Towards a Model-Based Federated Learning Architecture for Intrusion Detection and Mitigation in Small Office and Home Office Networks

This project aims to strengthen RabbitRun Technologies’ (RRT) cybersecurity capabilities in software-defined wide area network (SD-WAN) solutions for small office/home office (SOHO) environments. It addresses the critical need for secure and reliable connectivity in remote and small business settings by integrating federated learning into RRT’s systems. This approach enables RRT routers to collaboratively improve intrusion detection and mitigation while maintaining data privacy and minimizing communication demands.

The key goals of the project are: (1) developing a domain-specific language (DSL) to simplify the mining and specification of security-related network behaviours; (2) advancing state-machine learning techniques to better detect evolving cybersecurity threats; and (3) implementing a federated learning infrastructure for collaborative model refinement across distributed routers.

This initiative bridges academic research and industry innovation, providing the intern with hands-on experience in applied AI, cybersecurity, and model-based software engineering. It also lays the foundation for secure, dynamic networking solutions, benefiting both RRT and the broader Canadian digital economy.

Faculty Supervisor:

Mehrdad Sabetzadeh;Shiva Nejati

Student:

Partner:

RabbitRun Technologies

Discipline:

Computer science

Sector:

Information and cultural industries

University:

University of Ottawa

Program: