DR: Divided Ring Leader Election Algorithm

Abstract

Extending distributed system require new strategy for management and maintenance. For managing distributed system we need one leader like other process in system. Problem is when leader like other process crashed. For solving these problems like these researchers has proposed a number of local based strategies that is each region try to do its work by itself and the results will be gathered in a distributed manner. Leader election is the process of determining a process as the manager of some task distributed among several processes (computers). We suppose like network topologies our distributed system topologies are token and star. We proposed new schemes and discussed about some aspects of them. We proposed divided algorithm in ring topology and we try to reduce latency in the broad networks optionally. Our algorithm reduces the latency against messages but a bit because latency is more important than number of messages. At the end we see how using of n in Divided (n).