Network switches are an attractive vantage point to serve various network applications and functions such as load balancing and virtual switching because of their in-network location and high packet processing rate. Recent advances in programmable switch ASICs open more opportunities for offloading various functionality to switches. However, the limited memory capacity on switches has been a major challenge that such applications struggle to deal with. In this paper, we envision that by enabling network switches to access remote memory purely from data planes, the performance of a wide range of applications can be improved. We design three remote memory primitives, leveraging RDMA operations, and show the feasibility of accessing remote memory from switches using our prototype implementation.
Daehyeok Kim is a third year PhD student in the Computer Science Department at Carnegie Mellon University, where he is advised by Professor Srinivasan Seshan and Professor Vyas Sekar. His research interests lie in the intersection of systems and networking with a current focus on making data centers faster and more efficient by designing novel network primitives with advanced networking hardware such as programmable switches and RDMA NICs. Before coming to CMU, he worked at KAIST for three years as a research scientist. He received his BS and MS degree in computer science and engineering from POSTECH.
This is a practice talk for HotNets-2018.