Early Scheduling in Parallel State Machine Replication.

SoCC '18: ACM Symposium on Cloud Computing Carlsbad CA USA October, 2018(2018)

引用 22|浏览49
暂无评分
摘要
State machine replication, a classic approach to fault tolerance, requires replicas to execute operations deterministically. Deterministic execution is typically ensured by having replicas execute operations serially in the same total order. Two classes of techniques have extended state machine replication to execute operations concurrently: late scheduling and early scheduling. With late scheduling, operations are scheduled for execution after they are ordered across replicas. With early scheduling, part of the scheduling decisions are made before requests are ordered; after requests are ordered, their scheduling must respect these restrictions. This paper generalizes early scheduling techniques. We propose an automated mechanism to schedule operations on worker threads at replicas, integrate our contributions to a popular state machine replication framework, and experimentally compare the resulting system to late scheduling.
更多
查看译文
关键词
Parallel State Machine Replication, Scheduling, Optimization Problems, Dependability
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要