Dr. Alexander Stolyar (Bell Labs, Alcatel-Lucent)
Room 333-Fishbowl (WERC)
The model is motivated by the problem of efficient “packing” of virtual machines into physical host machines in a network cloud data center. There is infinite number of servers (physical machines) and multiple flows of arriving customers (virtual machines) of different types. Each server can simultaneously serve several customers, subject to some “packing” constraints. Service times of different customers are independent — even if customers share a server. Customers leave after their service is complete. The underlying objective is to minimize the number of occupied servers. We show that some versions of a greedy packing strategy are asymptotically optimal as the system scale (the average total number of customers in service) goes to infinity.
Bio: Alexander Stolyar is a Distinguished Member of Technical Staff in the Industrial Mathematics and OR Dept. at Bell Labs Research (Murray Hill, New Jersey). His research interests are in stochastic processes, queueing theory, and stochastic modeling of communication and service systems. He received Ph.D. in Mathematics from the Institute of Control Science, USSR Acad. of Science, Moscow, 1989, and was a research scientist at the Institute of Control Science in 1989-1991. From 1992 to 1998 he was working on stochastic models in telecommunications at Motorola and AT&T Research. He joined Bell Labs in 1998, where he has been working on stochastic networks and resource allocation problems in a variety of applications, including wireless systems. He received INFORMS Applied Probability Society 2004 Best Publication award, SIGMETRICS’96 Best Paper award. He currently serves on editorial boards of Advances in Applied Probability and Queueing Systems.
Host: Dr. Kumar