Strategy And Study Of Tuning Database Performance Based On SQL Server

With the rapid development of computer science and digital information, almost every corporation creates their owner information system. As the core and basic component of these information systems, database is used more and more in these systems. So many attentions are paid to the database performance.The traditional performance testing is usually done manually, thus it is impossible to simulate the actual condition when large quantities of concurrent users get access to the system. With the development of technologies, more and more softwares have been designed to simulate the concurrent-user scenes and carry out performance testing, which explores the system performance successfully.LoadRunner is a professional performance testing tool. By simulating the condition that thousands of users request and operate the application simultaneously, possible task pressure in the real condition will occur randomly in the lab environment, by analyzing the information and data acquired in the testing process, the problem causing software performance problem will be found and solved.The general process of LoadRunner performance testing to the system is described, and its working principle is presented. An actual system has been tested by employing LoadRunner, and the detailed process which includes creating script, operating scheme, monitoring scheme, and analyzing results has been given.Database performance tuning problem is researched in the paper. First, the research and development of database performance tuning and study value of this thesis are introduced in exordium. Second, the basic research frame of database performance tuning, including conception, aim, strategy and approach are introduced. And then LoadRunner testing process is fully described and analyzed, the test on a target server is created. Using terminal emulation of concurrent users on SQL Server 2000 database system operation, a detailed testing and analysis is carried out. Through the detailed tests, the optimization goal is attained. And then apply a top to down methodology, and propose some tuning suggestions from the application layer, database layer, and operating system layer and hardware layer aspects. Especially for the tuning applications, a variety of specific optimization ways is given and verified. After all the adjustment, LoadRunner is adopted again to conduct a comprehensive testing work. The response time, throughput and average number of clicks per second has apparently improved, which achieves the expected goal.