An Improved Raft Consensus Algorithm Based on Asynchronous Batch Processing
Hao Li, Zihua Liu, Yaqin Li · Lecture notes in electrical engineering · 2022
Abstract The consensus algorithm has been popular in current distributed systems as it is more effective in solving server unreliability. It ensures a group of servers can form a coordinated system, and the entire system continues to work when a part of the service point fails. Raft is a well-known and widely used distributed consensus algorithm, but as it has a built-in purpose of comprehensibility, it is always compromised in terms of performance as a trade-off. In this paper, we mainly aim to improve the traditional Raft consensus algorithm’s performance problem, especially in high concurrency scenarios. We introduce a pre-proposal stage on top of the algorithm to achieve efficiency optimization through batch asynchronous log replicated and disk flushing. The experiment proved that the improved Raft could increase the system throughput by 2–3.6 times, and the processing efficiency for parallel requests can be increased by 20% or more.