DRD──基于诊断的高可靠分布式计算机系统的设计  

DRD──Design of a Diagnosis Based on High Reliable Distributed Computer Systems

作　　者：左德承[1] 高巍[1] 杨孝宗[1] 

机构地区：[1]哈尔滨工业大学计算机科学与工程系,黑龙江哈尔滨150001

出　　处：《计算机应用研究》2001年第4期23-25,共3页Application Research of Computers

基　　金：国家自然科学基金资助项目!(69873013)

摘　　要：提出了一种基于Ｆａｉｌ－ｓｉｌｅｎｔ节点故障行为的高可靠分布式计算机系统，提出了减少上述系统节点故障检测延迟的两条有效途径：利用单机有限的故障检测能力和增加任务特征状态比较。利用节点故障的诊断技术，将系统中的故障节点与正确节点隔离．通过节点故障恢复和系统重构，以提高整个系统的可靠性。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.

关 键 词：分布式计算机系统 设计 DRD 容错 故障诊断算法 

分 类 号：TP338.802[自动化与计算机技术—计算机系统结构]