World Wide Web is consist of large amount of information and provides it different kinds users. Everyday number of users use log on internet. Internet information growing enormously. Users accesses are documented in web logs. As huge storage log files are growing rapidly. One of the application of data mining is web usage mining works on users logs. It consists of various steps such as user identification, session identification and clustering. Again removing robot entries. In previous years data preprocessing analysis system algorithm on web usage mining has been used buts algorithm lacks on scalability problem.