Business-Aligned Raft Algorithm in the Consortium with Multiple Permissioned Blockchains

Business-Aligned Raft Algorithm in the Consortium with Multiple Permissioned Blockchains

Raft is a well-known consensus algorithm with extensive application scenarios thanks to its comprehensibility and scalability in permissioned blockchain. But some issues about consensus efficiency still exist in Raft. For example, Raft detaches from business, and the random variable about leader identity has the largest information entropy in the leader election process. In this paper, we propose a Raft-like consensus algorithm, named Business-aligned Raft (BARaft) algorithm, which supports high transaction throughput and business adaptability. BARaft aims at optimizing the performance of the consensus algorithm in the consortium with multiple permissioned blockchains. BARaft combines actual application scenarios and specific characteristic of nodes to improve the adaptability and scalability of business. It enhances the countdown mechanism of leader election in Raft to reduce the randomness of leader election and build up business efficiency with fewer forwarded transactions. It also introduces new node states to reduce the scale of consensus cluster and improves the consensus efficiency. Experimental results show that BARaft provides a higher transaction throughput than Raft by 46.95% in a cluster of five consensus nodes.