Component Conditional Fault Tolerance Of Hierarchical Folded Cubic Networks

Xueli Sun,Jianxi Fan,Baolei Cheng,Zhao Liu,Jia Yu

DOI: https://doi.org/10.1016/j.tcs.2021.06.001

IF: 1.002

2021-01-01

Theoretical Computer Science

Abstract:For the sake of achieving higher reliability, conditional connectivity has gradually become well-known. Component connectivity, as a kind of conditional connectivity, is an extension of the classical connectivity. Given a simple and undirected graph G and a nonnegative integer r, the (r + 1)-component connectivity of graph G, say c kappa(r+1)(G), is the minimum number of a vertex cut whose removal causes the surviving graph to have at least r + 1 components. Another basic parameter of reliability, diagnosability is often related to the number of components in the surviving graph. The (r + 1)-component diagnosability, say ct(r+1)(G), is the maximum size of fault sets subject to at least r + 1 components in the surviving graph, provided that all faulty vertices can be detected. A graph is t/k-diagnosable if all faulty vertices can be isolated into a set in which up to k vertices are fault-free, provided that the number of faulty vertices is at most t. As a two-level interconnection network, the hierarchical folded cubic network HFQ (n) possesses a great deal of nice properties. In this paper, we first show that the n-dimensional hierarchical folded cubic network is tightly super connected. Then, we explore that c kappa(r+1)(HFQ(n)) = r(n + 1) - (r/2) + 1 (n >= 4, 1 <= r <= n - 4), and obtain that ct(r+1)(HFQ(n)) =(r+ 1)n - (r/2) + 2 (n >= 4, 1 <= r <= n - 4) under the PMC and MM* models. Last, we show that the hierarchical folded cubic network HFQ(n) is [(k + 1)n - (k/2) + 2]/k-diagnosable, where n >= 4 and 1 <= k <= n - 4. (C) 2021 Elsevier B.V. All rights reserved.