Autonomic Knowledge Representation Construction for Software Protection Systems

The increased complexity, scale and capabilities of current software elements make current security tools and technologies incapable to protect software against attacks. It also makes it very hard to proactively detect their vulnerabilities that might be exploited by attackers. Our focus will be to investigate and study the construction of a near human knowledge representation of the state of a system that is running an application of interest, in order to protect that application from reverse engineering and tampering attacks. The underlying assumption here is that through continuously monitoring a correct and sufficient set of applications, system resources, and communications metrics, the AKRCSPS will be able to determine any hostile behavior. Once the hostile behavior is detected and characterized, the AKRCSPS will plan what reaction is to be taken to perform protection and countermeasure actions. Furthermore, through the use of autonomous agents, we can destroy the ability of the attacker in launching attacks against our system. The capabilities that will be developed in Phase I of this project will be critical to construct the autonomic representation knowledge about the software’s state and defend against software attacks to achieve superiority for current and future software protection system.