k-pairwise disjoint paths routing in perfect hierarchical hypercubes (original) (raw)
Abstract
Hierarchical hypercubes (HHC), also known as cube-connected cubes, have been introduced in the literature as an interconnection network for massively parallel systems. Effectively, they can connect a large number of nodes while retaining a small diameter and a low degree compared to a hypercube of the same size. Especially (2_m_+m)-dimensional hierarchical hypercubes (\(\mathit {HHC}_{2^{m}+m}\)), called perfect HHCs, are popular as they are symmetrical, which is a critical property when designing routing algorithms. In this paper, we describe an algorithm finding, in an \(\mathit{HHC}_{2^{m}+m}\), mutually node-disjoint paths connecting k_=⌈(m+1)/2⌉ pairs of distinct nodes. This problem is known as the k_-pairwise disjoint-path routing problem and is one of the important routing problems when dealing with interconnection networks. In an \(\mathit{HHC}_{2^{m}+m}\), our algorithm finds paths of lengths at most 2_m+1+m(2_m+1+1)+4 in O(25_m_) time, where 2_m_+1 is the diameter of an \(\mathit{HHC}_{2^{m}+m}\). Also, we have shown through an experiment that, in practice, the lengths of the generated paths are significantly lower than the worst-case theoretical estimations.
Access this article
Subscribe and save
- Starting from 10 chapters or articles per month
- Access and download chapters and articles from more than 300k books and 2,500 journals
- Cancel anytime View plans
Buy Now
Price excludes VAT (USA)
Tax calculation will be finalised during checkout.
Instant access to the full article PDF.
Similar content being viewed by others
References
- TOP500 (2011) Japan’s K computer tops 10 petaflop/s to stay atop TOP500 List. http://www.top500.org/lists/2011/11/, November 2011. Last accessed December 2012
- Laudon J, Lenoski D (1997) System overview of the SGI origin 200/2000 product line. In: Proc IEEE compcon ’97, San Jose, CA, USA, pp 150–156
Google Scholar - Yang X, Megson GM, Evans DJ (2006) An oblivious shortest-path routing algorithm for fully connected cubic networks. J Parallel Distrib Comput 66(10):1294–1303
Article MATH Google Scholar - Lai P-L, Hsu H-C, Tsai C-H, Stewart IA (2010) A class of hierarchical graphs as topologies for interconnection networks. Theor Comput Sci 411(31–33):2912–2924
Article MATH MathSciNet Google Scholar - Gu Q-P, Peng S (1998) An efficient algorithm for _k_-pairwise disjoint paths in star graphs. Inf Process Lett 67(6):283–287
Article MathSciNet Google Scholar - Akers SB, Krishnamurthy B (1989) A group theoretic model for symmetric interconnection networks. IEEE Trans Comput 38(4):555–566
Article MATH MathSciNet Google Scholar - Duh D-R, Chen G-H, Fang J-F (1995) Algorithms and properties of a new two-level network with folded hypercubes as basic modules. IEEE Trans Parallel Distrib Syst 6(7):714–723
Article Google Scholar - Malluhi QM, Bayoumi MA (1994) The hierarchical hypercube: a new interconnection topology for massively parallel systems. IEEE Trans Parallel Distrib Syst 5(1):17–30
Article MathSciNet Google Scholar - Wu J, Sun X-H (1994) Optimal cube-connected cube multicomputers. J Microcomput Appl 17(2):135–146
Article Google Scholar - Ghose K, Desai KR (1989) The HCN: a versatile interconnection network based on cubes. In: Proc 1989 ACM/IEEE conf on supercomputing, Reno, Nevada, USA, pp 426–435
Chapter Google Scholar - De Bruijn NG (1946) A combinatorial problem. Proc K Ned Akad Wet 49:758–764
MATH Google Scholar - Wu R-Y, Chen G-H, Kuo Y-L, Chang GJ (2007) Node-disjoint paths in hierarchical hypercube networks. Inf Sci 177(19):4200–4207
Article MATH MathSciNet Google Scholar - Bossard A, Kaneko K, Peng S (2011) A new node-to-set disjoint-path algorithm in perfect hierarchical hypercubes. Comput J 54(8):1372–1381
Article Google Scholar - Bossard A, Kaneko K (2012) The set-to-set disjoint-path problem in perfect hierarchical hypercubes. Comput J 55(6):769–775
Article Google Scholar - Shiloach Y (1978) The two paths problem is polynomial. Technical Report CS-TR-78-654, Stanford University
- Karp RM (1975) On the computational complexity of combinational problems. Networks 5:45–68
MATH MathSciNet Google Scholar - Gu Q-P, Peng S (1997) _k_-pairwise cluster fault tolerant routing in hypercubes. IEEE Trans Comput 46(9):1042–1049
Article MathSciNet Google Scholar - Seitz CL (1985) The cosmic cube. Commun ACM 28(1):22–33
Article MathSciNet Google Scholar
Acknowledgements
The authors sincerely thank the reviewers, especially Reviewers 1, 2, 4, 5, 6, and 9, for their insightful comments and suggestions that greatly improved the quality of this paper.
