Phd Position F - M Group Key Management In Decentralized Collaborative Systems H/F

INRIA

PhD Position F/M Group key management in decentralized collaborative systems
Le descriptif de l'offre ci-dessous est en Anglais
Type de contrat : CDD

Niveau de diplôme exigé : Bac +5 ou équivalent

Fonction : Doctorant

Contexte et atouts du poste

This PhD position is in the context of PILOT project of PEPR eNSEMBLE ().

The PhD student will be located in Nancy at The Inria Center of the University of Lorraine in the Loreley team. It will be supervised by Claudia-Lavinia Ignat, Research Director at Inria in Nancy (Loreley team), Mathieu Turuani, Researcher at Inria in Nancy (PESTO team) and Davide Frey, Researcher at Inria in Rennes (WIDE team).

The Inria Center of the University of Lorraine is one of Inria's nine centers and has twenty project teams, located in Nancy, Strasbourg and Saarbrücken. Its activities occupy over 400 people, scientists and research and innovation support staff, including 45 different nationalities. The Inria Center is a major and recognized player in the field of digital sciences. It is at the heart of a rich R&D and innovation ecosystem: highly innovative PMEs, large industrial groups, competitiveness clusters, research and higher education players, laboratories of excellence, technological research institutes, etc.

Mission confiée

Context

In large-scale collaborative environments, groups of users frequently join and leave shared workspaces while interacting with shared documents, communication channels, or collaborative tools. In such dynamic groups, ensuring secure communication among participants is a critical challenge. In particular, the management of cryptographic group keys becomes complex when membership changes occur frequently.

When a participant leaves or is removed from a collaborative group, the confidentiality of the collaboration requires that this user should no longer have access to future communications or shared documents. Consequently, a new group key must be generated and distributed among the remaining members. In traditional approaches, this process requires several rounds of communication between participants to update and redistribute cryptographic material. In large and dynamic groups, these operations can introduce significant performance overhead, increase communication costs, and lead to temporary interruptions in the workflow of participants.

For many years, solutions for secure n-party group communication have faced a trade-off between scalability and security. Some protocols offered strong security guarantees but did not scale well to large groups, while others achieved scalability at the cost of weaker security guarantees.

The recent standardization of the Messaging Layer Security (MLS) protocol [1] addresses many of these challenges. MLS is designed to enable secure, scalable, and efficient group communication, supporting dynamic membership changes while ensuring strong security properties such as forward secrecy and post-compromise security. MLS relies on advanced group key agreement mechanisms that allow the cost of membership changes to scale logarithmically with the group size [2].

However, practical deployments of MLS rely on an additional component called the Delivery Service (DS). The Delivery Service acts as an intermediary that coordinates message exchange between group members and ensures that group operations are delivered in a consistent order. While the MLS protocol assumes that this service is untrusted with respect to message confidentiality, it nevertheless plays a critical role in maintaining group consistency and availability.

This centralized Delivery Service introduces potential availability, reliability, and security risks. A compromised or malicious Delivery Service could disrupt message ordering, block progress of the group, or create inconsistent views among participants. Furthermore, centralization raises concerns regarding scalability and resilience in large-scale collaborative environments.

Objectives

This PhD project aims to investigate the security and resilience of the Delivery Service component in the MLS architecture, with the objective of designing a decentralized and Byzantine-resilient Delivery Service suitable for large-scale collaborative systems.

First, we aim to analyze the security limitations of the current centralized Delivery Service model used in MLS deployments. Using both existing implementations and formal verification techniques, we will demonstrate that the centralized architecture can be vulnerable to several classes of attacks, including message ordering manipulation, consistency violations, and denial-of-service scenarios.

Second, we will design a decentralized Delivery Service architecture capable of coordinating message exchange among MLS clients without relying on a single trusted infrastructure. This architecture will aim to maintain the key consistency properties required by MLS while tolerating failures or malicious behavior of some participants.

In particular, we will focus on achieving resilience against Byzantine faults, where nodes in the system may behave arbitrarily or maliciously. Our goal is to develop a Delivery Service design that guarantees:

- consistent ordering of MLS messages,
- agreement among group members on the sequence of operations,
- robustness against malicious participants,
- scalability to large and dynamic groups.

Third, we will validate the proposed architecture through formal verification and experimental evaluation. Formal methods will be used to verify the correctness and security properties of the proposed protocol, while practical experiments will evaluate its performance and scalability in realistic collaborative scenarios.

