Hardware-Based Security Laboratory

Project Summary

Securing computation persists in being a significant unsolved hard problem in our nation's information infrastructure. A simple look at history---or the most recent issues of BugTraq or even The New York Times---show that, over and over again, society cannot manage to build and deploy computing applications that actually are secure.

When a problem persists in being unsolvable, it's time to consider changing the problem. In this case, an inescapable fact of computation is that it must take place on computing hardware. Consequently, a promising approach to making this hard problem easier is to change this basic hardware. This idea is not just a pie-in-the-sky lab dream, but rather is something coming in the next wave of real systems. Trusted Platform Modules (TPMs) are already shipping, and the Trusted Computing Group (TCG) consortium continues to crank out new specifications; Intel will be shipping CPUs enabling virtualization (the VT chipset) and secure hypervisors (LT); AMD has its own alternatives. IBM is shipping the multicore CELL processor that uses hardware structure to protect user processes from malicious kernels; Intel promises that multicore will soon give us more processors at the client than we'll know what to do with.

• Principal Investigator: Sean Smith, Department of Computer Science
• Project Researchers
• Project Publications
  
• Funding: Department of Homeland Security, National Cyber Security Division (DHS-NCSD)