Samira Khan Earns National Science Foundation Grant to Create New Networks Fast Enough to Maneuver Big Data
The National Science Foundation places a priority on inventing new computing and networking technologies. The need is urgent, because such technologies will help researchers use big data sets to find solutions for complex global challenges.
The problem is that the amount of data available globally has outpaced the processing power needed to analyze it. International Data Corporation predicts that the collective sum of the world's data will grow to 175 zettabytes – 175 trillion gigabytes – by 2025, a massive data explosion compared to 4.4 zettabytes available in 2015.
Samira Khan, assistant professor in the Department of Computer Science at the University of Virginia School of Engineering, is developing revolutionary computer architectures that will make problem-solving with big data possible. She has received a CAREER Award from the National Science Foundation in recognition of her groundbreaking research.
“Datasets are so large they must be broken up into bundles across multiple computers in a data center,” Khan said. “Computations get bottlenecked as larger and larger data packets get moved from computer to computer in progression to a single processor.”
Khan’s research aims to redesign programmable switches and smart network interface cards to allow data to be processed in transit instead, a fundamental redesign of outdated computer infrastructure. Her research team has built the first protype network that uses the revolutionary architecture, making data requests four times faster. In the real world, this would mean people could update their social media or make online transactions, like purchasing tickets, lightning fast compared to today.
“Reducing the amount of data that needs to be moved to that single point of processing dramatically speeds things up and fuels the entire system’s capacity,” Khan said. “We are expecting that processing in the reconfigured network will achieve more than 10-fold increases in processing speeds for scientific and machine-learning workloads.”
Khan is presenting the prototype this month at the annual International Symposium on Computer Architecture conference, where computer scientists from around the world meet to discuss the most important advancements in their field.
"Samira’s work will enable dramatic improvements in processing efficiency, which will have profound implications for discoveries in health care and many other human endeavors, such as smart cities and autonomous transportation.” said Kevin Skadron, Harry Douglas Forsyth Professor and chair of UVA Engineering’s Department of Computer Science. “She is forging entirely new system architectures to support the scale of computation needed for analyzing massive amounts of data.”
An undergraduate project at the Bangladesh University of Engineering and Technology—where Khan earned a bachelor’s degree in computer science and engineering—inspired her to choose computer hardware as a career path.
“I was involved in a microprocessor project where we built a CPU. It was so different than projects that only involved software,” Khan said. “When you are redesigning memory and network components, there is a hands-on aspect of the engineering where you are actually connecting and disconnecting parts. I was drawn to the tactile aspects of the research.”
Khan went on to earn a Ph.D. in computer architecture from University of Texas at San Antonio in 2012, and completed her post-doctoral research at Carnegie Mellon University in 2015. That same year, she joined the UVA Engineering Department of Computer Science.
Khan was tapped in 2018 to oversee research contributing to the UVA-led, 11-university Center for Research in Intelligent Storage and Processing in Memory, or CRISP. The $29.7 million research effort focuses on removing the “memory wall,” in which data gets bottlenecked and causes performance problems in outdated systems.
As an expert in computer architecture and its implications for software systems, Khan guides the center’s efforts in rethinking how the many layers of hardware and software in current computer systems work together.
Khan shares her expertise with future engineering leaders in an undergraduate Computer Architecture course in the School of Engineering. Through this CAREER Award, which also recognizes her as a role model in education, she is piloting a new experiential learning curriculum that is a real-world simulation in how to speed up a computer network using nano drones.
Students in the program operate nano drones that gather data about the same environment. The nano drone data is sent to a single computer that uses all of it to create one three-dimensional map. By incrementally moving successively larger data packets to be processed in transit—rather than waiting for all of it to be processed at the computer—students can quantify the direct relationship between data movement and network efficiency.
The Spring 2021 pilot was targeted to UVA undergraduates, but Khan plans to offer the program to local high school students too when she expands it into an annual summer program.
The prestigious National Science Foundation CAREER Award is one of six early career accolades Khan has received since 2012. While completing her doctoral work at Carnegie Mellon, she was selected to present at the Rising Stars Workshop for women in electrical engineering and computer science. She also received the National Science Foundation Grant Opportunities for Academic Liaison with Industry, or GOALI, Award.
Since joining UVA Engineering’s Department of Computer Science, Khan has received the National Science Foundation Computer and Information Science and Engineering Research Initiation Initiative, or CRII, Award, the National Science Foundation Scalable Parallelism in the Extreme, or SPX, Grant and a Google 2020 Rising Faculty Award.