We will be receiving Gilles Trédan from CNRS/LAAS Toulouse who will give a talk about networks and failures (corresponding to papers published in Infocom, DSN, Sigcomm CCR and TON). on Wednesday, April the 3rd at 2pm in room Lipari.
Abstract: In this talk, I will present some results related to local fast failover in networks. The problem is fairly simple: given a (highly connected) network, design a routing algorithm that can cope with future link failures. Furthermore, we require the solution to be local, in the sense that each router’s decision only depends on the state of its adjacent links. This property enables fast reactions to network failures as opposed to solutions where a centralized view of the network is required (typically milliseconds vs seconds). Ideally, we would also love our solution to avoid unnecessary detours, and balance load across redundant links. But the world we live in is not ideal, even for datacenter operators. Depending on time, I might also talk about network tomography: how to monitor a network infrastructure by deploying a minimal number of monitoring equipment à la traceroute.
Bio: I obtained a PhD degree in computer science from University of Rennes 1 in November 2009. I prepared my degree in the ASAP team, under the supervision of Achour Mostefaoui. From January 2010 to September 2011, I worked as a postDoc in the FG Inet group , Berlin. Since November 2011 I work as a full time researcher in the TSF group.