The Division of Cyber Security is one of three Divisions in the School of Design and Informatics.
We have developed strong links with government, Police and industry to collaborate on a range of research projects including improving the security of SMEs, training in cybercrime response using games technologies, and cybersecurity into the Software Development Lifecycle. Research in the Division is a major part of the Security Research Theme and is structured into four overlapping areas: responding to prevailing challenges of system security; vulnerability detection and the threats introduced by Internet-connected devices; the usability of security prevention measures; and using visualisation techniques to improve security.
Too often, the areas of computer security and software development are regarded separately, each with minimal consideration for the other. However, a large number of security flaws are caused by a limited understanding of how poor coding practices can be exploited by a malicious hacker. In partnership with industry, we are exploiting design patterns, a well-established approach to promote best practice in software engineering, to increase secure coding awareness in software developers, through all aspects of the Software Development Life Cycle.
We have analysed and catalogued security threats and vulnerabilities in order to better understand their root cause and identify appropriate techniques to improve communication of security problems. This combination of root cause and the security problem are encapsulated in vulnerability anti-patterns as a means to transfer knowledge from the cybersecurity community to the software development community to ensure that secure software is developed from the outset. Our findings show that these anti-patterns can improve software developers’ ability to recognise vulnerabilities in systems and how they can be exploited. Further research will measure the longer term impact of this improved awareness on the overall security of the developers’ software.
Any computer network is potentially vulnerable to cyber attacks. Every network has an attack surface, i.e., the set of devices on the network and the ways in which the surface may be attacked, and threats can occur at any point on that landscape. In addition to this general threat, many common household devices are now being connected to the Internet, and newly developed devices are also being introduced into people's homes. The number of devices connected to the Internet is expected to reach 50 billion by 2020. In a number of cases, these devices have security flaws that can compromise the privacy of the owners, or can be subverted to be used as a means to attack other systems.
We are developing novel solutions to effect pervasive security and privacy for networks in general and IoT devices in particular. For example, we have used artificial neural networks, a machine-learning technique, to build an intrusion detection system able to detect a range of Distributed Denial of Service attacks. We are also exploiting off-the-shelf massively parallel architectures such as GPUs by exploring distributed computation approaches and refactoring the underlying data to significantly reduce its size and so improve on existing algorithm performance.
Cyber security is in part a technical challenge and in part a human challenge: cyber security depends on the interplay between users and security technology in societal and industrial contexts.
A key aspect of cyber security is encouraging users to behave safely online. Many online activities attract risks; some of these are known to the user and some are not. We have drawn on techniques from nudge theory and affective computing to encourage safe behaviour online. We have successfully nudged users through visual cues in a web browser into choosing longer and stronger passwords during a system enrolment task. We have developed a system that automatically detects risky online behaviour and provides feedback on risky behaviour in real time.
Extending human-centred security beyond cyber security, and in partnership with industry we have investigated the perceived influence of social presence at self-service checkouts by staff and its perceived effect on dishonest customer behavior. Our findings show that the perceived motivational and situational factors contributing to theft are complex, and surveillance in its current form does not appear to provide a sufficient social presence to prevent potential theft at self-service checkouts.
Digital devices play a huge role in our everyday lives and the nature of the data stored on these devices will paint a vivid picture of the life of its owner. Patterns of behaviour and social connections are deeply embedded in much of this data. For institutions such as law enforcement, this digital forensic data can serve as invaluable evidence. However, the sheer volume and complexity of this data means that analysing it in an effective way is challenging.
We are combining our expertise in digital forensics with the knowledge in the Division of Computing and Mathematics in augmented and virtual realities to discover new ways of exploring these highly complex digital forensic datasets. It is envisaged that a new way to display and interact with this information in a way that makes effective use of 3D spaces will lead to marked increases in investigative efficiency.
Find out more about the work our researchers do and how you could join us to study for a PhD, MPhil or Masters by Research