Less than 2 percent of all cars sold in America today are electric, but that will soon change. The internal combustion engine’s firm grip on American transportation is loosening and for the first time... Read more about Prepare the grid now to power electric vehicles
Welcome Christina Delimitrou
- New Faculty Year: 2016
Christina Delimitrou, assistant professor in Cornell’s School of Electrical and Computer Engineering (ECE), did not have a computer until she was 17. Now she spends much of her time working to improve the design and management of large-scale datacenters. Delimitrou joined the faculty at Cornell in the summer of 2016. She is the John and Norma Balen Sesquicentennial Faculty Fellow and a member of the Computer Systems Laboratory (CSL) at Cornell.
Delimitrou’s main interests are in computer architecture and computer systems. Specifically, she works to improve the resource efficiency of large-scale datacenters through Quality-of-Service (QoS)-aware scheduling and resource management techniques. Delimitrou is also interested in designing efficient server architectures, distributed performance debugging, and cloud security.
Delimitrou grew up in Larissa, Greece and was interested in math, physics, and chemistry in high school. She attended the National Technical University of Athens and earned her undergraduate degree in Electrical and Computer Engineering. Her mentor in Athens, Professor Nectarios Koziris, helped Delimitrou in her search for a graduate program, which led her to Stanford University, where she earned both a Master’s and a Ph.D. in electrical engineering.
Delimitrou’s focus on making large-scale computing more efficient started just as she was finishing her undergraduate diploma thesis. Fortunately for Delimitrou, her Ph.D. advisor at Stanford was developing a strong interest in the same challenge at the same time. While working on her Ph.D. with Professor Christos Kozyrakis, Delimitrou built practical systems for cluster management and scheduling in warehouse-scale computers. “With resource efficient cloud computing, my goal is to get as much compute capabilities as we can from existing systems,” says Delimitrou, “without sacrificing performance or building new datacenters.”
Delimitrou knew early in her school career that she wanted to be an academic. “I have always had in the back of my mind that I would like to be a professor,” says Delimitrou. “I love working with students, watching them become experts in a field they did not know much about initially, and hopefully have a positive impact on their careers the way my mentors had on mine. I also like the freedom academia allows to define the problems I am working on.” She came to Cornell because of the quality of its students and faculty, its reputation for collaboration, and for the chance to be part of the Computer Systems Laboratory (CSL).
“The CSL was an important draw for me,” says Delimitrou. “There are colleagues here working on everything from architecture to distributed systems and my work includes the whole range, which presents a lot of opportunities for collaboration.” Delimitrou sees now as the perfect time to bring people in various computer fields together. “I think we can change the way things are done—rather than each field doing its own thing in a vacuum. This approach leads to inefficiencies. Instead, if we understand each piece and see where there is room to make computing more efficient we can provide better performance, increased responsiveness, and more computational capabilities for the same cost.”
In addition to the energy, time, and cost savings that accrue from making existing computer systems more efficient, Delimitrou sees an added beneficial side-effect. Systems that were working at capacity now have excess capacity that can be used to run other applications that might be used for the greater good. The example Delimitrou gives is one she recently started having an impact on. “We use the excess system capacity that resulted from our work and we overlaid international shipping routes in the Pacific Ocean with known whale migration routes. We are starting to find ways that we can adjust routes such that they keep goods moving efficiently and also protect whales.”
This fall, Delimitrou is teaching a graduate level ECE course in cloud computing. When she is not at work, Delimitrou plays piano and electric guitar, and also enjoys painting miniatures and canvases.