MAP Mobility Modeling

Realistic modeling of movement patterns of mobile nodes in mobile ad hoc networks using Markovian arrival processes

Mobility models are essential for evaluating different design alternatives for protocols and applications in mobile ad hoc networks (MANET). Mobility models specify the rules, according to which nodes in a MANET move. This enables the quantitative evaluation of protocols and applications for MANET under various movement patterns. Known mobility models exhibit weaknesses in the realistic modeling of movement patterns of pedestrians or cars.

The goal of the project is the derivation and evaluation of a realistic model for movement patterns of mobile nodes in a MANET. In contrast to known mobility models, where movements are constructed according to rules with a memoryless property, the long term dependencies will be taken into account. Furthermore inter-contact times and durations of mobile nodes will be used as parameters for the aggregated modeling of the movement patterns. As a result, the project will provide a synthetic generator for movement patterns as well as Markovian arrival processes for determining these parameters, which will reproduce or approximate the long term dependencies from reality. A synthetic workload generator for peer-to-peer file-sharing systems [1] and a model of aggregated IP traffic using a specially tailored Markovian arrival process [2] will serve as a starting point.

[1] A. Klemm, C. Lindemann, M. Vernon, and O. Waldhorst, Characterizing the Query Behavior in Peer-to-Peer File Sharing Systems, Proc. Internet Measurement Conference (IMC 2004), Taormina, Italy, October 2004. PDF

[2] A. Klemm, C. Lindemann, and M. Lohmann, Modeling IP Traffic Using the Batch Markovian Arrival Process (extendend version), Performance Evaluation, 54, pp. 149-173, 2003. PDF

Funding

German Science Foundation (DFG): January 1, 2008 till December 31, 2009