DRD──Design of a Diagnosis Based on High Reliable Distributed Computer Systems
Zuo De · Jisuanji yingyong yanjiu · 2001
This paper presents a fail-silent fault behavior based on high reliable distributed computer system, and uses two effective ways to reduce the fault detection latency: making use of the limited capability of fault detection of a single processor, and increasing the comparison of task signature state. With diagnosis of faulty nodes, all faulty processors are removed from the system service. To make the system more reliable, the faulty nodes must be recovered or the system must be reconfigured.