作者: Alan Burns , George M. A. Lima
DOI:
关键词:
摘要: Abstract A timely protocol to solve the distributed consensusproblem that tolerates process crashes and message omis-sions is described. The optimal in terms of thenumber communication steps needed achieve consen-sus. model on which based relies apriority-based network, a kind networkcommonly used practice support real-time systems.Keywords: Distributed Consensus, Fault Tolerance,Hard Real-Time System 1. Introduction Background consensus, fundamental problem dis-tributed systems, can be formulated as follows: all correctprocesses from set processes propose value thenthey have decide common value. con-sensus server provide basic building block imple-ment other fault-tolerant services [8, 3, 13, 9].More formally, consensus specifiedin following properties [12]:Termination: Every correct eventually decidessome value.Validity: If decides