Mitigating dynamic attacks using multi-agent game-theoretic techniques
Mahsa Emami-Taba, Mehdi Amoui, Ladan Tahvildari · Computer Science and Software Engineering · 2014
Emerging technologies such as mobile and cloud computing have given rise to new security vulnerabilities and challenges. At the same time, attackers utilize these technologies to initiate sophisticated attacks and exploit known and unknown vulnerabilities. A unique characteristic of recent attacks is their dynamic nature which allows attackers to stay stealth from Intrusion Detection Systems (IDSs). The proactive and dynamic nature of these security attacks make their detection and consequently their mitigation challenging. This demands fast reacting adaptive systems that are capable of detecting and mitigating threats on the fly. Our novel approach aims at addressing this demand by engineering a Self-Protecting Software (SPS) that incorporates attacker's possible strategies when selecting countermeasures. To achieve this goal, we propose a technique to model objectives of the attacker and SPS by the aid of multi-agent concepts, and utilize game theory to model the competition between the adaptation manager in SPS and the attacker.