Skip Navigation

Colloquium Details

The Judgment of Forseti: Economic Utility for Dynamic Heap Sizing of Multiple Runtimes

Author:Jeremy Singer and David White University of Glasgow, Scotland
Date:June 12, 2015
Time:13:30
Location:220 Deschutes

Abstract

We introduce the Forseti system, which is a principled approach for holistic memory management. It permits a sysadmin to specify the total physical memory resource that may be shared between all concurrent virtual machines on a physical node, e.g. in a datacenter. Forseti seeks to maximize the combined throughput of the set of VMs based on concepts from economic utility theory. Our results show that Forseti enables dramatic reductions (up to 5x) in VM memory footprint without compromising application execution times.

Biography

Dr. Jeremy Singer is a lecturer in the School of Computing Science. His particular focus is in the area of complex systems engineering. His current research investigates the application of mathematical models to runtime system behaviour.

Previously, Jeremy was a postdoctoral researcher at the University of Manchester, where he worked on Java runtime systems for speculative parallelism on multicore architectures. Jeremy obtained his PhD in 2006 from the University of Cambridge. His thesis topic was Static Program Analysis based on Virtual Register Renaming. Jeremy's research interests include memory management, garbage collection, compilers, runtime parallelization, and software support for manycore parallelism. Jeremy is a member of the ACM.

Dr. David R White is a SICSA Research Fellow in Complex Systems within the School of Computing at the University of Glasgow. His work includes memory management, the creation and optimisation of software using heuristic search and evolutionary computing, and new ways of researching and teaching cloud computing. He received his PhD in Computer Science from the University of York in 2010. His research interests include memory management, garbage collection, genetic programming, evolutionary computation, and cloud computing.