Predicting Needs in Future Decentralized Networks through Analysis of Barrage Relay Networks

Abstract

Improved routing algorithms are needed for the rapid proliferation of inexpensive, wireless network devices. In certain scenarios, decentralized wireless networks are either necessary or preferred over centralized ones. The barrage relay network (BRN) is an emerging ad hoc, decentralized wireless network designed to address issues of network reliability and routing overhead. In BRNs, the routing of unicast transmissions is controlled by the cooperative formation of controlled-barrage regions (CBRs), based on a simple set of rules. In this paper, we simulate CBR formation to study the routing reliability and node utilization, with the goal of predicting BRN utility in future scenarios. We have three specific aims which have not been addressed in other BRN studies: 1) we study the impact of channel effects on BRN routing, 2) we study the impact of node density significantly above the theoretical minimum density required to guarantee a fully connected network, and 3) we employ large ensembles of random networks to account for the wide variability that ad hoc networks may encounter. We find that CBRs tend to grow spatially as network density increases, leading to significant network utilization. Furthermore, we find with the most realistic fading model employed, a random channel model, requires the most network resources in high density networks. Then, we investigate a trade-off in network resources and reliability. Understanding these effects will lead to the design of more efficient and robust routing algorithms for high density decentralized networks for use in disaster relief, military, vehicle-to-vehicle and wireless sensor network applications.