Design And Implementation Of The Deployment Environment In The Stress Testing Framework Supported For The Http Protocol

Under the situation that mobile terminals are exponential increasing; services and applications continuously deepen and refining; Web access sustained grow, the response speed of the servers would be slowed because of the CPU overload, NIC overload or other reason or even all the services would be interrupted. In order to discover and expose these situations, help software development engineers to locate the problems and estimate the peak correctly with improving the efficiency of the testing engineers, a well user-friendly stress testing framework that can truly simulate a large amount of concurrent HTTP requests is needed to do the stress testing, so that the data basis can be provided for the optimization and adjustment of product performance.The main purpose of this framework is to realize the automated deployment of the stress testing environment and automated collection of the stress test results. When the system pressure test finished, it can automatically push the test results to the user in time. In above functions, the authors mainly completed the design, implementation and testing works for the automated deploy pressure environment module.Environment deployment module can be subdivided into configuration file processing module, trust relationship build module and environment deployment module. Among them, the configuration file processing module is mainly responsible for reading configuration files and its correctness check, implemented by using C/C++lib. Trust relationship module is mainly implemented through Secure Shell (referred to as SSH) to establish a trust relationship with the target machine, and the environment deployment module use C/C++lib to backup, restore the original environment and deploy the files in the machine which had been deployed.Before this system is put into operation, test engineers have to spend about six hours to build the environment, manual monitor the machines’ performance metrics and collect the test results under the bare machine environment. While using this framework, the same operation only need about two hours, increasing the efficiency by65%.