Servers Investigate

Message Efficient Leader Election in Synchronous Distributed System With Failure Detectors

Download now Free registration required

Executive Summary

Leader election, simply election, is an important problem to construct fault-tolerant distributed systems. Depending on a network topology, many kinds of leader election algorithms to elect a high-priority leader have been presented so far. Some algorithms are based on complete topology and others based on ring topology or tree topology. Among those, as a classic paper, there is the Bully algorithm for synchronous systems based on complete topology specified by Garcia-Molina. The leader election algorithm is used usefully in those systems where a coordinator of the protocol is needed, such as replicated data management, atomic commitment, process monitoring and recovery.

  • Format: PDF
  • Size: 67.1 KB