计算机系统容错技术研究

Research on Fault-tolerant Technology for Computer System

针对计算机系统中软、硬件可靠性问题的不同特点,讨论容错技术的最新发展现状,分析计算机系统中的各种容错方法,包括传统的冗余设计、错误回卷恢复机制以及当前研究较多的一般化容错设计方法等,研究目前已有的一些容错方法在反应延迟、容错成本、精确量化、异构同步、可靠性建模等方面存在的缺陷以及待解决关键问题,并对如何进一步更好地完善和使用这些容错方法进行总结。

According to different acharacteristics of software reliability and hardware reliability,this paper discusses the current developping status about fault-tolerance technology,and analyzes various ideas and strategies of fault-tolerance methods in computer systems including redundance,error rollback recovery mechanism and generalized fault-tolerant design method etc.Some defects and key issues for some existed fault-tolerance methods on reaction delay,fault-tolerance cost,precise quantification,heterogeneous synchronous,reliability modeling are pointed out.How to improve and use these methods better is suggested.