AI-driven Malware-Tolerant Network for IoT (AIM)


Main funder

Funder's project number5832/31/2018


Funds granted by main funder (€)

  • 58 400,00


Funding program


Project timetable

Project start date01/11/2018

Project end date31/05/2019


Summary

We are developing network defense system which relies on software-defined networking (SDN)
and network function virtualization (NFV) solutions and allows customers to mitigate attacks
performed against their smart devices. The main idea is to employ recent advances in
reinforcement machine learning in order to enhance network security mechanisms by letting an
artificial intelligent agent evaluate risks related to an intrusion and make the most optimal real-time
crisis-action decision on the network security policy.On infrastructure level, our framework includes
cloud compute servers that allow the defense system to emulate elements of real infrastructure
(sensors, routers, gateways) and launch security network functions (firewalls, intrusion detection
systems, honeypots) on fly. Speaking of SDN, its main advantages are global visibility of the
network state, centralized control of network behavior, and run-time manipulation of traffic
forwarding rules. The defense system can request flow samples through the controller from the
data-path and, after security analysis, redirect the data-path elements to either block the traffic,
reroute to security middle boxes or restrict the traffic within a particular network jurisdiction.


Principal Investigator


Primary responsible unit


Last updated on 2022-06-07 at 12:40