Fuzzy Logic Controller of Gentle Random Early Detection Based on Average Queue Length and Delay Rate

This paper proposes a controller technique for early stage congestion detection at the router buffer in the networks. The proposed technique extends the well-known Gentle Random Early Detection (GRED) algorithm. Unlike GRED, which relies on parameter settings, such as minthreshold, maxthreshold and double maxthreshold, in order to obtain a satisfactory performance, the proposed technique depends on a fuzzy logic system which reduces the large dependency on parameter settings. The proposed technique uses the average queue length and the delay rate as input linguistic variables for a fuzzy logic system. The utilized fuzzy logic system produces a single output that represents a packet dropping probability, which in turn control and prevent congestion in early stage. The proposed technique and the well-know GRED and REDD1 algorithms were simulated using Java environment. The performance of the proposed technique has been evaluated and compared with regard to various performance measures, which are: mean queue length, throughput, average queuing delay, packet loss and packet dropping probability. The simulation results show that the proposed technique, in comparison with the existing algorithms, offers better performance results in terms of mean queue length, average queuing delay and packet loss. Therefore, this technique, generally, improves the network performance.