Assistant Professor
Networked Systems Laboratory, Director
Swiss Federal Institute of Technology, Lausanne, Switzerland
Webpage

 

 

 

 

 

Tuesday, 14 December 2010 (15h00-15h30), EPFL Polydôme

Responsive, Energy-Proportional Networks

The power consumption of the Internet and datacenter networks is already significant due to a large degree of redundancy and high idle power consumption of network elements. Therefore, dynamically matching network resources to the load is highly desirable. However, this is difficult because the traffic changes quicker than it is possible to compute the minimal network subset to carry the traffic demand. In this talk, I describe how we achieve responsiveness by decoupling the decisions taken by routing and online traffic engineering (TE), and propose Energy-Proportional Networks (EPN) -- networks which use the minimum amount of energy to carry the required traffic. EPN computes three sets of routing tables: i) always-on, ii) on-demand, and iii) failover. A simple energy-aware TE algorithm (in)activates network elements to achieve the goal of energy-proportionality. Our evaluation on ISP and datacenter topologies shows that EPN achieves the goal of energy-proportionality and saves up to 42% of power, without sacrificing responsiveness. Further, using a Click testbed we show that it is possible to: 1) quickly and efficiently use the EPN paths at runtime for energy-saving, 2) quickly tolerate faults. Finally, two representative applications running over EPN-chosen paths demonstrate EPN's marginal impact on application-level throughput and latency.

 

About the speaker:

Dejan Kostic obtained his PhD in Computer Science from Duke University in 2005, his M.Sc. degree in Computer Science from the University of Texas at Dallas, and his B.Sc. degree in Computer Engineering and Information Technology from the University of Belgrade (ETF). From 1995 until 1999, he worked at INET, Inc. (now Tektronix) and MakeLabs (now OpNet). He also held positions at the University of California, San Diego and Microsoft Research. Since January 2006, he has been working as a tenure-track assistant professor at the School of Computer and Communications Sciences at EPFL. His dissertation produced Bullet, a heavily referenced system for high bandwidth data dissemination (multimedia streaming and content distribution) for large-scale distributed systems. During the course of his Ph.D., he worked on broad aspects of constructing large-scale distributed systems. At EPFL, he started the Mirage project, an integrated software upgrade testing and distribution system, as well as the GateKeeper project for allowing streaming overlays to adapt to available bandwidth. His work on the CrystalBall project has shown that it is possible to leverage the per-node increases in computational power and bandwidth to predict and prevent inconsistencies in deployed distributed systems. Most recently, he has shown how to reduce the Internet's energy consumption.