Large-Scale Reconfigurable Optical Networks

Prof. Manya Ghobadi MIT Wednesday, June 30, 2021 @ 3:00 pm Virtual Hosted by: Prof. Katerina Argyraki

Abstract

Large-scale networks, such as datacenters and wide-area networks, are the foundation of modern online services. However, the design of today’s networks still follows the telephony model where network operators treat the physical layer of networks as a static black box without any reconfigurability in it. As a result, the network is provisioned to carry the worst-case traffic demand under all plausible failure scenarios, making it excessively inefficient and prohibitively expensive. In this talk, I will argue that modern networks have elastic applications that can benefit from a dynamically reconfigurable physical layer. I will highlight two concrete use cases: (1) reconfigurable clusters for distributed machine learning workloads; and (2) reconfigurable wide-area networks to recover from fiber cuts.

Bio

Manya Ghobadi is an assistant professor in the EECS department at MIT. Before MIT, she was a researcher at Microsoft Research and a software engineer at Google. Manya is a computer systems researcher with a networking focus and has worked on a broad set of topics, including data center networking, optical networks, transport protocols, and network measurement. Her work has won the best paper award at the Machine Learning Systems (MLSys) conference as well as best dataset and best paper awards at the ACM Internet Measurement Conference (IMC).