We have proposed new models and algorithms for effective QoS management in multi-high wireless networks and mainly in Mesh, ad hoc, and vehicular networks (collaboration with the Université de Sherbooke and the Université de Montréal). We were particularly interested in congestion control issues at the Mesh backbone level and admission control and traffic regulation at access nodes. The models developed, using among others the fuzzy logic learning technique, allowed two interesting results: (1) efficient use of bandwidth and (2) equity between network flows.



The first "security" topic we discussed was the detection of DDoS (Distributed Denial of Service) distributed denial of service attacks, as security solutions are still incomplete due to technical difficulties and cooperation between Internet operators. After a detailed study of this problem, we proposed a solution against these attacks using a fully distributed detection system at the level of edge routers. This solution was validated by a concrete implementation integrating detection probes into our intelligent agents and using a peer-to-peer protocol for routing information between agents. We currently have a DDoS attack detection platform that we used in our research project "Collaborative and Distributed Intrusion Detection System" as part of the European 2Centre project with Thales. The main actor in this approach left the UTT in 2014 for an EC position at Télécom ParisTech.