Bibliography

[1] R. Barnes, B. Beurdouche, R. Robert, J. Millican, E. Omara, and K. Cohn-Gordon, The Messaging Layer Security (MLS) Protocol, RFC 9420, Jul. 2023.

[2] K. Bhargavan, R. Barnes, and E. Rescorla, Treekem: asynchronous decentralized key management for large dynamic groups A protocol proposal for Messaging Layer Security (MLS), Inria Paris, Tech. Rep., 2018.

[3] B. Blanchet, Modeling and verifying security protocols with the applied pi calculus and Proverif, Foundations and Trends in Privacy and Security, vol. 1, no. 1-2, pp. 1-135, 2016.

[4] Timothé Albouy, Davide Frey, Mathieu Gestin, Michel Raynal, François Taïani: Contention-Aware Cooperation. OPODIS 2025: 9:1-9:19

[5] Ludovic Paillat, Claudia-Lavinia Ignat, Davide Frey, Mathieu Turuani, and Amine Ismail. 2024. Discreet: distributed delivery service with context-aware cooperation. Annals of Telecommunications 80 (2024), 357-374.

[6] Matthieu Nicolas, Victorien Elvinger, Gérald Oster, Claudia-Lavinia Ignat, François Charoy: MUTE: A Peer-to-Peer Web-based Real-time Collaborative Editor. ECSCW Panels, Demos and Posters 2017

Principales activités

The research will be conducted in several stages.

Security Analysis of Existing MLS Delivery Services

Using symbolic verification tools such as ProVerif [3], we will analyze how the Delivery Service interacts with MLS clients and identify potential vulnerabilities related to message ordering, proposal handling, and commit validation.

We will also study several existing MLS implementations to understand how Delivery Service assumptions are implemented in practice and how inconsistencies may arise in real deployments.

Design of a Decentralized Delivery Service

Based on the vulnerabilities identified in the first phase, we will design a distributed Delivery Service architecture that removes the reliance on a centralized server.

Possible approaches include:

- peer-to-peer message dissemination,
- distributed ordering protocols,
- Byzantine fault-tolerant consensus mechanisms such as CAC [4],

The design will aim to preserve the security guarantees of MLS while improving resilience and availability.

Formal Verification of the Proposed Architecture

Once the decentralized architecture is designed, we will formally analyze its security properties using formal verification tools such as ProVerif.

The analysis will focus on verifying properties such as:

- epoch agreement [5],
- epoch-content consistency [5],
- resilience to Byzantine nodes.

Implementation and Experimental Evaluation

To demonstrate the practical applicability of the proposed solution, we will implement the decentralized Delivery Service and integrate it into real collaborative systems.

In particular, we plan to integrate our solution into the MUTE [6] peer-to-peer collaborative editor or collaborative tools from La Suite Numérique, such as Docs.

This integration will allow us to evaluate the system in realistic collaborative scenarios involving dynamic group membership and large number of participants.

Case Studies

An important application domain for secure group communication is healthcare, which is already being investigated by several PhD students and postdoctoral researchers within the PILOT project. In this domain, multidisciplinary teams collaborate around patients while handling highly sensitive medical data. Physicians, nurses, specialists, and external consultants frequently join or leave care teams, requiring secure mechanisms to update access rights to shared medical information. Group key management mechanisms such as those provided by the Messaging Layer Security (MLS) protocol can ensure that only authorized participants can access patient-related communications, while preventing former members from accessing future discussions. A decentralized MLS infrastructure would further improve resilience and trust when collaboration spans multiple hospitals or healthcare institutions, enabling secure inter-organizational medical collaboration. We will adapt our solution to some healthcare use cases already identified by other PhD students or postdocs of PILOT.

Compétences

- Engineering and/or Master 2 degree in Computer science / Applied mathematics with an experience in computer networks.
- Theoretical expertise: distributed systems, security
- Good collaborative and networking skills, excellent written and oral communication in English
- Good programming skills
- Strong analytical skills

Avantages

- Subsidized meals
- Partial reimbursement of public transport costs
- Leave: 7 weeks of annual leave + 10 extra days off due to RTT (statutory reduction in working hours) + possibility of exceptional leave (sick children, moving home, etc.)
- Possibility of teleworking (after 6 months of employment) and flexible organization of working hours
- Professional equipment available (videoconferencing, loan of computer equipment, etc.)
- Social, cultural and sports events and activities
- Access to vocational training
- Social secu€rity coverage

Rémunération

€2300 gross/month