Research On Resource Scheduling In Cloud Testing

With the rapid development of Internet technology, the functions of software get more and more abundant, the scale of software becomes bigger, the performance of software turns to be stronger and users’ requirements of software product quality are more than before. Cloud computing and automatic technologies are adopted by Cloud testing to overcome problems faced in the traditional software test. Users could get Cloud testing services on the basic of their demand, set up virtual and reliable test environment quickly and complete test with lower cost.Vm scheduling strategies in Cloud testing both influence test efficiency and energy consumption. Therefore, it’s really important to select suitable Vm scheduling strategy during Cloud testing. In the context of Cloud testing, this paper mainly study the related testing scheduling, and the major research works and innovative points in this paper are:(1) This paper analyzes automatic software test technology and Cloud computing adopted by Cloud testing. It discusses the scope and limitations of automatic software test and studies the service level and basic structure of Cloud computing. It also expounds the related technologies of Cloud testing and analyzes the advantages and challenges of Cloud testing.(2) This paper studies the scheduling process during Cloud testing. The common Vm scheduling strategies are divided into three types after a detailed analysis:center on saving energy, center on load balance and center on QoS performance. The common Vm scheduling algorithms are analyzed in detail and a simulator called CloudSim is also introduced which is applied to the resource scheduling of cloud data center.(3) This paper analyzes the dependency relationships between test tasks and proposes single dependence and multi dependencies. The multi dependencies are divided into three types:One-to-Many, Many-to-One and Many-to-Many. In consideration of ants’ communication through pheromone, it takes the dependency relationships as reference for ants and applyies ACO into Vm scheduling in Cloud testing. When compared with RR and GA, the simulation results demonstrate that ACO improves test efficiency and essures load balance of Vms as well.(4) This paper takes advantages of CloudStack, Selenium, TestNG and Struts2to develop automated Cloud testing platform based on IaaS. According to the software engineering’s development process and Cloud testing platform’s requirments, the funtional requirments are prestented. The overall structure, background realization and database are also designed respectively. Then, the realization and test process of overall platform are introduced in terms of customers detailedly.Finally, this paper gives a summary and an outlook about the future work.