Analysis of web data compression and its impact on traffic and energy consumption
- While connection speeds are increasing slowly, some ISPs mention plans about possible traffic limitations in the near future which would keep internet traffic expensive. In addition, Green IT became more important especially over the last few years. Besides on-demand content like video live streams, HTTP traffic plays an important role. Thinking of textual web content, compression quickly comes to mind as a possibility to reduce traffic. The current HTTP/1.1 standard only provides gzip as an option for content encoding. HTTP/2.0 is under heavy development and numerous new algorithms have been established over the last few years. This paper analyzes HTTP traffic composition on a production server and concludes that about 50% is compressible. It further examines the effectiveness of custom and existing algorithms with
regards to compression ratio, speed, and energy consumption. Our results show that gzip is a sound choice for web traffic but alternatives like LZ4 are faster and provide competitive compression ratios.