摘要： 本文根据分布式系统的运行原理提出了分布式系统中以通讯点作为断点保留的思想.并就此为出发点,讨论了系统诊断、断点保留、故障检测和系统恢复的协调设计思路.

关键词: 分布式系统, 断点, 通讯关系表, 一致性, 系统恢复

Abstract: According to the priciples of distributed systems, the paper presents the idea of checkpoint saving when communications occur. Continuing this thought it gives the design methods among system diagnosis, checkpoint saving, fault detection, and system recovery.