Research in Autonomic Computing
GrammaTech's research in autonomic computing is focused on the creation of systems that can monitor, protect, and defend themselves without human intervention. A number of GrammaTech technologies work together under this umbrella to enable complex systems to police themselves automatically.
Cyber Grand Challenge
In the Cyber Grand Challenge, 104 teams from around the world began the competition to develop a security system capable of automatically defending against cyber-attacks as fast as they are launched. GrammaTech was part of a smaller group of 7 teams selected to receive funding from DARPA to develop automated network defense technology for the challenge, and is now part of a group of 7 finalists, all of whom will compete to win the competition at DEF CON 2016 in Las Vegas.
The Cyber Grand Challenge is aimed at solving a major cyber-security issue that we are starting to face with alarming frequency – the reliance on expert programmers to uncover and repair weaknesses in an attacked system. Repairing weaknesses only after the system has been attacked, and after hackers have fully taken advantage of these weaknesses to steal data or otherwise impact processes, is dangerous for any system.
GrammaTech is working to solve this problem, while collaborating with the University of Virginia. Our system will provide automatic and adaptive protection of a network service (implemented as an x86 binary) and automatically evaluate network defenses by generating proofs of vulnerability. The system includes breakthrough technology for automated analysis, repair, and protection of binaries and an autonomous cyber reasoning component that dynamically adapts, adjusting resource allocation in response to evolving circumstances.
The system will leverage many innovative technologies previously developed by GrammaTech Research, including: CodeSonar’s binary analysis technology, PEASOUP, and Neptune.
The U.S. Defense Advanced Research Projects Agency (DARPA)
Multi-Abstractions System Reasoning Infrastructure toward Achieving Addaptive Computing Systems (SyRes)
The complexity of modern computer systems has grown to the point of stressing human ability to understand their behavior completely. The sheer number of software components (and the myriad interactions between them) that are present on a single desktop computer presents a difficult security challenge that continues to confound modern protection technologies. Every day, new exploits are created that take advantage of obscure combinations of software bugs and unexpected behavior to sidestep existing defenses. Response is slow, requiring human effort to diagnose and develop new counter-measures to each new threat.
GrammaTech envisions a new paradigm for building system security that endows the computer itself with the tools to diagnose new attacks, reason about their impact to the system, and implement countermeasures in an automatic fashion. Autonomy-oriented computation offers the potential to allow complex computer systems to police themselves, detecting intrusion, performing self-healing, and directly countering cyber threats.
Computer security is critical to both national security as well as the private sector. This technology will provide next-generation technology for creating autonomic systems capable of detecting and closing breaches in security in an automated fashion.
The U.S. Office of Naval Research
Progressive Model Generation for Adaptive Resilient System Software
Today’s software applications, expected to operate with minimal human oversight, have the potential to act erratically after being hacked, exposing critical data or harming physical equipment and humans. To solve this issue, future autonomic ‘intelligent’ software applications will operate with self-awareness. Such autonomic applications monitor for deviations from expected behavior, assess mission health, and react to preserve objectives – raising alerts, initiating recovery processes, or shutting the system down. The Scenario-Based Modeling and Checking (SMAC) project is aimed at helping software developers efficiently create models of ‘correct behavior’ that their applications will be required to follow.
Most systems are just implemented in code. A central problem in this approach to cyber security is defining a precise model that lets the monitor distinguish good behavior from bad. When the implementation is complete, there is no higher-level description of correct behavior that can serve as a model. The essence of GrammaTech's approach is to help programmers define the models while they are coding. This will facilitate ‘smart monitoring’ that will allow the Navy, DoD, and other software users to deploy systems with more confidence that their systems will survive attacks and behave as intended.