An Empirical Study on a Web Server Queueing System and Traffic Generation by Simulation
In this study, we first analyze the log data of the web-server system at a Korean company and confirm strong self-similarities that have been reported in a wide range of internet data. Then, we propose a simulation approach to generate arrivals of web page requests based on the conventional familiar probability distributions. We also present a forecasting model to generate future arrivals and test the validity of our approach through real data. Finally, we present our software tool that we developed to analyze web log data.
Unable to display preview. Download preview PDF.