Laura Keys laurak {AT} eecs.berkeley {DOT} edu RAD Lab, 4th Floor Soda Hall UC Berkeley Berkeley, CA

I am a second-year Computer Science PhD student in the UCBerkeley EECS department, advised by Randy H. Katz. My research interests include energy-efficient and energy-proportional computing, and subsequently, SSDs and flash memory. I am involved with the LoCal research group and sit in the RadLab.

Research
I am currently working with Yanpei Chen on an energy-efficient MapReduce and using Hadoop to better understand performance/power tradeoffs in clusters of machines. Additionally, I will be working on finding the right workload to do power predictions for data centers from and characterizing workloads in terms of benchmarks like SpecPower, on my way towards a Masters degree.

I am in the process of finishing up from a summer interning at Microsoft Research SVC with John Davis on a project that involved predicting power usage for multi-machine systems from single-machine runs.

Other CS-related interests include Linux, PERL, and playing with hardware (and pretending I know what I'm doing!). I am the impromptu sys admin / benevolent dictator of an Atom-processor cluster put together by Ken Lutz. (4 nodes live on my desk, and 12 more are on their way.)

Non-CS interests include dancing (ballet, swing, salsa, tango, etc.) and cooking. I also teach geometry once a week at San Quentin prison with the Prison University Project.

Publications / Scribblings

• Notes from the Energy Star Server Workshop
• Towards Energy Efficient MapReduce (Tech Report)

Class Projects

• ClickCAM: Modeling power/performance tradeoffs for TCAMs with the Click software router (Spring 2009)
• IotaFS: Examining performance of simple file systems of old in combination with SSDs of new (Fall 2008)

Presentations

• End of Internship talk at MSR (e-mail me if you're interested in this one)
• Toward Energy Efficient Hadoop: Presentation at Hadoop Summit 09 with Yanpei (Janpei?) Chen