本文试图建立计算机可靠性领域的统一观点,其中有关于可靠性设计、冗余技术和诊断方面的考虑。为了叙述这些,本文分成三个部分。在第一部分,讨论一种重要的设计,这种设计对于判定元件是固有的。这就是建立合适的设计准则,以便把故障的影响减到最小。第二部分,我们研究失效检测和纠正,进而掩蔽它对系统的影响。在最后一部分论述故障系统的故障诊断和维护系统对失效组合件的隔离。 This article attempts to establish a unified view of the field of computer reliability with considerations for reliability design, redundancy techniques, and diagnostics. In order to describe these, this article is divided into three parts. In the first part, we discuss an important design that is inherent to the decision element. This is to establish appropriate design guidelines in order to minimize the effects of failures. The second part, we study the failure detection and correction, and then mask its impact on the system. In the last section, we discuss fault isolation and fault isolation in faulty system.