Detail

Learning Attack Sequences and Pre- and Post-Conditions for Attack Graphs

An attack graph captures the sequences of attack steps that will lead to a breach in security in a network. Traditionally attack graphs are constructed and analyzed by hand by security experts, but such graphs can also be automatically generated, provided the pre- and post-conditions of each attack step are known a priori. This process is prone to error and omissions, time consuming, and cannot respond to novel attacks that have not been pre-encoded. We will discuss a method to automatically learn attack sequences and pre- and post-conditions of attack steps from network traffic data.