The Research About Key Technologies On Performance Test Platform Of Android-based Mobile Application

As a product of the combination of mobile communication and Internet, mobile internet has brought infinite possibilities to human life. A variety of mobile applications facilitate people’s life, and people’s behaviors and habits are thus gradually changed. While relevant enterprises enjoy higher commercial values brought by mobile internet, they are also involved in fiercer competition. More and more customers start using and paying their attention to mobile applications. Usually, there are millions of users accessing one application, which requires for higher performance to the application. The mobile internet environment has posed new challenges to traditional software testing method because of its characteristics, such as the diversity of mobile terminals, the flexibility of access, the great number of users, and the complexity of application structure. By studying these characteristics of mobile performance test, this thesis designed and developed a performance test automation platform of android mobile application. The platform is aiming at helping testers to do performance test better under the mobile internet by providing perfect automatic testing process.The mobile performance test automation platform is mainly for performance test of the server of android mobile application. After researching on the features of mobile performance test, this paper designed a B/S architecture-based test automation platform, and real devices running test models are supplied to meet demand. The paper devotes to doing research on the key technologies to implement the test platform, including: 1) provide the server monitoring during the execution of the performance test, so that testers can better manage and monitor all kinds of servers, and locate the performance bottlenecks, 2) study and realize the test execution model based on multiple agent, thus greater concurrency could be achieved. The test execution, which is based on the actual fleet of mobile equipment, can truly and effectively simulate users’ behavior and operating environment. Meanwhile, more perfect automatic test process is available out of this model, such as hybrid scenario settings and script management, 3) since test tasks of the platform share the limited device resources of the whole platform, an effective task scheduling strategy is in need to improve the resource utilization rate and the success rate, with which this paper provides. It applies load balancing strategy to allocate tasks to specific testing machines, thus avoids executing all tasks on a single test machine. Comprehensively taking account of the real-time load and task load of each device, the strategy allocates tasks in a way that load balance of the whole test fleet is guaranteed.