雲端測試可以利用虛擬化技術將測試案例分散到多個虛擬測試機器執行測試,透過平行化的方式提升測試效率,並可利用混合雲之架構調用成本較低的運算資源以降低雲端測試整體成本。然而平行測試的時間和成本與使用虛擬機器數量多寡和租用成本有關,如何達到較佳的時間和成本效益比是一個值得關注的議題。本論文提出一個雲端測試時間與成本效益分析流程,將測試平行化,並探討不同因素,如虛擬機器數量、測試排程等,對於測試時間和成本的影響,且實作一個雲端平行測試平台進行驗證,並以實驗來觀察雲端平行測試的時間成本效益。實驗結果顯示,測試平行化可有效提升測試效率,然其加速比率並未隨虛擬機器數量增加而呈線性成長,透過測試時間與成本效益分析流程,可以有效協助使用者決定雲端平行測試所需之虛擬機器數量以滿足其測試需求,並達到較佳的測試時間與成本效益比。
Through using virtualization, cloud testing allows to improve test efficency by distributing test cases to different virtual machines and by performing testing concurrently. In addition, through using multi-cloud architecture, cloud testing can reduce the overall test cost by taking the advantages of lower-cost cloud computing resources. The required time and cost of cloud testing would depend on the number of virtual machines used and the rental costs of the virtual machines. Thus, it is very important to determine how many virtual machines should be used for cloud testing in order to achieve better testing cost-benefit. This thesis proposes a cost-benefit analysis process for cloud testing in which the testing is performed concurrently. The factors that could influence the time and cost of cloud testing, such as the number of virtual machines and the scheduling of testing jobs, are considered in the analysis process. A cloud testing platform is developed and several experiments are conducted to validate the effectiveness of the proposed process. The experimental results show that concurrent testing over cloud can signifiantly improve testing effiency although the speedup ratio does not increase linearly as the number of virtual machines increases. By using the propose process, users can have a better way to determine the number of virtual machines needed for satifying the test requirements while achiving perferable test cost-benefit ratio.