NCRAWL: Network Coding for Rate Adaptive Wireless Links

Intersession network coding (NC) can provide significant performance benefits via mixing packets at wireless routers; these benefits are especially pronounced when NC is applied in conjunction with intelligent link scheduling. NC however imposes certain processing operations, such as encoding, decoding, copying and storage. When not utilized carefully, all these operations can induce tremendous processing overheads in practical, wireless, multi-rate settings. Our measurements with prior NC implementations suggest that such processing operations severely degrade the router throughput, especially at high bit rates. Motivated by this, we design {\bf NCRAWL}, a Network Coding framework for Rate Adaptive Wireless Links. The design of NCRAWL facilitates low overhead NC functionalities, thereby effectively approaching the theoretically expected capacity benefits of joint NC and scheduling. We implement and evaluate NCRAWL on a wireless testbed. Our experiments demonstrate that NCRAWL meets the theoretical predicted throughput gain while requiring much less CPU processing, compared to related frameworks.