MASPTA-O: Multiagent System Planning to Avoid Threats Optimally
Vandana Gandotra, Archana Singhal, Punam Bedi · Information Security Journal A Global Perspective · 2011
Extensive use of modern technologies with expanding connectivity has created high risk threats requiring innovative strategies to counter them. In the literature, multiagent system planning has been used for threat avoidance where agents are employed in all the identified attack paths to avert the threat. This paper presents the two-level optimal approach MASPTA-O, or Multiagent System Planning to Avoid Threats Optimally. The first level identifies optimum number of threats to be mitigated using the Layered Threat Elimination model. At the second level, each identified threat is optimally mitigated by generating various mitigation plans using Mitigation Plan Generation algorithm and selecting minimum number of paths to avoid the threat. Agents are then inducted in these minimum number of attacker's paths to mitigate the threats. In the proposed work, a Master Agent has been introduced to continuously monitor the performance of the active agents for any abnormality and provides an alert signal to save the system from being compromised. Experimental study using the Java Agent Development Environment (JADE) has been conducted to test the proposed approach.