Alejandro Cano – Random Topologies in Interconnection Networks: Theory and Simulation

As data centers scale to meet the demands of next-generation computing, designing efficient, low-cost network architectures becomes increasingly critical. This talk provides a comprehensive introduction to low-diameter network topologies, exploring fundamental network properties, topological limits such as Moore’s bound, and the application of random graphs to interconnection networks.

Moving from theoretical topology design to evaluation, the second part of the talk focuses on network simulation, specifically examining the traffic models used in the CAMINOS network simulator. Implementing realistic traffic simulation models is necessary to accurately assess these next-generation networks under distributed workloads. Ultimately, this presentation will feature a brief evaluation of how random graphs behave compared to other topologies.