雲端運算中動態調整虛擬機器運算資源機制

DYNAMIC ADJUSTMENT MECHANISM OF THE VIRTUAL MACHINE COMPUTING RESOURCE IN THE CLOUD COMPUTING

胡志凱(Chih-kai Hu)

指導教授 ： 李良德

大同大學/電機資訊學院/資訊工程學系所/碩士(2010年)

摘要

在雲端運算IaaS (Infrastructure as a Service)架構下，是以提供基礎設施CPU、記憶體、網路等之服務，而虛擬化技術所扮演的角色，是將基礎設施封裝成檔案型式，以提供電腦運算資源，此通常是一個虛擬化環境的平台如：KVM、XEN、VMware。在這樣的平台上，依據應用程式運算資源需求的不同，其動態資源調整的自動化、運算資源的分配機制等，都是尚待解決的議題。
本論文提出一運行於雲端運算環境中，可動態調整虛擬機器運算資源的機制，稱之為DAVMCR (Dynamic Ddjustment of the Virtual Machine Computing Resources)，這是建構在KVM(Kernel-based Virtual Machine)全虛擬化架構下的動態調整技術。此設計架構為虛擬機器在CPU高負載情況下，以不中斷的服務方式，即時增加CPU數量以提高運算效能，而在高負載狀況解除後能即時回收CPU資源，以增加整體系統資源之利用度。
實驗結果顯示，透過本研究所提出之機制，在虛擬機器負載高時，可以有效提升虛擬機器的運算效能及縮短連線請求的反應時間。

關鍵字

虛擬機器 ； 動態資源調整 ； 雲端計算 ； 虛擬化

並列摘要

IaaS (Infrastructure as a Service) architecture in the cloud computing is to provide infrastructure, CPU, memory, and network, etc., for the service. The virtualization is to package a file type of infrastructure, provision of computing resources. This is usually a virtual environment platform, such as KVM, XEN, and VMware. In such a platform, based on the application of different computing resource requirements, dynamic resource adjustment in part, there are still many research issues to be resolved. For example: automated dynamic resource adjustment mechanism of calculating the allocation of resources.
In this thesis, a dynamic adjustment mechanism of the virtual machine computing resources, called DAVMCR, in the cloud computing environment has been proposed. It is a dynamic adjustment technique constructed under the KVM (Kernel-based Virtual Machine) virtualization architecture. By applying the proposed mechanism, when the loading of a virtual machine is heavy, the number of CPUs in the virtual machine will be increased automatically with non-stop service to increase the system performance. Once the system loading is no longer heavy, the augmented CPUs will be released immediately to increase the overall system resource utilization. Experimental results show that by applying the proposed mechanism in a heavy loaded virtual machine, the computing performance of the virtual machine, can be improved effectively and the response time of connection requests can also be reduced significantly.

並列關鍵字

dynamic resource adjustment ； Virtual machines ； Virtualization ； cloud computing

參考文獻

[5] Evangelia Kalyvianaki, Themistoklis Charalambous, Steven Hand, “Self-adaptive and self-configured CPU resource provisioning for virtualized servers using Kalman filters,” Proceedings of the 6th International Conference on Autonomic computing.

[7] “Cloud computing,” available at

[8] Liang-Jie Zhang, Qun Zhou, “CCOA: Cloud Computing Open Architecture,” IEEE DOI 10.1109/ICWS, 2009.

[10] Ashok Raj, “CPU hot-plug,” available at

[11] “RedHat virtual machine manager,” available at