Abstract: In the last decade there has been an explosion in the computing power and capabilities of both personal electronics and data centers. This has come thanks to the huge parallelization of the workloads. In order to take full advantage of parallelization, a central design problem is how to balance the workload among a large number of parallel servers, that is, how to decide which server should process each incoming job. Due to the massive scale of these systems, resource usage in the implementation of load balancing policies has emerged as a critical concern besides conventional performance metrics. In this talk we explore fundamental tradeoffs between performance (in terms of the delay of a typical job and the stability region of the policy), and implementation resources (in terms of communication overhead and memory usage).

Bio: I am a postdoc in the Department of Mathematics and Computer Science at the Eindhoven University of Technology, hosted by Sem Borst, and in the Korteweg-de Vries Institute for Mathematics at the University of Amsterdam, hosted by Michel Mandjes. Born in Montevideo, Uruguay, I received a B.S. degree in Electronics Engineering (2012) and a M.Sc. degree in Engineering (2014) from the Universidad ORT Uruguay, and a Ph.D. in Electrical Engineering (2019) from the Massachusetts Institute of Technology, under the supervision of David Gamarnik and John Tsitsiklis.
My research primarily focuses on the modeling, analysis, and control of large-scale stochastic decision systems, inspired by applications in computer networks and other service systems. I am particularly interested in exploring the fundamental tradeoffs between performance and efficiency that arise in such systems, with a special emphasis on the role that information plays in these tradeoffs.