雲端運算中以可拓理論為基礎之虛擬機器資源動態調整機制

摘要在雲端運算軟、硬體架構中，最下層的IaaS (Infrastructure as a Service)架構，是提供運算、儲存及網路等資源的雲端基礎架構。為了整合資源及可彈性調整資源之運用，且可服務大量不同的需求，可藉助虛擬化技術來完成整合，以有效地管理及應用資源。虛擬化技術是將基礎設施封裝成檔案型式，提供電腦運算資源，一個虛擬化的平台上，執行不同的應用程式運算，可依運算需求去彈性調整資源，而如何適時去調整資源達到最好的效率，則為一重要的研究課題。本論文提出一執行於雲端運算環境中，以可拓理論為基礎之動態虛擬機資源調整機制，在KVM(Kernel-based Virtual Machine)全虛擬化架構下，利用可拓理論建立關係函數，據以判斷CPU需求的程度，進而調整虛擬機上提供的CPU數，即時增加或減少CPU數量。實驗結果顯示，所提出之機制可以有效地改善整體系統運算能力及資源分配。

關鍵字

虛擬機器；動態資源調整；虛擬化；可拓理論

並列摘要

Abstract In the cloud computing environment, the basic layer of IaaS (Infrastructure as a Service) is responsible for supporting computation, storage and internet access. In order to integrate the resources, utilize the resources flexibly, and to meet the demand of large scale data flow management and applications, a positive solution is applying the virtualization technology to reach this achievement. The virtualization technology means to pack a basic device or equipment into the file format like a packet and featuring as a workable and calculable source for computer operation to complete the task of integration. Various applications performed on a virtual platform, the required resources for individual application can be adjusted flexibly and effectively. How to adjust all resources dynamically for achieving the best performance is an important issue to develop and analyze. In this thesis, an extenics-based dynamic resource adjustment mechanism for supporting cloud computing is proposed. Based on a KVM (Kernel-based Virtual Machine) technology, a relation function has been established by applying the extenics theory to determine the degree of CPU requirement for each application. According to the degree of CPU requirement, the proposed resource adjustment mechanism can provide a proper number of CPUs for each virtual machine dynamically. The experiment results show that the proposed mechanism can improve the performance of the system and distribute the source information in a real time with efficiency.