Synthesis of dependable and self-driving communication networks

Communication networks have become a critical infrastructure of our digital society. The resulting stringent reliability requirements however stand in stark contrast to today’s manual and error-prone approach to operate networks. In this talk, I will present opportunities for using formal methods to build highly dependable communication networks. In particular, we show how the policy-compliance of important network protocols can be verified in polynomial time, and explore synthesis approaches to efficiently generate and update network configurations. We will also discuss opportunities of AI/ML methodologies in this context.