Lessons learned from real MANET experiments and simulation-based evaluation of UDP and TCP

Abstract

Although more than a decade of research has been done but pure general-purpose MANET is still not available rather than few prototypes within laboratory due to both technical and socio-economic point of view. Lacking in appropriate guidelines for realistic user traces, mobility models, routing protocols and considering real-life challenges, it is difficult to reproduce any typical scenario in reality apart from simulation. In this paper, difficulties faced to regenerate real-life scenarios have been discussed to clearly identify the gaps in simulation and real-time experiments. Four laptops are used in an open field environment for different scenarios to evaluate a TCP based streaming video application using real OLSR implementation within a IEEE 802.11g wireless network. Corresponding simulations are performed in ns-2 based on the realistic setup parameters achieved from real experiments and finally a comprehensive analysis identifies the generic gaps between these two approaches to evaluate network protocols. Simulation results show better performance than the real-life results due to differ in external influences and protocol implementation although maintaining realistic simulation setups.