即時偵測與動態虛擬機器資源管理

在雲端運算中的IaaS服務概念下，伺服器提供資源（CPU、網路與儲存等資源）租借服務，雲端中的資源是由許虛擬化技術整合多台實體主機資源組成的虛擬資源，客戶可以依照自己的需求使用雲端環境中所提供的資源，而資源的需求通常是取決於系統中的工作量。然而在雲端運算環境裡，客戶對於資源的需求會因工作量時常改變，如果要重新配置客戶系統的資源，以傳統的靜態配置來看，勢必要將系統關閉才能配置，但是不能時常關閉客戶系統，如此效能反而變差。動態配置是基於虛擬技術下的一種概念，運作中的客戶系統可以直接動態調整資源，不須事先將客戶系統關閉。本論文針對CPU部分提出一套基於KVM虛擬環境下的動態CPU資源管理系統，當虛擬機器提出需要更多資源需求時，可以動態增加虛擬機器的VCPU，若主機資源已被配置完，則會以動態遷移的方式將需求資源的VM遷移到足夠配置資源的實體主機上，並且配置資源予該VM；而當虛擬機器的VCPU使用率低時，管理系統會減少虛擬機器的VPCU配置，以達到提高虛擬機器效能，並提高實體資源使用率。經實驗結果顯示，透過動態CPU資源管理系統適時地配置虛擬機器，可以有效提升虛擬機器的效能。

關鍵字

虛擬機器；動態資源管理；雲端運算

並列摘要

One of the important services in cloud computing is Infrastructure as a Service (IaaS). The provision model is the offering of compute power and storage space on demand of guest’s resource requirement. The resource requirement is depending on workload of task which will be changed frequently. However, if guest OS in IaaS model is start up, the resource configuration of the Guest system is difficult to change with traditional method of resource allocation. Guest OS must be rebooted to change the resource configuration. However, the performance of the system will be degraded when rebooting the guest OS frequently. Dynamic resource allocation is a method based on the virtualization technology, the resource configuration of the guest system can be changed without rebooting OS. In this thesis, a dynamic CPU resource allocation mechanism in KVM virtual environment has been proposed. When a virtual machine with high CPU utilization has been detected by the proposed mechanism, a VCPU will be provided to the virtual machine dynamically. Once there exists no more CPU can be provided, the requested virtual machine will be migrated to another virtual machine with enough resource for improving the system performance. In contrast, when a virtual machine with low CPU utilization has been detected, the VCPU of the virtual machine will be deducted. The experimental results show that by applying the proposed mechanism for managing CPU resource in the virtual machine dynamically can improve the performance of the virtual machine system effectively.

並列關鍵字

virtual machine ； dynamic resource management ； Cloud computing

參考文獻

[3] J. Brandt, F. Chen, V. De Sapio, A. Gentile, J. Mayo, P. Pebay, D. Roe, D. Thompson, and M. Wong, “ Combining Virtualization, Resource Characterization, and Resource Management to Enable Efficient High Performance Compute Platforms Through Intelligent Dynamic Resource Allocation, ” 2010 IEEE International Symposium on Parallel & Distributed Processing, Workshops and Phd Forum (IPDPSW), April. 2010, Pages 1-8

[4] J. Che, Q. He, Q. Gao, and D. Huang, ”Performance Measuring and Comparing of Virtual Machine Monitors,” IEEE/IFIP International Conference on Embedded and Ubiquitous Computin( EUC '08), Dec. 2008, Pages 381-386

[5] Q. Li, Q. Hao, L. Xiao, and Z. Li, “Adaptive Management of Virtualized Resources in Cloud Computing Using Feedback Control,” International Conference on Information Science and Engineering (ICISE), Dec. 2009, Pages 99-102

[6] Y. Li, W. Li, and C. Jiang, “A Survey of Virtual Machine System: Current Technology and Future Trends,” 2010 Third International Symposium on Electronic Commerce and Security, Jul. 2010, Pages 332-336

[1] B. Zhang, S. Gao, L. Xia, J. He, and K. Miao, “Resource management policy for cloud testbed of China railway,” International Conference on Computer Application and System Modeling (ICCASM), Oct. 2010, Pages v4-375 - v4-379

Google Scholar

國際替代計量

即時偵測與動態虛擬機器資源管理

未授權

主題瀏覽