Flow-correlation-aware opportunistic network coding for video streaming over wireless networks

Network coding is an emerging technique that can improve the performance of video communication in wireless networks. However, most of existing network coding mechanisms omit the characteristics of video traffic and the correlation among different video flows, resulting in degradation in throughput gain and bandwidth utilization. This work begins with a thorough investigation and understanding of the performance limitations of packet mixing schemes of existing wireless network coding mechanisms. On this basis, we propose a Flow-correlation-aware Opportunistic Network Coding (FONC) to improve the performance of real-time video communication in wireless networks. We take the correlation among different video flows into account, building upon which a novel packet mixing approach is present to enhance the amount of original data carried of a coded packet. We implement the FONC in ns-2 and carry out extensive performance evaluation. The simulation results demonstrate that FONC could enhance the quality of video transmission greatly, and utilize network bandwidth resource efficiently.