ALLONE: A New Adaptive Failure Detector Model for Low-power Lossy Networks

Fatima Zohra Benhamida (1) , Yacine Challal (2, 3) , Mouloud Koudil (1)

1 LMCS - Laboratoire de Méthodes de Conception de Systèmes
2 Heudiasyc - Heuristique et Diagnostic des Systèmes Complexes [Compiègne]
3 Labex MS2T - Laboratoire d'Excellence "Maîtrise des Systèmes de Systèmes Technologiques"

We consider the problem of failure detection in networks with energy and communication constraints. Most of current implementations of unreliable failure detectors (FD) use mechanisms to notify process failures in fully connected networks with reliable communication links.This assumption is not applicable to lossy networks. Furthermore, such implementations do not consider resource limitations in terms of energy, storage and bandwidth. This paper presents a new failure detection model for Low-power Lossy Networks(LLN). Our approach defines an adaptive timer-based paradigm. Besides, we introduce two techniques based on the proposed general model. We evaluate all contributions using implementation on Omnet++/Mixim framework. Simulation results demonstrate that our FD enhances detection performances for LLN compared to traditional FD.

Domaines: Réseaux et télécommunications [cs.NI]

Fatima Zohra Benhamida, Yacine Challal, Mouloud Koudil. ALLONE: A New Adaptive Failure Detector Model for Low-power Lossy Networks. IEEE Global Communications Conference, 2013, Atlanta, United States. pp.113-118. ⟨hal-00871191⟩