Feb. 20, 2023, 14h-16h30
Amphi Gaston Planté, 2 rue conté, 75003 Paris
Nous organisons un cycle de séminaires avec trois séminaires de Sofiene Jelassi, Patient Ntumba, José Lopez Esteves.
QoE-aware AI-based resilient softwarized networks
Sofiene Jelassi, IRISA, Rennes.
Network softwarization has been proposed as an alternative way for computer networking. This environment has opened up a new era in computer networks where innovative solutions are sought. Undoubtedly, network softwarization is a primary driver behind the thriving 5G and Edge/Core systems. In a software network environment, there is a sharp distinction between pictured and physical networks. A pictured network is an instantiation of a real network that exists for a certain period of time over a physical network. A pictured network, which is ad—hoc and volatile, utilizes customized functions and policies. In that context, innovative approaches and solutions should be proposed for the sake of efficient management, control and monitoring.
In this talk, I will discuss principal measurement and monitoring strategies proposed for softwarized network environments. Next, I will introduce our QoE-based monitoring solution that has been designed specifically for this kind of environment. The proposed architecture aims at providing a straightforward, automatic and evolving QoE-based control-loop of instantiated networks. I will shed light on our efforts in order to measure and estimate specific parameters required by a given QoE function. The estimation is performed by means of AI-based inference techniques that use a reduced, primitive easily-accessible dataset in order to derive valuable and insightful metrics. In the last section of the presentation, I will describe recent results about resiliency over 5G RAN. I will discuss our ongoing work about the use of simulation after experimentation in order to provide strategies for building resilient softwarized networks.
Sofiene Jelassi is an Assistant Professor at the University of Rennes. Currently. He is a member of the Ermine research team located at IRISA/Inria in Rennes. He received respectively his MSc in 2005 and his PhD in 2010 from the University of Monastir and the Sorbonne University. His research interest includes network measurement and monitoring, automatic QoE assessment, network and service automation, and 5G RAN resiliency. He is the author of more than 20 scientific papers in well-reputed journals and conferences.
Adaptive Scheduling of Continuous Operators for IoT Edge analytics
Patient Ntumba wa ntumba, ROC team, Cnam/Paris
In this talk, we address the problem of adaptive scheduling data stream processing and analytics (DSPA) applications in the Edge-Fog-Cloud continuum. We are paying particular attention to the dynamic workload of DSPA applications due to variability of IoT data stream rates and of the Edge-Fog-Cloud resource capacities that may be shared across multiple applications (DSPA or other).
In this respect, we propose a new model for assessing the usage cost of both computational and network resources in a Edge-Fog-Cloud architecture that takes into account the inherent trade-offs from their joint usage. We additionally consider a response time model that captures constraints over windowing processing of IoT data streams. Based on these models, we formulate the problem of adaptive scheduling DSPA applications over a hierarchical Edge-Fog-Cloud architecture as a resource sharing optimization problem with response time constraints. To solve the problem we introduce a heuristic algorithm, called aTSOO-H, that adaptively schedules continuous operators of a DSPA application by taking into account its current deployment state.
Via extensive simulation experiments, we show that aTSOO-H exhibit optimal trade-offs between resource usage cost, response time and the execution cost.
Patient Ntumba is a postdoctoral researcher at le Cnam of Paris, in the ROC team. His research mainly focus on data streams, Internet of things, middleware, distributed systems, dynamic systems and Optimization. Before joining the Cnam, Patient worked as postdoctoral researcher at the Inria research center of Paris in the MiMove team after receiving his Ph.D in 2022 from Sorbonne University, conducting his Ph.D research at Inria Paris. He received his MSc in 2016 from Université de Franche-Comté in Besançon and in 2013 his BSc from Université de Kinshasa. Before starting his Ph.D, he has been working as R&D engineer at Inria Paris in the framework of the European project Chorevolution.
Intentional Electromagnetic Interference and risks for information security
Jose LOPES ESTEVES
Intentional electromagnetic interference (IEMI) relates to the exploitation of the electromagnetic susceptibility of a target electronic device in order to disrupt, confuse or damage it.
This kind of electromagnetic perturbation is historically investigated through the electromagnetic compatibility lens, with a focus on the estimation of parasitic currents and voltages or on the impacts on the availability of the main function.
This approach is therefore not well adapted for information security analysis, and new approaches are needed for the detection and the characterization of the impacts of IEMI on the information at the logical layer, necessary for exploitability and risk analysis.
In this talk, the possibility of using fault models is investigated as a way to both enable the detection of effects impacting information and analyze the exploitability of the effects from an information security perspective.
To illustrate the approach, the assessment of the susceptibility of smartphones will be explained, and a remote exploitation of voice assistants with IEMI will be discussed.
José is an information security researcher and the research coordinator in the wireless security lab at ANSSI.
His topics of interest include the security of wireless communication protocols, electronic warfare and electromagnetic security, which relates to the threats arising from interactions between electronic devices and their electromagnetic environment (e.g. TEMPEST, IEMI).
He is currently finalizing a PhD with the CNAM, under the supervision of Pierre Paradinas.