论文部分内容阅读
本文讨论实现多处理机系统高可靠性的一些技术。这种多处理机系统由许多小型、专用的处理机构成。文中展示出来用多个处理机和连接器勿于达到高可靠性的系统结构,此外,提出了实现故障—安全设计的几种技术。主要的技术是:处理机间应急通信机构;处理机功能动态改变机构。文章还介绍了这些技术如何用于处理机故障的检测、重构和恢复处理,在这些处理中,系统使用一些轻负载处理机执行出故障处理机所分配的任务,而不是使用通常的备用硬件。
This article discusses some techniques for achieving high-reliability multiprocessor systems. This multiprocessor system consists of many small, dedicated processors. The article shows a system architecture that does not achieve high reliability with multiple processors and connectors. In addition, several techniques for implementing fail-safe design are proposed. The main technologies are: to deal with the emergency communication mechanism between machines; to change the mechanism dynamically by the processor function. The article also explains how these techniques are used to detect, reconstruct, and recover processor failures where the system uses some light-load processors to perform the tasks assigned by the failed processor instead of using the usual spare hardware .