The Second Virtual Workshop Day of the French-German Center for Cybersecurity will take place on Thursday, 20th May.
02:00 p.m. to 02:05 p.m. Introduction by Prof. Marine Minier and Dr Antoine Joux (Directors of the French-German Center for Cybersecurity)
02:05 p.m. to 02:30 p.m.: Dr Jérôme François (Inria Research Scientist with RESIST Team): In-network security monitoring
Abstract: Monitoring is a critical task of network operations to serve as support as many other ones such as fault management, security, trouble-shooting, QoS… Recently, network softwarization empowers the paradigm of programmable networks where network functions or even tasks of these functions can be programmed with high flexibility to fit user or operator needs independently of hardware used. Programming network data-plane allows to program the processing of each packet and is supported by novel technologies like P4. In this talk, we will bring the gap between data-plane programming with P4 and network monitoring by highlighting what possibilities exist to support an enhance in-network security monitoring rather than offloading the latter to a server as usual. In particular, we will discuss how complex (malicious) behaviors can be tracked with extended finite state machines and how network switch could support real number operations that is a requirement for expecting AI to be run in-network.
02:30 p.m. to 02:40 p.m.: Q&A
02:40 p.m. to 03:05 p.m.: Dr Antoine Joux (Faculty at CISPA): Subset-sum collisions
Abstract: In this talk, we revisit the subset-sum collision problem in a parameter regime where lattice reduction technique are expected to be costly. For this, we turn to combinatorial algorithms based on the representation technique and show how they can be applied to the subset-sum collision problem. After a survey of the state of the art, we present two algorithmic strategy: first, a reduction to the resolution of ordinary subset sum problem and, second, a direct application of the representation technique. Quite surprisingly, for a given number of elements in the subset-sum, the subset-sum collision problem is (with current algorithms) harder that the subset-sum inversion problem. This contrasts with the general intuition that collisions are easier to find than pre-images.
03:05 p.m. to 03:15 p.m.: Q&A
03:15 p.m. to 03:30 p.m.: Final discussion
03:30 p.m. to 04:00 p.m. Virtual Coffee on gather.town