Masters Thesis

Machine Learning Method for Cyber Security Intrusion Detection for Industrial Control Systems (ICSS)

Industrial Control Systems, or ICSs, are modernizing everyday with new advances in technology. This modernization is being done while old traditions of air gapping, as the main line of defense from security attacks, is still maintained. ICSs are modernizing to new topologies in which devices are being connected directly to the outside world. Unfortunately, this leads to the idea of air gapping to fail. New Security methods are needed to help modernizing ICSs protect themselves from attacks. This work introduces the idea of an intelligent Intrusion Detection System, in which Snort and a machine learning model is created to help detect anomalies in an ICS. In this work, I implement a simple intelligent IDS using Snort and a Support Vector Machine. This intelligent IDS is then implemented into a modern Industrial Internet of Things based topology for an ICS. This implementation is rather slow when training but accurate when detecting anomalies in the ICS. This implementation shows how feasible an Intelligent IDS can be as security for an ICS.

Le relazioni