Attack graph analysis using parallel algorithm

Each attack graph shows a set of scenarios of penetrating a computer network. A penetration scenario actually defines the order of steps that an intrudor should take to achieve his goal, and each step is characterized to show which host must be abused. A specified weight is attributed to every abuse in each stpes by analyst. The atributted weight is proportional to a required cost to prevent the abuse. In this paper disadvantages of a network are gathered in a graph and analysed by a parallel algorithm. In a way that allthe available path in the graph which represent successful scenarios processed concurrently, and from each path a node with the least possible cost is selected which correspondes to an abuse. Eventually we have minimal set of disadvantages with the least prevention cost, and by preventing the happening of this mininal set no penetration scenario is applicable on our analytical network

Keywords:

attack graph , network graph , parallel algorithm