Princeton University

School of Engineering & Applied Science

Rethinking Datacenter Management with Game Theory

Speaker: 
Benjamin Lee, Duke University
Location: 
Computer Science 105
Date/Time: 
Monday, November 6, 2017 - 4:30pm

Abstract:  Sharing datacenter hardware improves energy efficiency, but whether strategic users participate in consolidated systems depends on management policies. Users who dislike allocations may refuse to participate and deploy private, less-efficient systems. We rethink systems management, drawing on game theory to model strategic behavior and incentivize participation. We illustrate this perspective for three fundamental challenges in datacenters.  For workload colocation, we design matching games to produce stable coalitions in which more users share with preferred co-runners. For resource allocation, we use Cobb-Douglas utility functions to produce fair allocations that incentivize users to share cache and memory. For power delivery, we design sprinting games to produce equilibria in which users selfishly draw power for performance boosts yet avoid oversubscribing the shared supply. Collectively, these varied solution concepts provide a rich toolkit for rigorously managing datacenters shared by strategic and competitive users.