Wireless mesh networks promise a flexible and cost-effective solution for bringing high-bandwidth low-latency applications to the home. Two orthogonal but immensely attractive approaches for designing high-performance mesh networks are network coding and cross-layer optimization. The former exploits the interfering wireless media as a broadcast channel by ingenious information processing, while the latter intelligently allocates and shares network resources across the layers. This project investigates the long-overdue and highly challenging synergistic union of network coding and cross-layer optimization.
The characterization of intersession network coding (INC), coding across different network flows, is NP-hard. Nonetheless, the problem can be made tractable under important practical settings. Based on new path-based characterizations, this project rigorously quantifies the network capacity in presence of different practical INC schemes. A new network-coding based cross-layer paradigm is developed for controlling mesh networks that takes into account practical considerations such as robustness, scalability, and time-varying wireless broadcast environment. New distributed, low-overhead protocols are constructed and validated over a mesh testbed on the Purdue University campus. Undergraduates and under-represented students are actively recruited for the project. Research results are disseminated via the web and through publications in conferences and journals. The results of the project, spanning from theoretic exploration to protocol deployment on a practical testbed lead to the significant performance gains necessary to realize the full capability of mesh networks, which in turn has the potential to revolutionize the telecommunications industry.
This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).
