Home

IMG_2338_Gerardo Rubino

Senior Researcher at Inria

Contact:
Tel.: +33 (0) 2 99 84 72 96
Email: gerardo.rubino@inria.fr

Research institutions: Dionysos team-projet INRIA/IRISA

Abstract

I work for Inria, the main French public research institution in computer science and applied mathematics. I am the scientific leader of the Dionysos team at Inria’s center in Rennes, in the French Brittany area. Dionysos’ research themes are the design and the analysis of communication networks and systems. I was previously the head of the Armor team, with similar research interests.

At Dionysos, we work mainly with wireless infrastructures and with optical networks. From the analysis viewpoint, our focus is on the quantitative aspects of these systems: we develop techniques allowing their analysis from the performance or dependability viewpoints, extended to Quality of Experience, to vulnerability aspects, to performability. We also work on the analysis of other types of systems with similar complexity levels.

My own expertise areas are the analysis of the Quality of Service and of the Quality of Experience of applications and services built on top of a network, and especially of the Perceptual Quality when a media such as voice or video is involved. I am the author of a technology called PSQA allowing the automatic measure of this essentially subjective metric. On the quantitative aspects, my domain is the analysis of probabilistic models, in particular of Markovian ones. This includes of queues and networks of queues, in particular in the transient regime, and dependability models. I also work on reliability diagrams, from the analytical or from the numerical points of view, as well as in problems appearing when the analysis is done by simulation (discrete event simulation, Monte Carlo techniques). On the latter topic, I mainly work on the development of methods for dealing with rare events. From the networking viewpoint, I’ve been working on wireless technologies (access control, for instance, or ressources management) and of wired ones, mainly on the design of optical infrastructures. Last, I use Machine Learning tools (mainly in Supervised Learning and in Reinforcement Learning) and I also develop generic tools in the area (for the training phase in Supervised Learning, and for time series prediction in the Reservoir Computing field).