Reliable Broadcast (RB) is a fundamental primitive for building fault-tolerant distributed systems, with applications in critical infrastructures, financial systems, and blockchain technologies. In this presentation, we explore how RB can be implemented in asynchronous environments prone to process- and network-related failures. Namely, we introduce a novel computing model combining Byzantine faults with a Message Adversary (MA). We then define the Message-Adversary-tolerant Byzantine Reliable Broadcast (MBRB) abstraction and prove its optimal resilience condition. We present three key algorithms implementing this abstraction: a simple signature-based MBRB algorithm, a new primitive called k2l-cast for cryptography-free MBRB implementations, and an erasure-coding-based MBRB algorithm optimizing communication complexity.
Timothé Albouy is a PhD student from the IRISA/Inria center of the University of Rennes, France. His work spans distributed algorithm design, performance optimization, and theoretical modeling of distributed systems.
This event will be conducted in English