SCS, ECE Scores Best Paper Award


Devin M. Young

Communications Assistant

Sidebar Content
No sidebar content submitted.

Summary Sentence:

Research Paper Featuring Researchers from the School of Computer Science and School of Electrical and Computer Engineering Wins Coveted Best Paper Award.

Full Summary:

No summary paragraph submitted.

  • Iot Side Channel Chip Iot Side Channel Chip

Georgia Institute of Technology researchers have won a best paper award during the 49th Annual IEEE/ACM International Symposium on Microarchitecture (MICRO-49), a top-tier conference in computer architecture.

The paper, "Spectral Profiling: Observer-Effect-Free Profiling by Monitoring EM Emanations," was authored by third-year School of Computer Science (SCS) Ph.D. students Nader Sehatbakhsh and Alireza Nazari with support from SCS Professor Milos Prvulovic and School of Electrical and Computer Engineering Assistant Professor Alenka Zajic.

In a deviation from tradition, the highly-discerning best paper committee selected two papers to receive its annual best paper award during this year’s conference due to the high quality of research presented in both of the winning papers.

“The paper acceptance process for MICRO is very selective,” said Prvulovic. “I have read hundreds of MICRO papers over the years, and I can honestly say that it is a privilege just getting to present your paper at MICRO. So, being selected for the best paper award is incredibly humbling.”

The paper describes what the researchers call “spectral profiling,” a new method for program profiling, a technical term that refers to identifying which parts of the application’s code are responsible for most of its running time. Unlike past profiling approaches that rely on changing the application itself to record when certain parts of its code are executing, “spectral profiling” works without changing the profiled application in any way by receiving and analyzing electromagnetic (EM) emanations unintentionally produced by the profiled system.

Through a two-phase implementation, the team’s method first trains the profiler with application information that is already known. This allows the spectral profiler to learn which signals correspond to which part of the application. From there, the profiler can monitor the same application and precisely identify which part of the application is executing at any given time. This lets the team know how much of the overall time is spent in each part of the application.

This research is the first work of its kind and is supported in part through funding from the Air Force Office of Scientific Research and the National Science Foundation. Recently, Prvulovic and Zajic received a $9.4 million DARPA grant to further their research in this area.

Additional Information


College of Computing, School of Computer Science

Education, Student and Faculty, Student Research, Research, Computer Science/Information Technology and Security
Related Core Research Areas
Cybersecurity, National Security
Newsroom Topics
No newsroom topics were selected.
Computer Science, engineering, SCS, ECE, School of Computer Science, School of Electrical and Computer Engineering, Milos Prvulovic, Alenka Zajic
  • Created By: Devin Young
  • Workflow Status: Published
  • Created On: Dec 1, 2016 - 3:10pm
  • Last Updated: Dec 9, 2016 - 10:38am