LEADER ELECTION In distributed systems, sometimes we've to pick just one leader from some nodes and that we called it a Co-Ordinator. Choosing a node is the same as granting special permission to the node. the strategy to determine on a frontrunner is sometimes called Leader Election. The goal of leader election is that after leader election, every node including the leader itself knows exactly who may be a frontrunner. Why do we want a leader? A leader is largely expected to try some tasks which should be executed just once. Let’s say there are 3 processes. Now we would like to send some WRITE requests to the present distributed system. After the request, these WRITE requests must be applied to all or any of the nodes. Of course, there must not be a knowledge inconsistency. Because there are multiple nodes, after one node accepts a call for participation, it must propagate it to other nodes. If we implement ...
Comparative Study Of Election Algorithms In Distributed Computing