报, 2013, 34(7): 124-133.
[13]	闫继垒,李建东,赵林靖,董全. 认知无线网络中兼顾效用与公平的联合带宽和功率分配算法[J]. 通信学报, 2013, 34(10): 59-64.
[14]	闫继垒，李建东，赵林靖，董全. 认知无线网络中兼顾效用与公平的联合带宽和功率分配算法[J]. 通信学报, 2013, 34(10): 7-64.
[15]	范文浩,刘元安,吴帆. 异构无线网络中多模终端多接入选择机制研究[J]. 通信学报, 2012, 33(7): 183-190.