Presentation

The Network and Security (NS) theme focuses, on the one hand, on issues related to data transmission and routing protocols in communication networks, primarily wireless and IoT networks, and, on the other hand, on the security of these protocols and services. The research activities conducted within this theme can be summarized as follows:

• Design of communication protocols and distributed algorithms for communication networks: This line of research addresses both short-range and long-range communication technologies. The investigated problems include resource allocation and interference management in LoRaWAN, Wi-Fi, cellular, and satellite networks, as well as interoperability across heterogeneous networking technologies, with the objective of building more robust and flexible communication infrastructures. As networks become increasingly complex and dense, dynamic performance optimization is becoming critical. In this context, part of our work relies on Artificial Intelligence (AI) models to dynamically optimize network performance by improving resource management and reducing interference. Additional research addresses the coordination of robotic swarms and the study of the computational power associated with different mathematical models of robots with limited capabilities (e.g., Look-Compute-Move model, luminous robots). These contributions target applications ranging from environmental monitoring to industrial use cases.
  
  
• Securing and improving the resilience of networks and communication protocols: Research efforts in this area focus on the design and validation of secure protocols or protocols guaranteeing resilience properties. First, the use of formal methods enables us to design secure protocols and to automatically prove their security properties. Several tools, such as ProVerif and Tamarin, are employed within the Dolev–Yao attacker model. A wide range of use cases is addressed, including electronic voting, messaging protocols, ticketing systems, blockchain technologies, and banking payment systems. Another key research area concerns the evaluation of the effectiveness of machine learning (ML)-based intrusion detection systems (IDS) in real network environments. Particular emphasis is placed on assessing the gap between IDS performance on training datasets and their behavior in real-world deployments. Finally, we also investigate the ability of systems to withstand unexpected or undesirable events, through the study of properties such as self-stabilization and tolerance to benign or malicious faults.

• Design of cryptographic primitives: Finally, part of our research is dedicated to the design of new cryptographic primitives that provide stronger privacy guarantees for users. Anonymous signature schemes and anonymous certificates are studied in order to enhance communication security. Other work has led to the development of a distributed data storage system spanning multiple cloud providers, in which users do not need to remember cryptographic keys, while ensuring data confidentiality with respect to the involved parties.

In addition to these three core research directions, the NS theme is also involved in several related projects and collaborations with other research teams and institutions. In collaboration with Institut Pascal, we work on the implementation of deep neural networks on FPGA architectures, with strong optimization constraints in terms of both performance and hardware resource utilization. A second research axis focuses on the security of complex Large Language Model (LLM) architectures—incorporating RAG pipelines, agents, and fine-tuning—within an inter-theme collaboration with the DSI theme of the SIC axis. This collaboration is structured around the endowed chair Reliable and Trustworthy Use of LLMs. Finally, a third research axis investigates zero-knowledge proofs based on physical objects (e.g., card games, Roberval balances). This work, conducted in collaboration with the Tokyo University of Electro-Communications in Japan, complements cryptographic primitives in environments where computing is not assumed, such as ballot-based elections with physical ballot boxes.

We organize a monthly seminar: https://sancy.iut.uca.fr/~durand/seminaire/