Computer Science Faculty Researchers Earn NSF Grant

The grant will allow the researchers to study energy use in virtual environments.
Oct. 7, 2008
Chris Gniady, assistant professor in the UA's computer science department
Chris Gniady, assistant professor in the UA's computer science department
John H. Hartman, associate professor in the UA's computer science department
John H. Hartman, associate professor in the UA's computer science department

University of Arizona computer science faculty members Chris Gniady and John H. Hartman have won a three-year, $300,000 National Science Foundation grant.

Gniady, an assistant professor of computer science, will work with Hartman, an associate professor of computer science, on research focusing on the intersection of energy management and virtual machine environments.

Current energy management algorithms exist in operating systems, and they rely on detailed applications as well as system and device monitoring to help manage devices while also improving their energy efficiency.

What virtual machines tend to do is to make energy management more complicated in several ways.

First, a virtual machine lacks the application context that is available for an operating system.

Also, a guest operating system is confined to its virtual machine, and therefore lacks global information.

Another reason is that virtual machines typically present virtualized devices to the guest operating systems so that a guest operating system no longer has detailed energy profiles of the physical devices.

In their project, Gniady and Hartman's will investigate the proper distribution of functionality between the guest operating system, virtual machine and underlying virtual machine monitor. In doing this, the two researchers hope to improve overall system energy efficiency.

To ensure success, Gniady and Hartman will figure out a way to make the guest operating systems collaborate with the virtual machine monitor, allowing the two to exchange information about current context, the state of the device and other information.

Overcoming such a challenge is critical, the researchers said, since future systems will rely heavily on both energy efficiency to prolong battery life and reduce energy costs, and virtualization to provide isolation and portability, among many other benefits.