The computer security research at the Department of Computer Science and Engineering is founded on a firm conviction that security must be applied to all parts and at all levels of the system, as opposed to traditional security methodologies focusing on boundary protection. Security considerations should also follow the system throughout its lifetime, starting from the specification phase, throughout construction, operation and maintenance. We call this approach in-depth security and our research spans all those phases.

For example, we have begun looking into the problem of tackling security problems at the earliest stages in the design of systems, before code is produced. In the next stage we consider programming-language based security, where we investigate the application programming language technology to ensure that only secure applications can be built or run. In collaboration with highprofile academic and industrial partners, we develop this technology for securing networks of extensible and heterogeneous devices (such as mobile phones) within Mobius, an EU integrated project on Mobility, Ubiquity and Security.

Despite our best intentions, many security mechanisms fail or are circumvented by malicious attackers, and detecting how and when this happens helps us understand and improve subsequent versions of the systems. Thus, another main branch of our current research is devoted to the next generation of intrusion detection systems (IDS), which can assist us in recovery after an intrusion. An important specialised area of IDS is fraud detection, focusing on economic crime. Here, our latest research results have led to immediate improvements in commercial systems.

Many of the challenges of computer security are best countered together with industrial partners. Therefore, as a complement to our basic research strengths, we have recently established an Open Arena in Security at Lindholmen Science Park, a common forum for applied research and development in close cooperation with such partners.