实时协同编辑系统的容错可靠性研究

容错系统是提高实时协同编辑系统可靠性的有效手段。为了满足协同的实时要求，向前式恢复用全局快照解决系统的客错问题。失效的进程是基于逻辑向量时钟发送消息，通过检查点操作和日志记录进行状态的恢复，避免了多米诺现象的发生。整个系统采用复制式结构，系统的可靠性分摊到各个处理机上，而不是仅依赖服务器和备份服务器的可靠性。

FAULT-TOLERANT REAL-TIME COOPERATIVE EDITING SYSTEMS

If the Computer Supported Cooperative Work(CSCW) is to be effectively used in WANs,where the networking and computing resources are unpredictable,it should allow process or link failures.In particular,processes should be able to leave and rejoin groups and recover from failures.Disconnected users should be able to rejoin groups without having to restart from scratch and reduce usage overhead as much as possible.Fault-Tolerant systems,which significantly improve the reliability of the real-time cooperative editing system,are highly desirable.Forward recover is used to cope with process and link failures.To achieve fault-tolerance,the processes record global snapshots.Upon a failure,the system recovers from the recent snapshot.Our scheme's effectiveness is shown by mathematical analysis.