Abstract
The performance of a physical cluster ultimately depends on two factors. One is the capability of individual computing nodes and the other is the networking speed among them. Recent processors are being greatly developed in terms of enhancing the processing speed while lowering the monetary cost. As for the networking technologies, nowadays dominant solutions have disadvantages such as high installation price and low protocol efficiency. Such drawbacks become the bottleneck of improving the ‘performance per cost’ ratio of the cluster as a whole. This paper proposes an alternative system interconnection device especially for application in small-scale clusters. The non-transparent bridges in PCI Express technology are employed to allow PCI Express packets to directly transmit across networked computing nodes. The performance is measured under two kinds of data transmission schemes, with two different benchmarking tools, respectively. Currently the proposed device delivers a peak unidirectional throughput of 8.6 gigabits per second.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
50 Years of Moore’s Law. http://www.intel.com/content/www/us/en/silicon-innovations/moores-law-technology.html
Interconnect Family System Share. http://www.top500.org/statistics/list/
Bencivenni, M., Bortolotti, D., Carbone, A., Cavalli, A., Chierici, A., et al.: Performance of 10 Gigabit ethernet using commodity hardware. IEEE Trans. Nucl. Sci. 57(2), 630–641 (2010)
Interconnect analysis: 10 GigE and InfiniBand in high performance computing. White Paper, HPC Advisory Council (2009)
Koop, M.J. Huang, W., Gopalakrishnan, K., Panda, D.K.: Performance analysis and evaluation of PCIe 2.0 and quad-data rate InfiniBand. In: 16th IEEE Symposium on High Performance Interconnects, pp. 85–92, August 2008
PCI Express Base Specification, Revision 3.0, PCI-SIG, November 2010
Ravindran, M.: Extending cabled PCI express to connect devices with independent PCI domains, SysCon 2008. In: IEEE International Systems Conference, Montreal, Canada, April 2008
Hanawa, T., Boku, T., Miura, S., Okamoto, T., Sato, M., et al.: Low-power and high-performance communication mechanism for dependable embedded systems. In: International Workshop on Innovative Architecture for Future Generation High-Performance Processors and Systems, pp. 67–73 (2008)
Hanawa, T., Kodama, Y., Boku, T., Sato, T.: Interconnection network for tightly coupled accelerators architecture. In: IEEE 21st Annual Symposium on High-Performance Interconnects, San Jose, USA, pp. 79–82, August 2013
Byrne, J., Chang, J., Lim, K.T., Ramirez, L., Ranganathan, P.: Power-efficient networking for balanced system designs: early experiences with PCIe In: HotPower 2011 Proceedings of the 4th Workshop on Power-Aware Computing and Systems, Article No. 3 2011
Krishnan, V.: Evaluation of an integrated PCI express IO expansion and clustering fabric. In: 16th IEEE Symposium on High Performance Interconnects, Stanford University, USA, pp. 93–100, August 2008
Ren, Y., Kim, Y.W., Kim, H.Y.: Implementation of system interconnection devices using PCI express. In: IEEE International Conference on Consumer Electronics, Las Vegas, USA, pp. 300–301, January 2015
Percival, D.: PCI express clustering. In: PCI-SIG Developers Conference, Israel (2011)
Bu-Khamsin, A.: Socket direct protocol over PCI express interconnect: design, implementation and evaluation. MS thesis, Simon Fraser University (2012)
Cooper, S.: Using PCIe over cable for high speed CPU-to-CPU communications. PCI-SIG Developers Conference (2008)
Acknowledgments
This work was supported by the ICT R&D program of MSIP/IITP. [10038768, The Development of Supercomputing System for the Genome Analysis].
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2016 ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering
About this paper
Cite this paper
Ren, Y., Kim, Y.W., Kim, H.Y. (2016). A System Interconnection Device for Small-Scale Clusters. In: Zhang, Y., Peng, L., Youn, CH. (eds) Cloud Computing. CloudComp 2015. Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering, vol 167. Springer, Cham. https://doi.org/10.1007/978-3-319-38904-2_36
Download citation
DOI: https://doi.org/10.1007/978-3-319-38904-2_36
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-38903-5
Online ISBN: 978-3-319-38904-2
eBook Packages: Computer ScienceComputer Science (R0)