Optical switching for data centers and advanced computing systems [Invited]
Giannis Patronas, Nikos Terzenidis, Prethvi Kashinkunti, Eitan Zahavi, Dimitris Syrivelis, Louis Capps, Zsolt-Alon Wertheimer, N. Argyris, Athanasios Fevgas, Craig Thompson, Avraham Ganor, Julie Bernauer, Elad Mentovich, Paraskevas Bakopoulos
Abstract
We explore optical switching to extend network programmability to the physical layer and discuss applications of a Layer-1 software-defined network (SDN) in AI/HPC clusters. In this context we identify two applications for optical circuit switches (OCSs): failure resilience and reconfigurable topologies for deep learning workloads. We present experimental results from a DGX-based testbed towards improving failure resilience and a simulation analysis for efficient deep learning training in AI clusters.
Topics & Concepts
Computer scienceOptical switchOptical burst switchingTelecommunicationsOptical performance monitoringElectronic engineeringOptoelectronicsEngineeringWavelength-division multiplexingMaterials scienceWavelengthOptical Network Technologies