On modeling and impact of geographic restrictions for human mobility in opportunistic networks

Abstract

Human mobility has been shown to have a significant impact on the performance evaluation in Opportunistic Networks. As shown in previous work, the integration of geographic restrictions is (1) a further step towards more credibility of Opportunistic Network simulative performance evaluation, and (2) influences relevant metrics, such as contact duration and number of contacts. In this paper, modeling challenges related to geographic restrictions are in our focus. We investigate the impact of the road network underlying the scenario’s map extract. The diversity in road network structure has less impact on the Inter-Contact Times but causes considerably shorter and less contacts for dense road networks. This makes dense road networks an especially challenging urban scenario for forwarding algorithm performance evaluation. As integrating geographic restrictions is an additional effort, we examine the runtime for trace generation and contact computation. It increases significantly but should be accepted as a trade-in for more realistic and credible modeling.