Speedup techniques for analyzing scalability of large scale mobile ad hoc network management systems

Abstract

The need for scalability analysis of mobile ad hoc networks (MANETs) is both obvious and critical. However, the complexity of MANETs coupled with the unpredictable environment they are designed to operate (e.g., highly dynamic link conditions and random mobility patterns) renders use of closed form solutions infeasible unless over-simplifying assumptions are made. The use of modeling and simulation (M&S) techniques that allow us to retain the complexity of the system therefore become invaluable while analyzing such systems. While current discrete event simulation (DES) tools render the analysis of complex MANETs feasible, they are often times limited to analyzing small (O(100)) node networks. Scalability analysis (e.g., O(500)+ nodes) with current DES tools however tend to require very long simulation times. In this paper we describe approaches to speeding up the M&S phase while analyzing MANETs. More specifically, we propose techniques to analyze the scalability performance of a dynamic policy-based network management system for MANETs that are instrumental in maintaining the "health" of the underlying MANET.

Simulating CSMA/CA behavior for performance evaluation of multi-hop wireless networks

“…Network simulators are often required to be efficient in terms of memory space and processing power because they should usually deal with largescale networks such as vehicular ad-hoc networks and wireless sensor networks, or they should be run a number of times for exhaustive tests of applications. In particular, it is known that event scheduling in the MAC-layer frame transmission consumes a lot of computational resources in wireless network simulations [1], [2]. Accordingly, several techniques have been proposed to deal with this issue [3], [4], [5], [6].…”