文章基本信息

标题：State Machine Replication Is More Expensive Than Consensus
本地全文：下载
作者：Karolos Antoniadis ; Rachid Guerraoui ; Dahlia Malkhi 等
期刊名称：LIPIcs : Leibniz International Proceedings in Informatics
电子版ISSN：1868-8969
出版年度：2018
卷号：121
页码：1-18
DOI：10.4230/LIPIcs.DISC.2018.7
出版社：Schloss Dagstuhl -- Leibniz-Zentrum fuer Informatik
摘要：Consensus and State Machine Replication (SMR) are generally considered to be equivalent problems. In certain system models, indeed, the two problems are computationally equivalent: any solution to the former problem leads to a solution to the latter, and vice versa. In this paper, we study the relation between consensus and SMR from a complexity perspective. We find that, surprisingly, completing an SMR command can be more expensive than solving a consensus instance. Specifically, given a synchronous system model where every instance of consensus always terminates in constant time, completing an SMR command does not necessarily terminate in constant time. This result naturally extends to partially synchronous models. Besides theoretical interest, our result also corresponds to practical phenomena we identify empirically. We experiment with two well-known SMR implementations (Multi-Paxos and Raft) and show that, indeed, SMR is more expensive than consensus in practice. One important implication of our result is that - even under synchrony conditions - no SMR algorithm can ensure bounded response times.
关键词：Consensus; State machine replication; Synchronous model