Doowon Lee receives Rackham Predoctoral Fellowship to improve the reliability of computer systems
His research addresses the problem of diminishing reliability in future computer systems.
Doowon Lee, a CSE PhD candidate, has been awarded a Rackham Predoctoral Fellowship to support his research while he completes his dissertation, which is entitled, Low-Cost Comprehensive Robustness for Modern Heterogeneous Systems. The Rackham Predoctoral Fellowship is awarded to outstanding doctoral candidates in the final stages of their program who are unusually creative, ambitious and risk-taking.
Doowon Lee’s research addresses the problem of diminishing reliability in future computer systems. As computer hardware systems are becoming more complex, they are integrating hundreds of different components into monolithic chips. Even with recent advances in computer engineering, these systems are too complex to be completely verified in the short time frame allowed for designing a system, leaving many, if not most, interactions among components unverified. Unverified interactions, unfortunately, are often erroneous, and do jeopardize the entire system. Failing computer systems, unless patched properly, can lead to significant financial losses, tarnish a company’s reputation, and even cause loss of lives.
In his dissertation, he proposes a novel approach to improve the reliability of computer systems. He investigates a unified methodology from two different research areas: fault-tolerant computer architecture and computer hardware verification. While those research areas have traditionally been studied by separate groups of researchers, he observed that comprehensive solutions that consolidate ideas from these two research fields can address many of the reliability problems we expect in future systems.
His observation lies on the fact that in today’s extremely-integrated designs, design bugs may slip into computer chips that are deployed in the field, leading to runtime computation errors that very much resemble failures due to transistor faults. This is true even though the mechanisms of transistor failures, such as transistor wear-outs or alpha particle strikes, are much different from design bugs. More importantly, prior research on fault-tolerant architecture assumes that fault-free designs are essentially bug-free. However, in today’s highly integrated design flows, this assumption does not hold because no design is truly bug-free. Furthermore, both design bugs and faults can be addressed by similar methodologies, such as the one that he is proposing in his dissertation.
His research strives to improve the reliability of the complex interactions among these units, especially in the presence of functional bugs or faults. By completing his dissertation, he expects to be able to reduce the cost of reliability down to a viable 5% overhead.
Lee is advised by Prof. Valeria Bertacco.
About the Rackham Predoctoral Fellowship
The Rackham Predoctoral Fellowship supports outstanding doctoral students who have achieved candidacy and are actively working on dissertation research and writing. They seek to support students working on dissertation that are unusually creative, ambitious and risk-taking.