The Malicious and Accidental Fault Tolerance for Internat Applications (MAFTIA) project was shortlisted in the field of information sciences and aims to make large network infrastructures, such as the Internet, more secure.
The project was co-ordinated by Dr Robert Stroud of Newcastle University, in co-operation with research teams in in co-operation with research teams in France, Germany, Portugal, Switzerland and the UK.
The MAFTIA project is designed to provide a new way to think about and build networks that provide a bridge between "dependability and security", according to the project's Web site .
"We can't build perfect systems, and we have to be able to tolerate attacks -- even some partially successful attacks. You have to avoid a single point of trust and therefore a single point of failure," said Stroud, who is a reader in computer science at Newcastle.
"The Internet itself was highly distributed, to avoid being destroyed in a nuclear attack, but some of the protocols in the Internet do create single points of failure. You want to prevent a single failure from bringing down the system," said Stroud.
MAFTIA maps out a set of mechanisms and protocols for achieving intrusion tolerance. These include middleware protocols for secure group communication; an architecture for a large-scale distributed intrusion detection system; a blueprint for building generic trusted third-party services; and the design and implementation of an intrusion-tolerant distributed authorisation service.
The MAFTIA researchers also tried to build a formal verification and assessment for their own middleware, using "a rigorous model for reactive cryptographic systems that allows for formal specification and verification of security properties under a standard cryptographic semantics".
According to Stroud, IBM's Zurich Research Labs, which was one of the project's original partners, has developed a set of protocols to replicate services over the Internet -- a difficult problem, as the Net is inherently unstable.
"IBM found a way to do it in a non-deterministic way, and built services such as a replicated certification authority, so trusted third-party services can be replicated in a way that means if they are under attack, you can have different implementations," said Stroud. "You are using diversity to make yourself more resilient to attack."
It's the fifth year that the EU Descartes Prize has been awarded. The teams short-listed this year come from 20 countries, and include people working in life sciences, engineering, chemistry and physics.
Sixty-five teams from 19 European and non-European countries have been awarded the prize for projects that range from basic sciences, chemistry and life sciences to electronics and physics.
The two winners of this year's award -- which is worth €1m (£0.66m) with €500,000 for each winner -- will be announced on 2 December in Prague.