Abstract
To lower the large-scale network cost and energy consumption, we proposed a hierarchical topology with low-radix switches. The hierarchical topology HyperFatTree is designed by combining Fat Tree topology and complete graph topology. Theory analysis of topology characteristic shows that the proposed topology can achieve high cost performance and high scalability. According to the characteristic of HyperFatTree, we designed a minimum path routing algorithm and a load balanced non-minimum path random routing algorithm for the WorstCase traffic. Evaluation shows that the saturated throughput of HyperFatTree is nearly 68.5 % at the scale of 83,232 nodes under Uniform Random traffic, which is higher than both Fat Tree and Dragonfly. It is also important that the throughput reduces only 5 % while the scale has increased 80 times, which shows good smooth scalability. Evaluation on energy efficiency shows that HyperFatTree performs better than Fat Tree and can achieve 3–7 times higher energy efficiency than Dragonfly.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Agarwal, A.: Limits on interconnection network performance. IEEE Trans. Parallel Distrib. Syst. 2(4), 398–412 (1991)
Johnson, K.L.: The impact of communication locality on large-scale multiprocessor performance. Comput. Archit. News 20, 392–402 (1992)
Barker, K.J., Benner, A., Hoare, R., et al.: On the feasibility of optical circuit switching for high performance computing systems. In: Proceedings of the IEEE Supercomputing Conference, pp. 12–18. Seattle, WA (2005)
Delimitrou, C., Sankar, S., Kansal, A., et al.: ECHO: recreating network traffic maps for data-centers with tens of thousands of servers. In: Proceedings of the IEEE International Symposium on Workload Characterization, pp. 14–24 (2012)
Duato, J.: A necessary and sufficient condition for deadlock-free adaptive routing in wormhole networks. IEEE Trans. Parallel Distrib. Syst. 6, 1055–1067 (1995)
Valiant, L.G.: A scheme for fast parallel communication. SIAM J. Comput. 11(2), 350C361 (1982)
Fan, Z., Cao, Z., Su, Y.: HiNetSim: A parallel simulator for large-scale hierarchical direct networks. In: Proceedings of the 11th IFIP WG International Conference, NPC, pp. 120–131. Ilan, Taiwan (2014)
Vetter, J.S., Mueller, F.: Communication characteristics of large-scale scientific applications for contemporary cluster architectures. In: Proceedings of the 16th International Parallel and Distributed Processing Symposium (IPDPS) (2002)
Arista Networks: Arista 7148SX Switch, Avaiable: http://www.aristanetworks.com/en/7100_Series_SFPSwitches
Leiserson, C.E.: Fat-trees: universal networks for hardware-efficient supercomputing. IEEE Trans. Comput. 34, 892–901 (1985)
Al-Fares, M., Loukissas, A., Vahdat, A.: A scalable, commodity data center network architecture. In: Proceedings of the ACM SIGCOMM 2008 Conference on Data Communication, SIGCOMM ’08, pp. 63–74. New York , NY, USA (2008)
Farrington, N., Rubow, E., Vahdat, A.: Data center switch architecture in the age of merchant silicon. In: 17th IEEE Symposium on High Performance Interconnects, pp. 93–102. IEEE (2009)
Kim, J., Dally, W., Abts, D.: Flattened buttery: a cost-efficient topology for high-radix networks. In: Proceedings of the 34th Annual International Symposium on Computer Architecture, pp. 126–137. ACM (2007)
Kim, J., Dally, W., Scott, S., Abts, D.: Technology-driven, highly-scalable dragony topology. In: Proceedings of the 35th Annual International Symposium on Computer Architecture, pp. 77–88. IEEE Computer Society (2008)
Chen, D., Eisley, N.A., Heidelberger, P., et al.: The IBM Blue Gene/Q interconnection net-work and message unit, In: Proceedings of 2011 International Conference for High Performance Computing, Networking, Storage and Analysis, 26 (2011)
Ajima, Y., Sumimoto, S., Shimizu, T.: Tofu: a 6D mesh/torus interconnect for ex-ascale computers. IEEE Comput. 11, 36–40 (2009)
Faanes, G. et al.: Cray Cascade: a scalable HPC system based on a dragonfly network. In: International Conference for High Performance Computing, Networking, Storage and Analysis, SC (2012)
Acknowledgments
This Project is supported by the National Natural Science Foundation of China (Grant No. 61572464) and the National High Technology Research and Development Program of China (863 Program) (Grant No. 2015AA01A301).
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Su, Y., Wang, Z., Fan, Z. et al. HyperFatTree: A Large-Scale Tree-Based Network with Low-Radix Switches. Int J Parallel Prog 45, 172–184 (2017). https://doi.org/10.1007/s10766-015-0393-2
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10766-015-0393-2