AN ADAPTIVE TECHNIQUE FOR ELECTING THE MASTER NODE IN DISTRIBUTED SYSTEM VIA THE CLUSTERING PROCESS

The process of electing the master node in distributed systems is a common problem that requires a huge amount of time. To participate in solving such problem, this research paper presents an improvement version for bully algorithm which bases on clustering concept. The proposed framework divided the distributed system into main cluster and sub cluster, in which the main consists of master and assistant nodes, while sub cluster holds Demaster, Assistant nodes and leaf nodes. The constructed framework reduced the communications overload and single point of failure.