Peering infrastructures, namely, colocation facilities and Internet exchange points, are located in every major city, have hundreds of network members, and support hundreds of thousands of interconnections around the globe. These infrastructures are well provisioned and managed, but outages have to be expected, e.g., due to power failures, human errors, attacks, and natural disasters. However, little is known about the frequency and impact of outages at these critical infrastructures with high peering concentration.

In this talk, I will first present the tectonic changes in the peering ecosystem and the role of peering infrastructures. Then, I will present our novel and lightweight methodology for detecting peering infrastructure outages. This methodology relies on the observation that BGP communities, announced with routing updates, are an excellent and yet unexplored source of information allowing us to pinpoint outage locations with high accuracy. I will also present Kepler, a system we built to locate the epicenter of infrastructure outages at the level of a building and track the reaction of networks in near real-time. Our analysis unveils four times as many outages as compared to those publicly reported over the past five years. Moreover, it shows that such outages have a significant impact on remote networks and peering infrastructures. Our study provides a unique view of the Internet’s behavior under stress that often goes unreported.

Short Bio

Georgios Smaragdakis is a Professor at TU Berlin, and a visiting researcher at MIT and Akamai. In the past, he was a researcher at Deutsche Telekom Labs and Telefonica Research. He earned his PhD from Boston University and his BEng from TU Crete. George’s research was awarded a European Research Council Starting Grant Award (2015), a Marie Curie International Outgoing Fellowship (2013), and best paper awards at IEEE INFOCOM (2017), ACM IMC (2016 and 2011) and ACM CoNEXT (2015). His research interests include Internet measurement, content delivery, and network and web analytics.