The Performance Testing And Tuning Of M2M Business Support Platform

M2M business support platform is developed for the integration of different data collect business to provide flexible and reliable collect business. It also gives a solution on cross-platform and cross-area data transferring in heterogeneous environment. M2M not only supports different data collect application business but also collects, manages and forwards business messages by separating the management processing and business processing.M2M business support platform's users are in a large scale multitude. It includes not only thousands of the end users and different administrators, but also diverse terminals and business application systems. This characteristic implies a high performance demand to the platform. But the business logic of M2M is miscellaneous and the control between layer and layer is complicated. It makes the system vulnerable to performance. So performance testing and tuning become more important than other small scale J2EE application during the developing period.According to the business characteristic of M2M, this paper sets up a load performance testing model from top to bottom. Based on this model, the load testing scheme is designed and implemented .Then system is tuned and re-testing to confirm the effectiveness.Firstly, the paper introduces the research background of M2M, and then it describes the current domestic and international research status of the M2M technology and its tendency and application prospects. Secondly, the paper researches on the theories and methodologies of the performance testing for the subsequent testing implementation. Thirdly, the paper describes the architecture, use case, function module, business processing and technology implementation of M2M business support platform.Based on the analysis of the characteristics of the system and the researching on the performance testing, the paper gives a top-down modeling approach and gets a actual model of the M2M business operation based on it. After that, the scripts and scenarios are designed according to the modeling result. The paper gives an analysis report after the testing.The performance tuning mainly includes data access tuning and JVM tuning. A testing re-executed after the tuning for validating the effectiveness of the tuning.At the end, the paper summarizes all the content and practical value of this research work, describes the work of the author and points out the direction of further research.