Congestion Control by Deviation Routing in Electromagnetic Nanonetworks

Pulse-based wireless nanonetworks differ in many ways from traditional wireless networks. This paper investigates congestion in multi hop nanonetworks, which do not behave as usual due to the specifics of the channel and the physical layer. Most protocols and network models assume that each node can listen to all of the traffic sent on their channel. In wireless nanonetworks, the capacity (in the order of Tb/s) of the shared channel is well beyond the processing capability of an individual node. Consequently, congestion arises from the limited buffers of individual nodes on the path instead of limited channel bandwidth. After defining congestion in this context, we propose a solution suitable to large, wireless nanonetworks. Instead of decreasing the sending rate to reduce overall traffic and congestion, we use the SLR routing protocol to find less saturated routes. Our evaluation demonstrates the effectiveness of this solution and shows that the throughput can be preserved with moderate activity overhead and latency cost.