MANET-SIM: Scalable Simulator for Mobile Ad Hoc Networks

The exploding demand for computing and communication on the move has led to a significant need for ad hoc networks. However, there is relatively little understanding of the behavior of such networks, as they are scaled up, particularly for traffic withstrict Quality of Service requirements. The eventual goal of this SBIR is to develop MANET-Sim, a simulator for Mobile Ad hoc NETworks, with the ability to simulate wireless networks with tens of thousands of nodes, in real time, on common desktopplatforms that are likely to become available within the next three years. In Phase I, we will identify the primary bottlenecks to achieving real time simulation capability for such networks, and explore curative mechanisms including parallel execution,use of abstractions, and use of appropriate propagation layer models to facilitate accurate and fast simulation of mobile networks with tens of thousands of heterogeneous devices. We will quantitatively compare both execution times and accuracy ofdetailed packet level simulation models with abstract models including analytical, fluid flow, and multi-resolution models. A prototype methodology will be developed to demonstrate the feasibility of real time simulation of large scale wireless networksusing selected unicast and multicast protocols to route QoS sensitive traffic.The phenomenal growth rate of the Internet is a clich¿. It is also widely accepted that the most dramatic future growth, in commercial and military contexts, will be in wirelessconnectivity. The demand for tools to evaluate and manage these networks is growing together with the complexity and size of the heterogeneous networks. A South China Morning Post Online, 10 Jan 2000 report predicts that deals conducted throughwireless devices will generate about $14 billion a year and identifies mobile ticket sales, vending machine payments, and Internet shopping as the most promising wireless commerce areas. The Gartner Group estimates the market for software for networkmanagement & analysis to go from $3.6B in 1999 to $7.1B in 2003. The research and development activity proposed by this effort will directly lead to products in this large and growing market.