| 本文已被：浏览 3949次 下载 9071次 |
码上扫一扫！ |
|
|
| 支持随机服务请求的云虚拟机按需物理资源分配方法 |
|
曹洁1,2,3, 曾国荪1,3, 匡桂娟1,3, 张建伟2, 马海英4, 胡克坤1,3, 钮俊5
|
|
1.同济大学 计算机科学与技术系, 上海 200092;2.郑州轻工业学院 软件学院, 河南 郑州 450002;3.国家高性能计算机工程技术中心同济分中心, 上海 200092;4.南通大学 计算机科学与技术学院, 江苏 南通 226019;5.宁波大学 计算机科学与技术系, 浙江 宁波 315211
|
|
| 摘要: |
| 针对云平台按负载峰值需求配置处理机资源、提供单一的服务应用和资源需求动态变化导致资源利用率低下的问题，采用云虚拟机中心来同时提供多种服务应用.利用灰色波形预测算法对未来时间段内到达虚拟机的服务请求量进行预测，给出兼顾资源需求和服务优先等级的虚拟机服务效用函数，以最大化物理机的服务效用值为目标，为物理机内的各虚拟机动态配置物理资源.通过同类虚拟机间的全局负载均衡和多次物理机内各虚拟机的物理资源再分配，进一步增加服务请求量较大的相应类型的虚拟机的物理资源分配量.最后，给出了虚拟机中心基于灰色波形预测的按需资源分配算法ODRGWF.模拟实验结果表明，该算法能够有效地提高云平台中处理机的资源利用率，对提高用户请求完成率以及服务质量都具有实际意义. |
| 关键词: 云计算 虚拟化 随机服务请求 灰色波形预测 按需资源分配 |
| DOI：10.13328/j.cnki.jos.005054 |
| 分类号: |
| 基金项目:国家高技术研究发展计划（863）（2009AA012201）；国家自然科学基金（61402244）；上海市优秀学科带头人计划（10X D1404400）；华为创新研究计划（IRP-2013-12-03）；高效能服务器和存储技术国家重点实验室开放基金（2014 HSSA10）；河南省科技创新人才计划（[2015]4）；浙江省公益技术应用研究项目（2014C31059） |
|
| On-Demand Physical Resource Allocation Method for Cloud Virtual Machine to Support Random Service Requests |
|
CAO Jie1,2,3, ZENG Guo-Sun1,3, KUANG Gui-Juan1,3, ZHANG Jian-Wei2, MA Hai-Ying4, HU Ke-Kun1,3, NIU Jun5
|
|
1.Department of Computer Science and Technology, Tongji University, Shanghai 200092, China;2.Software College, Zhengzhou Institute of Light Industry, Zhengzhou 450002, China;3.Tongji Branch, National Engineering and Technology Center of High Performance Computer, Shanghai 200092, China;4.School of Computer Science and Technology, Nantong University, Nantong 226019, China;5.Department of Computer Science and Technology, Ningbo University, Ningbo 315211, China
|
| Abstract: |
| Low resource utilization is becoming much more serious in cloud platform which allocates processor resources according to the peak load while providing single service application and facing dynamic variation of resource demand. To address the problem, this study uses cloud virtual machine (VM) center to provide a variety of reasonable service applications simultaneously. Gray wave forecasting algorithm is adopted to predict the future load of service requests and a VM service utility function is proposed by taking resource requirements and service priorities into account. Each VM inside a physical machine dynamically configures physical resources to maximize the service utility value of the physical machine. Besides, by applying the global load balancing and multi-time physical resource redistribution for each virtual machine in the same physical machine, the number of physical resources assigned to the VMs whose service request amount is much larger is further increased. In the end, on-demand resource reconfiguration algorithm ODRGWF based on grey wave forecasting is put forward. The simulation results show that the proposed algorithm can effectively improve processor resource utilization, which is of practical significance to improve user request completion rate and service quality. |
| Key words: cloud computing virtualization random service request grey wave forecasting on-demand resource allocation |
