Simulating Large Topologies in ns-3 using BRITE and CUDA Driven Global Routing

We present two new modules for ns-3. The first is a BRITE module that allows ns-3 users to take advantage of the topology generation features of the BRITE topology generator. This module allows users to quickly create highly customizable, large scale topologies for simulation within ns-3. The second module is a new Global Routing module specifically designed to work with large ns-3 network topologies. This module leverages the parallel processing abilities of GPUs using CUDA. Using this new routing module on large topologies, we measured speedups of over three orders of magnitude compared to ns-3's current global routing protocol when performing global route computations.