Title :
Multiprocessor fault diagnosis under local constraints
Author :
Das, A. ; Thulasiraman, K. ; Agarwal, V.K. ; Lakshmanan, K.B.
Author_Institution :
Montreal Univ., Que., Canada
fDate :
8/1/1993 12:00:00 AM
Abstract :
The authors study the fault diagnosis of multiprocessor systems when fault constraints in the local domain of each processor are specified. They use the comparison-based model. A multiprocessor system S is t-in-L diagnosable if, given a syndrome, all faulty processors can be uniquely identified provided there are at most t faulty processors in the local domain L(u i) ∪ {ui} of every processor, ui in S, where L (ui ) denotes the set of processors adjacent to ui. Certain basic results that lead to efficient conditions for unique diagnosis of a system when certain fault constraints are satisfied in the local domain of each processor in the system are presented. The t-in-L diagnosability of certain regular interconnected systems is examined under the assumption that less than half of the processors in the system are faulty. Diagnosis algorithms for these systems are presented
Keywords :
fault tolerant computing; multiprocessing systems; comparison-based model; fault constraints; local constraints; local domain; multiprocessor fault diagnosis; regular interconnected systems; t-in-L diagnosable; Distributed algorithms; Fault diagnosis; Graph theory; Hypercubes; Interconnected systems; Multiprocessing systems; Multiprocessor interconnection networks; Semiconductor device modeling; Very large scale integration; Wafer scale integration;
Journal_Title :
Computers, IEEE Transactions on