Computer systems are increasingly located in places, where they can be physically accessed by people who are not authorized to have unrestricted control of and access to these systems. Examples of this include company laptops carried by employees, robotic vehicles, gaming consoles with copyright protection mechanisms, and remote users of servers in data-centers. An emerging threat in such scenarios are hardware attacks, where snooping devices are attached to the system to directly read and/or modify data within the system. These attacks can circumvent all traditional security protections in the system, such as password checks, access permissions for data files, etc. To address these threats, researchers and processor makers have proposed or developed various types of secure processor architectures, which encrypt and continuously verify data in the system?s memory. However, these secure processors are far from being ready for widespread use, primarily because they focus on a single-processor, non-mobile system that is already up and running and executing a single application.
This research project will investigate how to overcome these limitations, focusing on mechanisms for secure boot-up and system configuration, secure communication between and migration of applications, secure access to peripheral devices (including the network). In essence, this project will build the intellectual framework that will be needed to make computers secure regardless of their physical location, preventing unauthorized access even if the system is captured, stolen, or actually owned by a potentially malicious entity. Other broader impacts of this project include improvements in education and workforce, by making computer hardware designers more aware of physical security and by making computer security experts more aware of implications of physical (hardware) attacks.
