SBIR/STTR Award attributes
Dover Microsystems, Inc. (Dover) is working towards a national defense-related mission-need in the area of cybersecurity. Software, despite our best efforts, continues to have errors (aka bugs) that can be exploited by attackers to modify the intended behavior of software-intensive systems. The only two ways to address this problem are to either (1) develop software without exploitable errors, or (2) detect and block attacks when they occur. While preventing errors in the first place is ideal, the history of software has shown that strict processes and extensive testing find many but not all exploitable software flaws. Furthermore, the formal methods to support “correct by construction” (which implies bug-free) are still in their infancy. The five main categories of technology used to detect and block attacks today are: signature-based, anomaly-based, cryptography, compartmentalization, and instrumentation. None of the aforementioned approaches addresses the fundamental problem facing software-intensive systems today: zero-day (not previously encountered) attacks on running systems can redirect execution, modify data, and exfiltrate data. Every system in the Air Force that is currently running software and is connected to a network is at risk of attack. The aforementioned security mechanisms either detect a small subset of possible attacks or merely attempt to limit the scope of damage of a successful attack. All USAF systems planned or in development can benefit from incorporating CoreGuard as a key element of their security. We believe technology development under the subject SBIR topic will contribute to solving a mission need. The main goals of our project are to demonstrate both the precision and the breadth of customizable, hardware-accelerated security policies using CoreGuard. Specifically, we will focus on defining and then monitoring safety automata for safety-critical, software-intensive systems. The mission impact of this project on the Air Force and the Department of Defense will be to add instruction level, fine-grained, non-subvertible, oversight of an execution processor used by the Air Force Nuclear Weapons Center to keep a finite-state machine-based safety critical system from ever entering a dangerous state. This capability addresses a serious and currently unmet gap in the cybersecurity defenses of USAF systems and can save both civilian and warfighter lives by preventing compromise of safety-critical systems.
