Abstract:
An ad-hoc network is a collection of wireless mobile nodes forming a distributed reconfigurable network topology, which can operate without the aid of any fixed infrastructure or centralized control. The flexibility of implementation makes it an attractive network system for temporary emergency communication usage, such as disaster recovery communication systems. Simulations using NS-2 have been used for the evaluation of network performance in most studies of ad-hoc networks. Two potential communication problems (associated with the IEEE 802.11 standard's MAC layer design) have been reported. Specifically, these problems (affecting only multi-hop networks) are highly variable throughout and unfairly shared bandwidth. However, the simulation topologies implemented in these studies are highly unrealistic and unable to properly represent network behaviour in real world applications. In order to obtain more accurate simulation results, a more realistic simulation topology must be used. Consequently, the first objective of this research was to improve the topology implemented in the ad-hoc network simulations, to be more representative of the node layout likely to be found in an emergency communication system used for disaster relief. The second objective was to investigate the performance of the network by examining the data throughput and transmission delay of the network for different scenarios based on the newly developed realistic topology. The realistic simulation topology was designed to divide the transmission nodes into several categories for easy performance analysis and allow multiple routing options for every node in the simulation environment. The NS-2 simulation results showed that with reactive routing protocols (AODV and DSR), the TCP instability problems noted in simulations using simple topologies did not occur if implemented using the realistic topology. On the other hand, results obtained using a proactive routing protocol (DSDV) showed inconsistent throughput values because of the large amount of routing overhead produced by the protocol. Those simulations conducted to investigate the channel unfairness problem showed that the multi-hop transmissions shared much less bandwidth than the single-hop transmissions for both Two Ray Ground and Fading channel models. The reason for this problem lies in the mechanism of the MAC layer protocol, and the degree of this unfairness can be reduced by varying the TCP window size. The transmission delay of the network was found to be proportional to the number of active nodes in the system. The characteristics of transmission delay period were modelled to calculate the maximum node capacity of the system.