Title: Predictable Performance Connectivity for AI Training and Inference.

Speaker : Prof Adrian Perrig, ETH Zürich, Switzerland
Date : 2026. 04. 10(Fri) 10:30
Location : #519, Jung Woonoh IT & General Education Center

Biography:
Adrian Perrig is a Professor at the Department of Computer Science at ETH Zürich, Switzerland, where he leads the network security group. He is also a co-founder of and board member at Anapaya Systems, an advisor to Mysten Labs, and a Distinguished Visiting Professor at City University in Hong Kong. He is a recipient of the ACM SIGSAC Outstanding Innovation Award, and is an ACM and IEEE Fellow. Adrian's research revolves around building secure systems -- in particular his group is working on the SCION next-generation Internet architecture.

Abstract:
As a next-generation Internet, SCION offers exciting opportunities for applications. With the operation of the global SCION network offering native connectivity to hundreds of thousands of users, an exciting direction is to add additional information to paths. This enables applications to learn about carbon emissions or propagation latency before sending a packet on that path. This property opens up new business models and opportunities that correspond to a next-next-generation Internet.

As an additional evolutionary step, we can provide future predictions of path information, such as latency, loss, jitter, etc. Such a framework enables completely new properties and opportunities. As AI training and inference require highly performant connectivity, the per-path performance prediction of SCION provides premium performance to demanding applications.