Measuring Latency at the Cloud Edge

Abstract

Quals Talk

Typical cloud architecture places backend resources in “regions” with private wide area networks providing fast paths between users and these regions. These have been shown to provide low latency for most of Europe and North America, but not for everyone. Recently, cloud providers started offering compute at the edges of their networks to provide even lower latency. In this talk I’ll highlight the improvement to latency made possible by using the cloud edge, how it can continue to improve based on cloud providers deployment plans, and why some locations still have high latency to the cloud edge.

Research area: Networking and Systems

Please join meeting in Cummings #302.

Zoom is not available for this event; disregard dial-in passcode included in email.