Projects‎ > ‎

STAIRS - Maximizing Web Server Performance with Selective Compression

Background: Enabling HTTP compression in a web server can significantly reduce the bandwidth usage by 60-80% when dealing with HTMLs. HTTP compression is widely used today, mostly in gzip format. It does not only reduce bandwidth usage but usually also provides faster experience on the end-user's side. On the down side, it also reduces the maximum load a server can handle and may actually reduce site performance.

Related material

Project description: Building a web server (or modifing an existing one) which compresses the output streams with gzip, but only when it seems beneficial. Unlike existing web servers, this one will consider the server's load and the content's characteristics before trying to compress. In this project the students will create the server, the compression module and special clients that can bring the server to its performance limits. At the end of the project the students will demonstrate how the selective compression performs better that predetermined compression like Apache and Linux perform today. 

Students: Alexander Tzigirintzev, Yevgeniy Sabin