A Survey on Methodologies for Multi-Step Attack Prediction

Abstract

Multi-step or multi-stage attacks are those attacks that occur in a series of steps. In these types of attacks, a single system can be compromised in a number of stages. These stages are nothing but a path followed by the attacker to reach the target system. The Intrusion Detection Systems (IDS) are not enough for predicting these types of attacks. In this survey, we have reviewed the most commonly used methodologies for the prediction of multi-step attacks. We have also compared these methodologies based on the performance metrics that remains essential for prediction. The performance metrics that are taken into consideration are efficiency, complexity, detection/prediction time, new attack mining, scalability, and simplicity. This survey will guide the researchers in choosing the most suitable methodology for different multi-step attack scenarios.