雲端運算平台上因應動態工作負載變動之服務部屬及排程方法研究

Dynamic service deployment and scheduling for time-varying workload on cloud computing platform

雲端運算和軟體即服務等新的運算模式，已經對人們開發及使用軟體的方式帶來了很大的改變，而且也帶來許多新的議題及挑戰。文獻中已有許多對於雲端服務的相關研究和討論，而本篇論文就針對其中動態資源供應下之服務部署與服務請求排程等問題進行研究。因應雲端平台上，隨著時間不同而有所變化的服務請求與工作負載特性，我們提出了三種能夠因應動態負載需求的服務部署方法。另一方面，為進一步提升複合式軟體即服務的執行效能，我們也提出了一個基於工作序列機制的新服務請求排程方法。我們所提出的方法能配合任何動態資源供應的方法一起運作，以根據雲端環境中動態的工作負載變化來及時調整服務部署的安排，以提升複合式軟體及服務的運行效能。我們藉由一系列模擬實驗來評估所提方法的效能，並與文獻中先前的方法進行比較。實驗結果顯示，相較於過往的方法，我們所提出的新方法更有效率，能夠大幅地降低服務的平均反應時間，顯著提升複合式軟體即服務應用的執行效能。

Cloud computing and the SaaS model have brought an innovative way for people to develop and use software nowadays. However, the innovation also raises a lot of new research issues. In this thesis, we investigate two such important issues, service deployment and service request scheduling, for composite SaaS applications in dynamic cloud environments. We present three load-aware service deployment approaches for time-varying workload, and one service request scheduling method based on task ranking mechanisms. The proposed approaches can collaborate with dynamic resource provisioning methods to adaptively deploy services according to the dynamically changing requests for composite SaaS applications, facilitating their efficient execution. A wide range of simulation experiments have been performed to evaluate the proposed approaches. The experimental results indicate that our approaches surpass the previous methods in the literature in terms of average service response time.