Jakub Szefer Wins NSF Grant for Cloud Computing Technology

11/01/2019

For his proposal to secure Field-Programmable Gate Arrays (FPGAs) for use in cloud computing, Jakub Szefer, assistant professor of electrical engineering and computer science has received a four-year grant for $457,160 from the National Science Foundation (NSF). It is part of a larger $1,000,000 NSF-funded project in collaboration with University of Massachusetts, Amherst. 

The general public can now easily get access to large-scale computer systems through what's known as cloud computing to perform challenging computations easily without upfront cost of buying computer hardware and managing the systems. These systems typically use microprocessors, but recently special reconfigurable computer chips called FPGAs have become accessible through Cloud Computing, creating a new so-called cloud FPGA computing paradigm. The FPGAs can allow users to create custom logic circuits for applications such as machine learning, and can be more efficient than using microprocessors. The downside is that cloud FPGAs introduce new security risks which could put users' data at risk.   

“This project explores how to secure FPGAs in distributed large-scale computer servers to prevent information theft and incorrect calculations,” said Szefer.

Szefer’s project begins by assessing how much of a risk malicious FPGA use poses, whether it’s data snooping, data leaking, or deliberate attempts by users to degrade or damage the FPGA infrastructure. It will also explore potential security vulnerabilities and demonstrate their expected impact. Then, to mitigate these vulnerabilities, Szefer and his collaborators will develop a monitoring system designed to continuously check the FPGAs, detect attacks, and limit malicious behavior at run-time. At the same time, it will compile-time rule checks will also be implemented to help detect malicious designs.

This layered approach will allow for truly secure cloud FPGAs, allowing sensitive data to be processed securely and cost-efficiently. Further, this work will enable multiple users to share the FPGAs in Cloud Computing data centers, creating secure multi-tenant FPGAs.

Szefer will also develop new courses on data centers and FPGA security that focus on scalable real-world systems, as well as a student-run research workshop to allow students to share research ideas and their work. 

“These activities teach undergraduate and graduate students how to leverage Cloud FPGA resources for a variety of hardware and software experiments,” he said.