Traffic optimization with redundancy elimination on mobile and wireless networks

Mobile video services dominate half of the cellular traffic and 40 % of the WiFi traffic today [4, 5]. However, spectrum resources in mobile and wireless networks are quite restricted. The Cisco Visual Networking Index [2] shows that the average mobile network downstream speed in 2013 was 1387 kbps, whereas the standard HD (720p) videos on YouTube are encoded as 5000 kbps [3]. Thus, to provide higher goodput performance and better quality of experience (QoE) for mobile video services, it is essential to use resources more efficiently.;Redundancy is prevalent in today's networks, and leads to waste of network capacity and degradation of network performance. The redundancy can be at the byte level-i.e., repeating bytes in flows, or it can be at the information level-i.e., inherent meaning of objects for users such as the content of videos.;In this thesis, we propose REfactor and iProxy to remove the redundancy.;REfactor eliminates the byte level redundancy at fine granularity and fully leverages the broad- cast nature of wireless radios. To realize the benefits of IP-layer content overhearing, various challenges must be overcome that arise from the probabilistic nature of wireless reception (which could lead to inconsistent state) and the limited resources of wireless devices. We overcome these challenges through careful data structure (self-addressing cache) and wireless redundancy elimination designs (model driven RE). We find that our design is highly effective and can improve goodput by nearly 25%.;We provide another solution, iProxy, to remove redundant videos in cache storage and adapt bitrates for clients to improve QoE. iProxy acts as a middlebox between video providers and clients. It elevates the traditional view of caching from "data" to "information", which optimally will reduce redundant videos in caches. It uses information bound references as video identification to collapse multiple related cache entries into a single one, improving hit rate and lowering the storage costs. Furthermore, it incorporates a novel dynamic linear rate adaptation scheme to ensure high stream quality with client and channel diversity. Our evaluation shows that iProxy can improve cache hit, start-up latency, and video quality compared to MPEG DASH.