Performance analysis of a decentralized network simulator based on IMUNES

IMUNES, a fast integrated network emulator/simulator, presents an excellent replacement for the general purpose network test-bed since it increases scalability while providing support for real applications and protocols, as well as decreasing the cost of equipment and management. However, working on one machine only the scalability of IMUNES is limited, leading to inefficient and complex management of large-scale network simulations. In this paper we present a distributed version of IMUNES, designed to circumvent this problem. Guided by the principles of scalability, transparency and security, the resulting system is built on top of a decentralized structure. We evaluate the resulting system performance in terms of scalability, responsiveness, latency and throughput. Based on the observed values, it can be concluded that the simulation capacity increase is proportional to the increase in the number of machines available in the distributed system.