A Stable Fault tolerance Algorithm for Leader Crash in Distributed Honeycomb Networks

Abstract

The failure of the coordinator (Leader) is one of the most serious issues in distributed systems. The coordinator's major responsibility is to keep the network's nodes in sync. Leader election algorithms can tackle this problem by converting the system from an unstable state in which all nodes are in the same state to a stable one in which one node is the coordinator and all other nodes in the network are aware of it.A new algorithm for replacing a failed honeycomb torus network coordinator with a fresh one is presented in this research. Honeycomb coordinates and network nodes placements are used in the proposed approach to direct election messages. Connection sequences are also used to define the message source and destination for one or more pathways. Concurrency and synchronization are taken into account by the new algorithm. The algorithm ensures that the problem will be solved in less than or equal O (n1.5) messages and O (n0.5) steps.