NOWIRE · Network Coding for Wireless Networks

Our goal is to develop fundamentally new architectures for wireless networks that offer the convenience of wireless communication while achieving the performance, predictability and security of wired networks. The wireless channel is inherently a shared medium characterized by limited resources and complex signal interactions between transmitted signals. The question we address is how do we transmit information over wireless and how do we exploit the wireless channel properties to share its resources. Ours is a fundamentally different approach to existing strategies, that builds on new physical and packet layer sharing and cooperation paradigms that we have been working on, to extract the optimal throughput and reliability performance from the wireless medium. These are recent breakthroughs in (i) network coding and (ii) wireless cooperation. Network coding is a new area bringing a novel paradigm for network information flow that enables cooperation at a packet level to optimally share the network resources. Deployment of the first network coding ideas in wireless have already indicated benefits as large as a factor of ten in terms of throughput. Complex signal interactions caused by the inherent broadcast nature of wireless channels, is traditionally viewed as an impediment to be mitigated. Recently it has been demonstrated that one can utilize interference to develop cooperation at the wireless signal level (physical layer) for arbitrary wireless networks. This can give significant capacity advantages over techniques that mitigate interference. Both these ideas can radically affect the way information is communicated, stored and collected, and can revolutionize the design of future wireless networks. In this project we plan to addess several fundamental questions that develop on these themes. We take a complete view of these ideas by not only developing the underlying theory but also through validation on wireless testbeds.