Based On Log Analysis, Web Load Testing Research

The conception and application of Web Load Testing have been past for many years, but with the rapid development of network and information, it has meet a new challenge. Its scope and complexity have been to a new high degree. Traditional software load testing is mainly exist in LAN/WAN and running with the predictable time and fixed users. But the Web applications are running in the open Network with inscrutable software/hardware configurations and user behaviors. At the same time, inferior performance and poor validity, which are caused by poorly designed Web application software are unacceptable. If we can't solve this problem in the competing network, sooner or later it will affect the company's benefit, the quotient of business and the brand of production. How to ensure the correctness and reliability of web application becomes an attentive problem more and more. As an important means of assuring the quality and reliability of software, Web load testing becomes an importance chain of software development process.In this thesis, we make a study of Web load testing and its related techniques. Having studied the performance indicators and the relationship between them, we analyzed in detail the Web log and its key variables and metrics, which are used to analyze and produce Web application load, identify the validity of testing. Then bringing forward the Web application log analysis (WALA). We can get two benefits from using this method, firstly, the variables and metrics can been used in a systemic manner, secondly, we can apply the method to the Web load testing process and enrich the present knowledge of Web load testing process. Finally, Web load testing tool (WLTT) has been developed. It can simulate thousands of users and their behavior, generate many levels of work load. With the help of WLTT, we can find bottleneck and make some performance optimization. We expect that our work will do some help in Web load testing.