Author information
Authors and Affiliations
- Graduate School of Engineering, Tokyo University of Agriculture and Technology, 2-24-16 Nakacho, Koganei, Tokyo, Japan
Antoine Bossard & Keiichi Kaneko
Authors
- Antoine Bossard
- Keiichi Kaneko
Corresponding author
Correspondence toAntoine Bossard.
Appendix: Solving the case m=2
Appendix: Solving the case _m_=2
We have restricted the algorithm of Sect. 3 to _m_≥3 to be able to apply Lemma 1 in Step 2.
We are in the case _m_=2, which means that _k_=⌈(m+1)/2⌉=2 pairs c 1=(s 1,d 1) and c 2=(s 2,d 2) need to be connected. We mark each subcube Q m (σ) with |T(σ)|≥2 as a fault and define F the set of faulty subcubes. So there are at most _k_=2 faults. If only one subcube is marked as a fault, then both k_≤2_m_−1 and |F|≤_k_≤2_m_−2_k+1 hold, so we can apply Lemma 1 in Step 2 of the algorithm of Sect. 3.
Assume that two subcubes, say Q m (σ 1) and Q m (σ 2), are marked as faults. If c 1⊂Q m (σ 1), then c 2⊂Q m (σ 2), and it is trivial to connect the nodes of c 1 (resp. c 2) by using a shortest path inside Q m (σ 1) (resp. Q m (σ 2)). The case c 1⊂Q m (σ 2) is handled similarly. Now assume that Q m (σ 1) and Q m (σ 2) both contain one node of c 1 and one node of c 2, say s 1,d 2∈Q m (σ 1) and s 2,d 1∈Q m (σ 2). For any node u_=〈_σ u ,π u 〉, let \(u^{\uparrow}= \langle\sigma_{u}\oplus2^{\pi_{u}}, \pi_{u}\rangle\) be the unique neighbor of u that is not inside Q m (σ u ).
Select the edge \(s_{1}\rightarrow s_{1}^{\uparrow}(=s_{1}')\) if \(s_{1}^{\uparrow}\notin Q_{m}(\sigma_{2})\) and the two edges \(s_{1}\rightarrow u\rightarrow u^{\uparrow}(=s_{1}')\) with u a neighbor of s 1 in Q m (σ 1)∖{d 2} otherwise. Say \(s_{1}'\in Q_{m}(\sigma_{s_{1}'})\). Select the edge \(s_{2}\rightarrow s_{2}^{\uparrow}(=s_{2}')\) if \(s_{2}^{\uparrow}\notin Q_{m}(\sigma_{1})\) and \(s_{2}^{\uparrow}\notin Q_{m}(\sigma_{s_{1}'})\), and the two edges \(s_{2}\rightarrow v\rightarrow v^{\uparrow}(=s_{2}')\) with v a neighbor of s 2 in Q m (σ 2)∖{d 1} otherwise. Say \(s_{2}'\in Q_{m}(\sigma_{s_{2}'})\). By Lemma 1 we can find in \(Q_{2^{m}}\) two disjoint paths connecting \(\sigma_{s_{1}'}\) to σ 2 and \(\sigma_{s_{2}'}\) to d 2. These two \(Q_{2^{m}}\) paths are converted to paths in \(\mathit {HHC}_{2^{m}+m}\) by using the algorithm CONV (see Fig. 7).
Fig. 7
Two mutually node-disjoint paths s 1⇝d 1 and s 2⇝d 2 in an HHC 6 (_m_=2)
Rights and permissions
About this article
Cite this article
Bossard, A., Kaneko, K. _k_-pairwise disjoint paths routing in perfect hierarchical hypercubes.J Supercomput 67, 485–495 (2014). https://doi.org/10.1007/s11227-013-1013-9
- Published: 19 September 2013
- Issue date: February 2014
- DOI: https://doi.org/10.1007/s11227-013-1013-9