Indian Journal of Science and Technology

Abstract

Background/Objectives: Congestion in wireless network causes loss of packets, longer delay and throughput degradation. So it is always a major concern to network. Many sources send data of different priorities in different applications like Multimedia. So weighted fairness for these applications is also required. Methods: To address these issues we propose a protocol named FPECODA for Mobile Ad hoc Network. Findings: It uses the GPU for parallel version of algorithm for finding the Dynamic Priority of packets, sorting algorithm etc. It also applies the fuzzy logic for rate control of node. Subsequently handling the congestion will become more effective by: a) Bottleneck node based source data sending rate control. b) Flexible Queue Scheduler for packets transfer. Improvements: Experimental results demonstrate that processing of packets becomes significantly faster than CPU only based algorithm. Also it will be very efficient in terms of packet delivery ratio, end to end delay and fairness in network.

Keywords: Congestion Control, GPU, Multimedia, Parallel Computation, Multiple Class of Traffic