Proactive In-Network Caching for Mobile On-Demand Video Streaming

Abstract

Mobile video streaming has become an essential application in mobile wireless networks, making up most of the mobile data of today's Internet traffic. Studies have shown that mobile video data is projected to make up about 78 percent of the global mobile data traffic, and that global mobile data traffic is expected to increase sevenfold by 2021. Massive small cell base station (SBS) deployments have emerged as a potential solution promising to fulfill these unprecedented mobile data demands, by offering great coverage enhancements and maintaining high quality of video streaming. However, due to relatively small cell sizes and high user mobility, mobile video streaming in dense SBS networks faces fundamental challenges such as intermittent connectivity and frequent handoffs, causing degradation in video streaming quality. In this paper, we tackle this issue by introducing a hybrid proactive in-network caching framework that stores some popular videos at the edge of the network, namely at the SBSs, while also pre-caching video contents in advance to better service mobile users. The proposed framework essentially reduces the need for bringing every requested video from the core (original) network, which results in alleviating network congestion by reducing back-haul traffic and in improving mobile video streaming experience by avoiding service discontinuity during handoffs. We develop a simulation framework using MATLAB to study the performance of the proposed caching technique and show that the proposed technique can effectively improve video quality of experience and reduce back-haul traffic.