Key points are not available for this paper at this time.
The RINSE simulator is being developed to support large-scale network security preparedness and training exercises, involving hundreds of players and a modeled network composed of hundreds of LANs. The simulator must be able to present a realistic rendering of network behavior as attacks are launched and players diagnose events and try counter measures to keep network services operating. We describe the architecture and function of RINSE and outline how techniques like multiresolution traffic modeling and new routing simulation methods are used to address the scalability challenges of this application. We also describe in more detail new work on CPU/memory models necessary for the exercise scenarios and a latency absorption technique that help when extending the range of client tools usable by the players.
Liljenstam et al. (Wed,) studied this question.