透過軟硬體協同設計的記憶體虛擬化之優化方式

虛擬化技術可將數個作業系統合併於一台實體機器上同時執行。它源自於過往利用數個單一使用者作業系統，建立一個多使用者分時作業系統。存在已久的虛擬化技術也與時俱進。除了在伺服器端的應用外，隨著嵌入式處理器能力的進步，虛擬化技術被更加廣泛地應用在不同類型的系統。由於這些新興應用的多元性，我們需要發展新的設計方法，以便將這些系統的特性一併納入考量。在這篇論文中，我們提出利用軟硬體協同設計來優化虛擬化環境的想法，並以優化記憶體虛擬化為例，展示軟硬體協同設計對於優化虛擬化環境的潛能。基於過往的研究，我們認為記憶體子系統是虛擬化環境中主要的效能瓶頸之一。因此我們以特定類別的虛擬化環境作為範例，專注針對其環境的記憶體虛擬化之優化。首先，我們以定量分析的方法，評估記憶體虛擬化的影響。隨後我們提出一個記憶體虛擬化的優化方式，並以正式證明進行定性分析；同時結合軟體擬真與硬體模擬，進行定量分析。實驗結果指出我們提出的方法效能勝過現有的方法。這個研究也指出運用軟硬體協同設計進行優化，在新興應用的虛擬化環境中深具潛力。

關鍵字

虛擬化；記憶體管理；轉譯後備緩衝區；快取記憶體；軟/硬體介面

並列摘要

Virtualization is a technology enabling consolidation of multiple operating systems into a single physical machine. It originated from the need to create a multi-user time-sharing operating system based on multiple single-user operating systems. This long-lasting technology has evolved constantly. In addition to the popular applications for server-side virtualization, the advances of the capabilities of embedded processors make virtualization available on various systems much wider than before. The diversity of the target systems demands new design approaches considering the characteristics of the systems. In this dissertation, we propose the idea of optimizing virtualization environments through hardware/software co-design, and demonstrate the potential power of hardware/software co-design through the development of a new optimization technique for memory virtualization. Based on the existing studies, we recognize the memory subsystem as a major bottleneck of a virtualization environment. Therefore, we concentrate our efforts on optimizing memory virtualization for a specific type of virtualization environments as a working example. We first present a quantitative analysis of the impacts of memory virtualization. We then propose an optimized memory virtualization technique along with a comprehensive evaluation including the qualitative analysis with a formal proof and the quantitative analysis based on software emulation and hardware simulation. The results suggest the proposed technique outperforms the existing technique. The research points out hardware/software co-design is a promising direction for optimizing virtualization for the emerging applications.

並列關鍵字

virtualization ； memory management ； translation lookaside buffers ； cache memories ； hardware/software interfaces

參考文獻

[1] R. A. Meyer and L. H. Seawright, “A virtual machine time-sharing system,” IBM Syst. J., vol. 9, no. 3, pp. 199–218, Sep. 1970.

[3] M. Pearce, S. Zeadally, and R. Hunt, “Virtualization: Issues, security threats, and solutions,” ACM Comput. Surv., vol. 45, no. 2, pp. 17:1–17:39, Mar. 2013.

[4] J. Daniels, “Server virtualization architecture and implementation,” Crossroads, vol. 16, no. 1, pp. 8–12, Sep. 2009.

[5] W. Vogels, “Beyond server consolidation,” Queue, vol. 6, no. 1, pp. 20–26, Jan. 2008.

[7] J. Shuja, A. Gani, K. Bilal, A. U. R. Khan, S. A. Madani, S. U. Khan, and A. Y. Zomaya, “A survey of mobile device virtualization: Taxonomy and state of the art,” ACM Comput. Surv., vol. 49, no. 1, pp. 1:1–1:36, Apr. 2016.

國際替代計量

透過軟硬體協同設計的記憶體虛擬化之優化方式

全文下載

主題瀏覽