Biography:

Dr. Ernest Foo is an active researcher in the area of information and network security.  Dr. Foo has extensive experience with computer networking having worked and taught in this area for over 15 years.  He has been responsible for the design and development of the QUT SCADA security research laboratory.  Recently Dr. Foo has been researching the IEC 61850 standard for controlling and automating electricity sub-stations.  This standard will be key component of the future smart grid.  Dr. Foo has an ARC Linkage grant with the electricity transmission authority Powerlink.  The Linkage project will extend knowledge of cyber security for control systems in electricity sub-stations.

 

 

Abstract:

Industrial control systems have been moving from isolated communications networks to IT networks connected to corporate networks, making it probable that these devices are being exposed to the Internet. Many industrial control systems have been designed with poor or little security features, making them vulnerable to potential attack. This talk will discuss recent incidents of successful cyber attacks against critical infrastructure that are changing the landscape of modern conflict.  Also the talk will discuss recent research that identifies and analyses several feature sets that have been used in studies related to industrial control system communication protocols in order to propose a well-defined initial feature set. We propose a procedure for data pre-processing necessary for machine learning including feature extraction and feature selection to prepare datasets for machine learning IDS experiments. We conduct several experiments to analyse and compare the performance of the introduced feature sets on two machine learning algorithms: Support Vector Machines and Neural Networks. Our analysis show promising results in term of detecting distributed denial of service attacks.