故障容许计算的概念计算机程序的正确执行传统上由以下步骤保证: 1.在现行的成本约束内,选择最可靠的元件; 2.使用被证明了的连接和封装技术; 3.广泛地进行模拟以确定和排除逻辑设计中的错误; 4.对完成的硬件进行功能和诊断试验。经验证明,这些步骤不能预防计算机正确操作因逻辑故障而终止。程序出错的原因可以是由元件故障引起的永久故障,也可以是只引起错误、但并不留下故障元件的暂时故障。对于这里的讨论,我们定义逻辑故障是“计算机硬件中,一个或多个逻辑变量偏离其指定的设计值”。在数字硬件中,观测到了暂态的(暂时的)和永久的两种逻辑故障。 The concept of fault tolerant computing The correct implementation of computer programs has traditionally been ensured by the following steps: 1. Selection of the most reliable components within the current cost constraints; 2. Use of proven connection and packaging techniques; 3. Wide-ranging simulation To identify and eliminate errors in the logic design; 4. Perform functional and diagnostic tests on the finished hardware. Experience has shown that these steps do not prevent the computer’s proper operation from being terminated by a logical failure. The cause of a program error can be a permanent fault caused by a component fault or a temporary fault caused only by an error but without leaving a faulty component. For the discussion here, we define a logical fault as “one or more logical variables in computer hardware that deviate from their assigned design values.” In digital hardware, transient (temporary) and permanent logic failures were observed.