Self-Sovereign Swarm: protecting decentralized IoT agents interactions
Geovane Fedrecheski, Laisa C C De Biase, Pablo C. Calcina-Ccori, Jan M. Rabaey, Marcelo Knörich Zuffo · 2025
Swarm Computing is a decentralized paradigm in which autonomous IoT agents dynamically discover and cooperate with one another to share resources and solve tasks collectively. However, enabling secure, spontaneous interactions in such open and dynamic environments remains a significant challenge, particularly for constrained devices. This paper introduces a lightweight, fully decentralized security framework for IoT agents, enabling autonomous secure communication and fine-grained access control without a centralized infrastructure. Our approach enables decentralized and secure communication, as well as access control for IoT agents through lightweight identity and trust mechanisms. The framework is implemented and evaluated on a real embedded platform, demonstrating low overhead in terms of runtime and storage. Security is analyzed under a realistic threat model, with mitigation strategies based on standardized cryptographic mechanisms. Our results confirm that secure and autonomous agent interactions in IoT Swarms are feasible and efficient, even on constrained devices.