Abstract: Many distributed algorithms require one process or node to act as leader node in distributed systems. Election algorithms are meant for electing such process or node that acts as leader node also called as coordinator from among currently alive processes in such a way that at any instance of time there will be single coordinator for all the processes in the system. Therefore, election algorithms are very important in any distributed system. Bully algorithm is one of the standard approaches for electing the coordinator in distributed systems. In this paper, we have compared base and efficient version of bully algorithm to minimize the number of messages when electing the coordinator and when a process recovers from a crashed state in distributed systems.