Internship Programmes
 
Traffic Characterisation for Embedded Networks

Discipline: Computer Engineering/Computer Science/Electrical Engineering preferred

Project Description:
Background.  Increasingly, TCP/IP-enabled devices are deployed in control area networks.  These embedded networks could be used for data collection or remote controls.   Due to the lack of resources, embedded networks typically communicate in unsecured protocols.  During cyber security assessment, one important question is the amount of information that can be derived from the network given a finite amount of time.  Also, embedded devices that exhibit abnormal network behavior be an indication that it has been compromised. 

Objectives.  The objective of this project is to identify important data points using some form of machine learning or data-mining techniques and demonstrate the technique via a proof-of-concept.  This study involves identification and characterization of critical control messages and the relative importance of data points.  The study shall also investigate & validate the possible features (such as network interaction, frequency, etc) that may be suggestive of possible attacks.

Potential Benefits to Intern

1) Hands-on exposure to distributed control systems which are commonly used in real-world critical infrastructure.
2) Understanding security concerns & mitigating measures of real-time control systems.
3) Opportunities to apply computer science techniques concepts (e.g. data-mining, machine-learning techniques) to real life problem.
4) Deserving candidates who successfully demonstrate the concepts/prototype would be funded to attend a relevant overseas/local conference.

Pre-requisite:
Non-technical
- Natural sense of curiosity, like to spend time outside curriculum requirements to explore how things works
- Self-driven to learn things that are not taught
- Strong independent thinking

Technical
- Familiarity with TCP/IP network programming
- Knowledge of network analysis tools (such as Wireshark)
- Good skills in C, PERL or other system programming language