SummaryProcessor fault diagnosis takes a key role in fault-tolerant computing on multiprocessor systems.The t∕k diagnosis strategy which is a generalization of the precise and pessimistic diagnosis strategies can significantly improve the self-diagnosing capability of the system. Using this tool, it is possible to deal with large faults in the system. This paper presents a t∕k diagnosis algorithm on n-dimensional hypercube-like networks (include Hypercubes, Crossed cubes, Möbius cubes, Locally Twisted cubes, and Twisted cubes) for any k ∈ [0, n − 2]. The algorithm can correctly identify all nodes except at most k nodes undiagnosed. It runs in O(N) time, where N = 2 n is the total number of nodes of n-dimensional hypercube-like networks. To the best of our knowledge, in the case k ≥ 4, there is no known t∕k diagnosis algorithm for general diagnosable system or any specific system.
KEYWORDShypercube-like networks, large connected component, system-level diagnosis, t∕k diagnosis strategy
INTRODUCTIONWith the rapid development of very large scale integration (VLSI) technology, a multiprocessor system may contain hundreds or even thousands of processors (nodes). As the number of processors increases, so does the expected number of processors being faulty at the same time. To ensure reliability, the system should have the ability to identify the faulty processors which are then isolated from the system or replaced by additional fault-free ones. The technique of identifying faulty processors by conducting tests on the processors and interpreting the test outcomes is known as system-level diagnosis.Preparata et al. 1 proposed the first model of system-level diagnosis, known as the PMC model. It assumes that each node in a system can conduct tests on its neighbors via the communication links between them. When a node tests its neighbor, the tester declares the tested node to be fault-free or faulty based on the test responses. This evaluation is reliable if and only if the testing node is fault-free.There are two classical diagnosis strategies, i.e., the precise diagnosis strategy 1 and the pessimistic diagnosis strategy. 2,3 However, due to their limited diagnosability, they are unable to deal with large faults in large multiprocessor systems. In order to greatly enhance the degree of diagnosability, Somani and Peleg proposed a new diagnosis strategy known as the t∕k diagnosis. 4 Under this strategy, all the faulty nodes can be isolated to within a set in which at most k nodes is fault-free. Specifically, a system S of N nodes is t∕k-diagnosable if, given any test syndrome produced by the system under the presence of a fault set F, all the faulty nodes can be isolated to within a set F ′ , out of which at most k nodes can possibly be fault-free, |F ′ | ≤ |F| + k, provided the number of faulty nodes does not exceed t. By definition, if k = 0, the t∕k diagnosis reduces to the precise diagnosis; if k = 1, the t∕k diagnosis is closely related to the pessimistic diagnosis. By choosing a larger k, the t∕k diagnosis can signif...