VMware Funds Coskun’s Server Consolidation Research


Assistant Professor Ayse Coskun (ECE)
Assistant Professor Ayse Coskun (ECE)

Energy consumption is a growing problem for data centers and high performance computing (HPC) clusters. Even some of the best methods for saving energy have their flaws.

Take the popular method of server consolidation, where resources are shared across multiple applications and virtual CPUs. Without careful optimization, consolidation may result in high temperatures that require better cooling, extended runtimes of applications, and reliability problems that require additional computations – all of which add to the energy cost.

Professor Ayse Coskun (ECE) believes this can be improved and VMware, a leading provider in virtualization software, recently awarded her $67K for one year of research on her project, “Energy-Efficient and Reliable Server Consolidation for the Cloud.” Additionally, VMware may opt to renew this funding after the first year.

In cloud computing, any user connected to the Internet can have access to vast computational or storage resources through a service provider, without the need of physically owning these resources. The backbone of cloud computing is a large set of data centers and HPC clusters.

Coskun aims to improve the energy efficiency of these shared systems by enabling performance-awareness, better thermal control, and proactive reliability management during server consolidation.

“Our ultimate goal is to achieve significant energy efficiency and tunable performance-energy tradeoffs in the cloud,” said Coskun.

VMware is a very appropriate partner for this project, said Coskun, because the work will be highly relevant to the company’s research interests, that include coordinated resource management, power management, and providing the fundamental tools and techniques to enable real-time monitoring of systems.

“We expect the results of this research to help new product and software development immediately and to provide valuable insight for future research in cloud computing in the longer run,” said Coskun.

-Rachel Harrington (rachelah@bu.edu)