TY - GEN
T1 - Fault-tolerance on Boolean n-cube architectures
AU - Yang, Chu Sing
AU - Wu, Shun Yue
N1 - Funding Information:
This work was supported by the National Science Council of the Republic of China under contract No. NSC83-0408-E-110-001.
Publisher Copyright:
© 1994, Springer Verlag. All rights reserved.
PY - 1994
Y1 - 1994
N2 - An approach to fault-tolerant Boolean n-cube architectures (FTBns) is proposed in this paper. We employ spares, including nodes, links and switches, to reconfigure a failed system so that system topology with its original dimension can be retained. The FTBn is designed in two levels. In the first level, we use a Boolean m-cube of 2m nodes with 2P, p ≤ m, spare nodes, and some switching elements to build a fault-tolerant module (FFM). Then an FTBn, n ≥ m, is built in the second level by taking 2n-m FTMs, and augmenting several switching elements between two adjacent FTMs. We will show that each FTM can achieve full spare utilization, and also that the degree of each node maintains a constant n. A two-phase reconfiguration algorithm is developed to allocate an adequate spare node to replace a faulty node. Finally, the reliability and costs of the FTBn are evaluated, and we then show that the FTBn can achieve higher or the same reliability as previous comparable systems at less extra hardware cost.
AB - An approach to fault-tolerant Boolean n-cube architectures (FTBns) is proposed in this paper. We employ spares, including nodes, links and switches, to reconfigure a failed system so that system topology with its original dimension can be retained. The FTBn is designed in two levels. In the first level, we use a Boolean m-cube of 2m nodes with 2P, p ≤ m, spare nodes, and some switching elements to build a fault-tolerant module (FFM). Then an FTBn, n ≥ m, is built in the second level by taking 2n-m FTMs, and augmenting several switching elements between two adjacent FTMs. We will show that each FTM can achieve full spare utilization, and also that the degree of each node maintains a constant n. A two-phase reconfiguration algorithm is developed to allocate an adequate spare node to replace a faulty node. Finally, the reliability and costs of the FTBn are evaluated, and we then show that the FTBn can achieve higher or the same reliability as previous comparable systems at less extra hardware cost.
UR - http://www.scopus.com/inward/record.url?scp=85027047226&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85027047226&partnerID=8YFLogxK
U2 - 10.1007/3-540-58426-9_157
DO - 10.1007/3-540-58426-9_157
M3 - Conference contribution
AN - SCOPUS:85027047226
SN - 9783540584261
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 546
EP - 559
BT - Dependable Computing - EDCC-1 - 1st European Dependable Computing Conference, Proceedings
A2 - Echtle, Klaus
A2 - Hammer, Dieter
A2 - Powell, David
PB - Springer Verlag
T2 - 1st European Dependable Computing Conference, EDCC-1 1994
Y2 - 4 October 1994 through 6 October 1994
ER -