A token-bus-based design method of the distributedfault-tolerant industrial network is presented in this pa-per.The dual-link network is of hot-redundancy.The performance of the network is also discussed.
Abstract In this paper, the problem of fault tolerant routings in fault tolerant networks is considered. A routing in a network assigns to each ordered pair of nodes a fixed path. All communication among nodes must ...Abstract In this paper, the problem of fault tolerant routings in fault tolerant networks is considered. A routing in a network assigns to each ordered pair of nodes a fixed path. All communication among nodes must go on this routing. When either a node or a link in a fault tolerant network fails, the communication from one node to another using this faulty element must be sent via one or more intermediate nodes along a sequence of paths determined by this routing. An important and practical problem is how to choose a routing in the network such that intermediate nodes to ensure communication are small for any fault set. Let C d be a directed cycle of order d . In this paper. The author first discusses connectivity of Cartesian product digraphs, then proves that the Cartesian product digraph C d 1 ×C d 2 ×...×C d n (d i≥2,1≤i≤n) has a routing such that at most one intermediate node is needed to ensure transmission of messages among all non faulty nodes so long as the number of faults is less than n . This is a generalization of Dolev et al's result for the n dimensional cube.展开更多
Let G be a k-connected simple graph with order n. The k-diameter, combining connectivity with diameter, of G is the minimum integer d k(G) for which between any two vertices in G there are at least k internally verte...Let G be a k-connected simple graph with order n. The k-diameter, combining connectivity with diameter, of G is the minimum integer d k(G) for which between any two vertices in G there are at least k internally vertex-disjoint paths of length at most d k(G). For a fixed positive integer d, some conditions to insure d k(G)≤d are given in this paper. In particular, if d≥3 and the sum of degrees of any s (s =2 or 3) nonadjacent vertices is at least n+(s-1)k+1-d, then d k(G)≤d. Furthermore, these conditions are sharp and the upper bound d of k-diameter is best possible.展开更多
文摘A token-bus-based design method of the distributedfault-tolerant industrial network is presented in this pa-per.The dual-link network is of hot-redundancy.The performance of the network is also discussed.
文摘Abstract In this paper, the problem of fault tolerant routings in fault tolerant networks is considered. A routing in a network assigns to each ordered pair of nodes a fixed path. All communication among nodes must go on this routing. When either a node or a link in a fault tolerant network fails, the communication from one node to another using this faulty element must be sent via one or more intermediate nodes along a sequence of paths determined by this routing. An important and practical problem is how to choose a routing in the network such that intermediate nodes to ensure communication are small for any fault set. Let C d be a directed cycle of order d . In this paper. The author first discusses connectivity of Cartesian product digraphs, then proves that the Cartesian product digraph C d 1 ×C d 2 ×...×C d n (d i≥2,1≤i≤n) has a routing such that at most one intermediate node is needed to ensure transmission of messages among all non faulty nodes so long as the number of faults is less than n . This is a generalization of Dolev et al's result for the n dimensional cube.
文摘Let G be a k-connected simple graph with order n. The k-diameter, combining connectivity with diameter, of G is the minimum integer d k(G) for which between any two vertices in G there are at least k internally vertex-disjoint paths of length at most d k(G). For a fixed positive integer d, some conditions to insure d k(G)≤d are given in this paper. In particular, if d≥3 and the sum of degrees of any s (s =2 or 3) nonadjacent vertices is at least n+(s-1)k+1-d, then d k(G)≤d. Furthermore, these conditions are sharp and the upper bound d of k-diameter is best possible.
