LAUSR

In Internet of Things (IoT) networks, congestion is growing with the increasing number of devices, and a large amount of collected data must be transferred. Congestion control is one of the most significant challenges for such networks. The Constrained Application Protocol (CoAP) has been adopted for the IoT to satisfy the demand for smart applications. However, CoAP uses a basic congestion control algorithm that operates only when congestion occurs. Thus, the basic CoAP and most similar loss-based congestion control schemes have remaining issues for burst data transfer in dynamic network environments. This paper proposes a novel rate-based congestion control scheme using fuzzy control for CoAP, called FuzzyCoAP. We use the round-trip time gradient and bottleneck bandwidth gradient as inputs for FuzzyCoAP to infer the degree of congestion. FuzzyCoAP uses this indicator to predict early congestion and adjusts the sending rate to avoid congestion. FuzzyCoAP uses the congestion degree to update the variable RTO for retransmissions. On the other hand, FuzzyCoAP dynamically checks for the available bandwidth to gain high performance for burst data transfer. Various simulation experiments have demonstrated the feasibility of the FuzzyCoAP in different traffic scenarios. We compared the proposed scheme with representative loss-based CoAP schemes, that is, the basic CoAP. The simulation results proved that FuzzyCoAP provides high performance in terms of delay, throughput, loss rate, and retransmissions compared with the basic CoAP.