Network coding aware queue management in multi-rate wireless networks

In this thesis we focus on network coding for unicast flows, in particular we take into consideration the COPE architecture. We study what happens when two nodes transmit with different rates and we propose a Markov chain to model this scenario. Furthermore, we propose a queue management algorithm to increase the coding opportunity and the throughput of the network. In COPE, a node codes two or more packets together when packets are headed to different nexthops. When there are just packets for the same nexthop, a node loses a coding opportunity. The queue management algorithm increases the coding opportunity prioritizing the channel access of sender nodes based on the queue information and on PER of the links. We simulate this algorithm on top of COPE in a multi rate scenario with NS 2 and we show that our algorithm yields throughput gains of up to 57\% compared to COPE