Genetic CoCoA++: Genetic Algorithm based Congestion Control in CoAP

CoAP is the protocol used in various IoT device communications. Congestion in the network is a reason for packet loss during the transmission. Packet loss can be reduced via reducing congestion in the network. CoAP uses a simple binary exponential backoff factor which does not consider the dynamic nature of the network. A major enhancement for congestion control in the IoT network is achieved via CoCoA which somehow considers the network dynamics a little bit. New enhancements over this protocol are made overtime but those enhancements are based on the static observation of RTT. In this paper, a dynamic approach based on the observation of the network for a fixed interval of time, and based on the genetic algorithm is discussed. This approach also does not store previous states of the network for congestion control and RTO estimation. We worked on this approach and a significant decrease in RTO and PFR is observed via this approach.