Enabling and Exploiting High-Speed In-Network Computing

Explore the potential of in-network computing in this lecture by Vishal Shrivastav, Assistant Professor at Purdue University's School of Electrical and Computer Engineering. Discover how to overcome the computational constraints of high-speed programmable network devices that limit the full potential of in-network computing. Learn about two innovative systems: Seer, which addresses limited on-chip memory through an online caching scheme leveraging network delays, and Leo, which efficiently utilizes constrained compute resources on programmable switches to implement ML-based traffic analysis models. Understand how these advancements enable real-time network intrusion detection systems capable of inspecting every packet at multi-Tbps speeds. The talk presents cutting-edge research from Shrivastav, a recipient of the NSF CAREER Award, Google Research Scholar Award, and other prestigious recognitions for his work in programmable networks and in-network computing.