The analysis of burst packet losses in saturated ad hoc networks using semi-Markov process

Loading...
Thumbnail Image
Issue Date
2011-12
Authors
Sharma, Samir
Advisor
Pendse, Ravi
Citation
Abstract

Medium contention in saturated ad hoc networks increase collisions and retransmissions. As the number of transmissions for a frame exceeds the maximum retransmission count in the IEEE 802.11 networks, the medium access control (MAC) layer drops the frame and forwards this information to the network layer. The network layer further drops all the subsequent packets destined to the same next hop node or passing through this node, assuming the node to be unavailable. Furthermore, the transmitting node propagates this information through the route error (RERR) packet. The receiving nodes drop their packets destined to the same next hop node or if the packets are destined to through it to another destination. The network, therefore, experiences burst losses at several nodes, leading to the degradation of quality of real-time applications such as Voice over IP (VoIP). Hence, burst losses must be reduced and this requires a well-defined model to explain the burst losses and their causes. This thesis uses semi-Markov process based model for illustrating the burst losses in the IEEE 802.11 saturated ad hoc networks. It describes the frame transmission process in the IEEE 802.11 networks and the effect of medium contention on burst losses. Furthermore, the proposed model also computes the average frame service-time in the saturated ad hoc networks. Simulations are conducted to analyze the effect of variations in the network size (number of nodes in the network) and the data transmission rate on the burst losses. The results show that an increase in the network size and data transmission rate, independently increases burst losses. Simulation results present a range of acceptable network sizes to minimize the burst losses in the saturated IEEE 802.11 networks.

Table of Content
Description
Thesis (M.S.)--Wichita State University, College of Engineering, Dept. of Electrical Engineering and Computer Science.
publication.page.dc.relation.uri