A locally-optimizing approach for multichannel assignment and routing

In this paper, we propose a distributed online multichannel assignment and routing algorithm called LOAM. Using LOAM algorithm, a node in the network assigns channels for the incoming traffic flows and forwards them based on the channel usage information of its nearby nodes. We also present a model to analyze the throughput of multichannel networks and prove LOAM can achieve more than a fraction of the optimal throughput. Furthermore, we compare the achievable throughput of LOAM with a centralized offline channel assignment algorithm in average cases considering different channel rate and traffic load scenarios, the results of which demonstrate that LOAM can achieve more than 90% of theoretical maximum throughput and has a more balanced load distribution. Finally, we evaluate LOAM in a realistically simulated 802.11a wireless multihop network, compared with a random channel assignment algorithm. The simulation results show that LOAM can deliver up to 25 percent more messages with a shorter delivery latency.