Overview
This project is funded within the International Technology Alliance programme in Network and Information Science between UK (MoD) and US (ARL). One key challenge in coalition networks is the management and provisioning of secure networking. Because of their dynamic ad-hoc formation, the lack of fixed infrastructure, and the integration of different wireless and ad-hoc technologies, coalition hybrid networks pose new unique challenges and vulnerabilities. Traditional fault detection solutions to localize and correct network problems; common intrusion detection systems to recognize and prevent attacks, and current network forensics techniques to trace and identify the origin of an attack are inadequate for coalition hybrid networks. Solutions for coalition hybrid networks have to be distributed and able to adapt “on-the-fly” in order to respond to changes in the topology.
Expected Outcomes
This task aims to address the problem of secure networking by providing efficient and flexible solutions to locate and correct faults, to detect ongoing attacks, and to guarantee the overall security whilst preserving the flexibility of dynamic coalition networks. Specifically, we will develop
1) general methods for the analysis of distributed fault localization, and intrusion detection algorithms in hybrid networks. These will be built upon our declarative networking framework, developed in our previous ITA project.
2) new distributed algorithms for the management of information flow in hybrid and dynamic networks. In particular, new collaborative distributed algorithms for database query planning under policy confidentiality constraints.
The project is in collaboration with Professor Jorge Lobo (ICREA -UBF), Dr Franck Le (IBM Thomas J. Watson Research Centre and Professor Peng Liu (Pennsylvania State University). Official website of the ITA Programme is available here.
Detailed results on our recently developed distributed intrusion detection algorithm for MPTCP can be found in this technical report “Multipath Signature Detection“.
