Heartbeat synchronization protocols attempt to have all active nodes in a distributed system to trigger a periodic event at the same time. This concept constitutes an important primitive of sensor networks, also known as Synchronicity, which can be defined as ``the ability for all nodes in the network to agree on a common period and phase for ring pulses'' For instance, in the specific case of sensor networks, synchronicity is particularly useful as a coordination primitive, in order to provide energy saving mechanisms (coordinate the wake-up time of sensor to communicate and diffuse data).

In the last years, several algorithms have been proposed to achieve heartbeat synchronization in distributed systems, considering different network settings (overlay networks, ad-hoc networks, wireless sensor networks, with or without message loss, churn and node crash). Most of those protocols are based on mathematical models inspired from the flash synchronization performed by certain species of fireflies. An ongoing Ph.D. project at U. Minho, by Paulo Jesus, is proposing robust averaging aggregation algorithms. We propose for this MSc research the study and development of a new envisioned approach based on averaging aggregation algorithms, in order to achieve a robust heartbeat synchronization (tolerant to non-byzantine failures and churn). The main purpose of this work will be the development of a new synchronicity algorithm.

Supervision: CBM, PSA, Paulo Jesus (PhD Student)