Getting to Know: Professor Guru Prasadh Venkataramani
Dr. Venkataramani conducts research in computer architecture with an emphasis on security and energy efficiency, and his research group has made pioneering contributions in investigating hardware timing channels on multicore processors. His research has been funded by the National Science Foundation, SRC, the Office of Naval Research, and Oak Ridge Associated Universities, and he has received an NSF CAREER award. Dr. Venkataramani joined the ECE department in 2010 and is now a full professor. Department staff member Robert Baden interviewed him recently about his research.
Please describe in simple terms the research that you conduct?
My research is focused on computer architecture and security. Computer architecture is about architecting and designing the next generation of computer processors which are the brains of any modern computer system. Traditionally performance was the primary fact of computer architecture that was researched. Over the last twenty years the focus from shifted from performance to power (energy). Power is an important dimension to consider because there is a finite energy source to do computations. Energy efficiency is a premium and energy hardware computing is an important thing.
The other major dimension that I focus on is security. We’ve been looking at the increasing number of security attacks over the past five years because security is an important and pertinent problem that needs to be addressed with the increased use of shared systems. My research looks at security from the angles of hardware security and software security. The National Science Foundation (NSF) supports my hardware security research that focuses on how to understand information leakage from the hardware perspective, where many of the hardware components are already being shared when applications run on them. My research looks into how one application can do something malicious to infiltrate or observe the secrets of other applications. I organized an NSF workshop on side and covert channels in computing systems that was held at GW, with 60 researchers participating from industry, government and academia. I also work with Office of Naval Research, trying to understand how to discover bugs and vulnerabilities in software applications and network protocols before they manifest. We’ve had success in this particular domain with a best paper award at the Securecom conference. I was recently elected as a European Alliance for Innovation Fellow.
What is the significance of your research in terms of practical application?
Energy efficiency and security are important and trendy problems from the perspective of having to deal with a limited energy source and running your computations efficiently. The amount of computations doubles every two to three years requiring investment into more capable processors that will require more energy. My research deals with sustainable energy-aware computing. We are trying to build hardware and software platforms that are geared towards providing more secured access to use the data. Our group is one of the first to work on hardware covert channels and where hardware processes are subjected to leakage attacks and we have studied methods on how to efficiently overcome this problem.
Which of your research areas is the most interesting or challenging? Why?
Security is the most challenging and most rewarding research field that I work on, because security inherently deals with malicious actors. All it takes is one loophole to be exploited. A defender’s outlook is how to keep everything secured to ensure that there are no loopholes.
How did you become interested in this particular field of research?
I have always loved computers since my high school years. Understanding the design of computers and what delivers high performance has always been fascinating to me. I am also into building things and understanding system designs. It is no surprise where I ended up and when it comes to my background in computer architecture. I had excellent professors and a PhD advisor to teach and train me in computer architecture research. I made sure to be surrounded by the most renowned scholars in the area. I’ve learned a lot from my students as well.
What are the challenges or limitations associated with this particular field of research?
There are a number of unique challenges that only computer architects can address. The landscape of applications has been constantly growing and rapidly outpacing other industries. It is a challenge to keep a finger on where and how the landscape is expanding and it is very unique how challenges arise. When you look at the 1990s it was all about performance, which was the number one challenge, and with the expansion of computing capabilities power supply and management became a pressing problem. In the last 10 years, security has become a major challenge because everyone has access to computers and they’re all interconnected. As a result the number of dimensions that you have to optimize in the processes and designs of computers has become exponentially hard. “How does a computer architect carefully balance the performance needs of users while having to make sure the power efficiency and security need are met?” is indeed a very grand challenge. There isn’t a silver bullet solution for this unfortunately.
What are the goals of your research and what are you hoping to learn/achieve from it?
Research is a constant learning process that is pushing the frontier of what’s known today even further. I view research from the aspect of what can I learn. I’ve always been a fan of how to understand the process of how to efficiently analyze data. Computer designers always look for the next innovative design. I’ve dealt with research by first being aware of the limitations from physics and science, while advancing the state-of-the-art and hopefully down the line, we will see greater design space solutions. My research objective is to obtain a good understanding of security from a quantitative perspective and to find out if this is doable. It has been a slow and painful progress made by several researchers in terms of understating security in quantitative terms. My research goal is finding a quantitative way to define the amount of security.
Is there anything about your research that makes you think it is unique compared to other research being conducted in your area?
My research is collaborative and I am interested in research as it pertains to related domains outside computer architecture. I’ve taken the liberty to talk to people in different fields from electrical engineering and beyond to understand the depth of the problem from several dimensions. Security overlaps several fields from my domain and involves other aspects. There are softer aspects that involve an in-depth combination of hardware and software. There is a need for pattern of recognition where signal processing can be a dimension added to several detection algorithms for malicious attacks. What goes into the behavior of the attacks to speed it up with some automation with machine learning? Several researchers have worked on several different areas that overlap in the security domain. Processor design needs a combination of physicists who look at materials and properties, material scientists who make manufacturing possible, electrical engineers who build the circuits, computer engineers who architect, and computer scientists who write applications to increase application performance.
What successes or milestones have you reaped thus far in your research?
Our group pioneered the first ideas in understanding how information leakage channels can be constructed in hardware and how modern processers that have shared hardware can contribute to information leakage spilling sensitive secrets belonging to users to unwanted parties. My former PhD student, Jie Chen, and I have received some recognition for this research. I have also received prestigious awards including the NSF Career Award. I’ve been fortunate to work with several collaborators and students and I’ve been blessed with a lot of good problems. Fantastic research problems are the key to making an impact in the real world.
With whom do you collaborate, and who funds your research?
I work with Dr. Suresh Subramaniam on data center energy efficiency, Dr. Milos Doroslovacki on hardware security, and Dr. Tian Lan on software security. My primary funding agencies are the National Science Foundation, the Office of Naval Research, and Semiconductor Research Corporation.
What role do GW students play in your research and what do you gain from their involvement?
GW PhD students have been the central part of defining my research career, and on the teaching front it is always exciting to meet students who are full of energy and questions. The students keep it very engaging during discussions in class. I teach computer architecture at the undergraduate and graduate levels and secure computer architecture. GW students have been a central reason why I have enjoyed my career so far. I encourage questions since research doesn’t progress if you don’t ask questions. As a result I’ve learned a lot from them, and they have been the catalyst behind everything that I have learned and will keep learning. I have gained tremendously from the hard work that the students have plugged in. They’ve produced a tremendous number of research papers and have made a pioneering set of contributions in top venues such as the IEEE/ACM Micro and IEEE HPCA conferences.