DNB-AODV: Dynamic Neighbors Based Contention Window Approach for IEEE 802.11 AODV MAC Protocol

Abstract

The Backoff algorithm is one of the most important algorithms which plays a significant rule in terms of data transmission. This paper introduces a new backoff algorithm which is the Dynamic Neighbors Based (DNB) backoff algorithm. DNB combines the standard binary exponential algorithm with the octal exponential algorithm depending on the number of neighbors around the sender node. When the sender node is located in a dense area, the possibility of competition on the network media becomes higher so DNB algorithm selects the octal exponential algorithm which relatively guarantees long back off time that is fair enough to check whether the medium is busy or not. But when the sender node is located in a sparse area, the possibility of competition on the transmission medium becomes less. Thus, DNB algorithm selects the standard binary exponential algorithm which relatively guarantees less backoff time than the octal exponential algorithm. The simulation results show that the DNB protocol achieves an enhancement on packet delivery ratio, up to 2%, as compared to the standard binary exponential and the octal exponential algorithm and up to 16% as compared to the logarithmic algorithm. This enhancement achievement is more tangible in more stable networks (high pause time values). The simulation results are obtained by the well-known Glomosim Simulator, version 2.03, without any distance or location measurements devices.