ABSTRACT
The World Wide Web (WWW) has undergone remarkable change over the past few years, placing substantially
heavy load on Web servers. Today’s web servers host web applications that demand high computational
resources. Also some applications require heavy database retrieval processing, making server load even more
critical. In this paper, performance of Apache web server running compute and retrieve-intensive web workloads
is analyzed. Workload files implemented in three dynamic web programming technologies: PERL, PHP and Java
Servlets are used with MySQL acting as a data source. Measurements are performed with the intent to analyze
the impact of application workloads on the overall performance of the web server and determine which web
technology yields better performance on Windows and Linux platforms. Experimental results depict that for both
compute and retrieve intensive applications, PHP exhibits better performance than PERL and Java Servlets. A
multiple linear regression model was also developed to predict the web server performance and to validate the
experimental results. This regression model showed that for compute and retrieve intensive web applications,
PHP exhibits better performance than Perl and Java Servlets.
Keywords: - Web performance analysis, Web Servers, compute and retrieve intensive