Fault Tolerance Analysis Of Complete Josephus Cubic

Fault tolerance is especially important for multiprocessor systems, since the growing size of the networks increases its vulnerability to component failures. Pro-cessor fault diagnosis plays an important role in multiprocessor systems for reliable computing, and the fault diagnosis of many well-known networks has been explored deeply.Josephus Cubic network has improved topological and exhibits better embed-ding and communications performance than the Binary Hypercube and several of its variants. Its link-augmented form, Complete Josephus Cubes, can also be ap-plied as node cluster in an optical-based architecture suitable for large-scale hierar-chical networks. This thesis mainly explores fault tolerance of complete Josephus cubic networks, some kinds of fault diagnosabilities of complete Josephus cubic net-works under the comparison model and PMC testing model. In addition, this thesis discusses the fault-tolerant path and cycle embedding in complete Josephus cubic networks.In the introduction, we introduce the research background and significance of network construction, fault tolerance, as well as system-level diagnosis. There are five chapters in the thesis.In the first chapter, we mainly introduce some basic knowledge, terminologies on graph theory and combinatorial network theory as well as theory of fault diagnosis. It also shows the construction and some basic properties of the network structure based on complete Josephus cubic network, which will be investigated in this thesis.In the second chapter, the fault tolerance analysis, which pave the way for the subsequent studies, of the complete Josephus cubic network is given.In the third chapter, we investigate classical diagnosability, conditional diag-nosability, local diagnosability and strong diagnosability of complete Josephus cubic networks under the comparison model and PMC testing model, respectively.In the fourth chapter, we explore the fault-tolerant path and cycle embedding in complete Josephus cubic networks.The fifth chapter gives concluding remarks and proposes some constructive but unsolved problems.