Towards Minimizing the Required Bandwidth for Mobile Web Browsing

Abstract—The number of smartphones has been increasing at a very fast pace over the last several years. At this rate, it is quite likely that mobile web traffic will take a serious portion of the global Internet traffic in the very near future. However, the improvement of mobile web access technology has been known to lag behind the current growth: we face issues such as limited bandwidth and computational power, and high latency. Additionally, users have financial incentives to use their limited data plan effectively. These facts motivate new ways of designing and building mobile web access technologies and networks. Specifically, new methods for reducing mobile web traffic and for utilizing the mobile computational and network resources more efficiently should be found. In this paper, we propose a novel technique for reducing the needed bandwidth for mobile web browsing. It automatically identifies redundancies in mobile website data to reduce the number of transmitted bytes. We leverage the standard dedu-plication techniques of partitioning website data into chunks and fingerprinting them for redundancy detection and efficient retrieval from a cache. The only additions to the current mobile browsing framework that our method requires is a dedicated proxy server and the installation of a mobile proxy-client. We implement our techniques in a sophisticated simulator and evaluate their effectiveness on a number of popular websites. Our empirical results show that, with an appropriate tuning of all system components, the needed bandwidth can be consistently reduced to less than 20 % in most cases. I