Performance Evaluation of IEEE 802 11 based Ad Hoc Networks in Presence of Buffer

Abstract

Wireless communication has become a core part of modern communication technology. Among various wireless technologies Wireless LAN has gained immense popularity. The Infrastructureless wireless network, commonly referred to as Ad Hoc networks, has attracted extensive research interest for past 30 years. In this work, the focus is on IEEE 802.11 network performance analysis of single hop Ad Hoc networks under non-saturated network conditions. newlineTo meet the increasing demand of multimedia, it is necessary to provide the quality of service in such networks. The QoS supporting components at upper layer such as signaling and routing assumes the existence of MAC protocol which solves the problem of medium contention, supports reliable communication and provides resource reservations, particularly bandwidth. However, such assumptions are not realistic. The problem becomes more challenging in case of ad hoc networks. The performance evaluation of DCF 802.11 in ad hoc networks is always an important research area. The current work presents the development of two analytical models for network performance analysis. newlineFirst a two-dimensional Markov Chain model is proposed for analyzing the performance of IEEE 802.11 networks in presence of first order memory buffer considering the short retry limit. The throughput is obtained through this model. Compared with other analytical models, the proposed model includes more correct details of IEEE 802.11, and accordingly the results are more accurate. The accuracy of the model is proved by simulation results. newlineThe current thesis incorporates more practical scenarios while analyzing the performance of IEEE802.11 based networks. The work in this thesis helps network administrators to choose the feasible number of nodes to maintain desired QoS performance according to user requirements. The proposed analytical models guide system administrator to fine tune the buffer size in fast and accurate manner to predict the performance of applications.