Taxonomy of Distributed Consensus Algorithms

Abstract – In distributed database computing transactions, group agreements of all the participating nodes is the crucial task. In such transactions, distributed algorithms are used for solving the problem of consensus. Distributed consensus algorithms are mainly of two types one is used in the case of link failures and second for process failures. For process failures, the algorithms used for solving the consensus problem are stopping, byzantine failures and commit protocols. The 2-phase commits (2-PC) and three-phase commit protocols (3-PC) are distributed algorithms that make all nodes in a distributed system agree to either commit or abort a transaction. The algorithms like FloodSet, EIGStop, EIGByz, PAXOS and E3PC are also used in last few years for solving this problem. This paper presents taxonomy of Distributed Consensus Algorithms. Different types of distributed algorithms for solving consensus problem in distributed database computing transactions are discussed here. Keywords- Distributed Consensus Algorithms; 2-phase commit protocols; 3-phase commit protocols; stopping failures; byzantine failures. 1