Fault Tolerance Schemes for Multiprocessor Systems: Algorithmic Issues

多处理器系统比传统的单处理器机器更容易发生故障。因此，将设计增强算法方面容错能力的方案，以解决在设计这些方案时出现的几个问题。这项研究的重点是最近提出的一些在线技术，这些技术可以在正常计算的同时完成故障检测。这项研究的主要目的是开发实用和有效的故障检测和定位方案，并更好地了解多处理器系统容错方案设计中涉及的问题。

Multiprocessor systems are more susceptible to failure than conventional uniprocessor machines. Thus the design of schemes that enhance the fault tolerance capabilities of the algorithmic aspects of several problems that arise in conjunction with the design of such schemes will be done. The main thrust of this investigation is on some recently proposed on-line techniques that accomplish fault detection concurrently with normal computation. The main goals of this research are to develop practical and efficient schemes for fault detection and location and to obtain a better understanding of the issues involved in the design of fault tolerance schemes for multiprocessor systems.