Predicting multi-stage attacks based on hybrid approach

Multi-stage attacks can evolve dramatically causing much loss and damage to organisations. These attacks are frequently instigated by exploiting actions, which in isolation are legal and are therefore particularly challenging to detect. Much research has been conducted in the multi-stage detection area, in order to build a framework based on an events correlation approach. This paper proposes a framework that predicts multi-stage attacks based on a hybrid approach, which combines two techniques; IP information evaluation and process query system (PQS). This paper shows the analysis of three multi stage attacks, detailing their steps and information hitherto unexploited in current intrusion detection systems. The paper also goes through the implementation of each technique used in the hybrid approach.