Skip to main content
SpringerLink
Log in

End-to-End Data-Flow Parallelism for Throughput Optimization in High-Speed Networks

  • Published:
Journal of Grid Computing Aims and scope Submit manuscript

Abstract

The increase in the data produced by large-scale scientific applications necessitates innovative solutions for efficient transfer of data. Although the current optical networking technology reached theoretical speeds of 100 Gbps, applications still suffer from inefficient transport protocols and bottlenecks on the end-systems (e.g. disk, CPU, NIC). High-performance systems provide us with parallel disks, processors and network interfaces. However the lack of orchestration of these end-system resources with the available network capacity results in underutilization of the network bandwidth. In this study, a model and two algorithms that use ‘end-to-end data-flow parallelism’ to optimize the use of network and end-system resources are proposed. This is achieved by using multiple parallel streams over the network; and multiple parallel disks and CPUs at the end systems. Our model predicts the optimal number of streams and disk/CPU stripes that maximizes the data transfer speed for any setting. Our algorithms use GridFTP parallel samplings and calculate the optimal level of parallelism based on our prediction model. The experiments conducted by using actual GridFTP transfers show that the predictions performed by our model and algorithms provide close-to-optimal performances with negligible overhead and use minimal number of resources. The end-to-end data transfer throughput is improved dramatically in existence of end-system bottlenecks compared to the non-optimized transfers.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

References

  1. Arra/ani testbed. https://sites.google.com/a/lbl.gov/ani-100g-network. Accessed 30 July 2012

  2. Fdt, fast data transfer. http://monalisa.cern.ch/FDT/. Accessed 30 July 2012

  3. FutureGrid. http://www.futuregrid.org/. Accessed 30 July 2012

  4. Super computing bandwidth challenge. http://sc09.supercomputing.org/?pg=bandwidth.html. Accessed 30 July 2012

  5. Udt, udp-based data transfer. http://udt.sourceforge.net/. Accessed 30 July 2012

  6. Ahuja, R., Magnanti, T., Orlin, J.: Network Flows. Prentice Hall (1993)

  7. Allcock, W., Bresnahan, J., Kettimuth, R., Link, M., Dumirescu, C., Raicu, I., Foster, I.: The globus striped Gridftp framework and server. In: Proceedings of the ACM/IEEE Conference on Supercomputing, p. 54 (2005)

  8. Altman, E., Barman, D., Tuffin, B., Vojnovic, M.: Parallel tcp sockets: simple model, throughput and validation. In: Proc. IEEE Conference on Computer Communications (INFOCOM’06), pp. 1–12 (2006)

  9. Chase, J.S., Gallatin, A.J., Yocum, K.G.: End system optimizations for high-speed TCP. IEEE Commun. Mag. 39(4), 68–74 (2000)

    Article  Google Scholar 

  10. Crowcroft, J., Oechslin, P.: Differentiated end-to-end internet services using a weighted proportional fair sharing TCP. ACM SIGCOMM Comput. Commun. Rev. 28(3), 53–69 (1998)

    Article  Google Scholar 

  11. Floyd, S.: Rfc3649: highspeed TCP for large congestion windows

  12. Gropp, W., Lusk, E., Thakur, R.: Using MPI-2: Advanced Features of the Message-Passing Interface. MIT Press (1999)

  13. Hacker, T.J., Noble, B.D., Atley, B.D.: The end-to-end performance effects of parallel TCP sockets on a lossy wide area network. In: Proc. IEEE International Symposium on Parallel and Distributed Processing (IPDPS’02), pp. 434–443 (2002)

  14. Hasegawa, G., Terai, T., Okamoto, T., Murata, M.: Scalable socket buffer tuning for high-performance web servers. In: International Conference on Network Protocols (ICNP01), p. 281 (2001)

  15. The Lustre file system. http://wiki.lustre.org. Accessed 30 July 2012

  16. Jain, M., Prasad, R.S., Davrolis, C.: The TCP bandwidth-delay product revisited: network buffering, cross traffic, and socket buffer auto-sizing. Tech. Rep., Georgia Institute of Technology (2003)

  17. Jin, C., Wei, D.X., Low, S.H., Buhrmaster, G., Bunn, J., Choe, D.H., Cottrell, R.L.A., Doyle, J.C., Feng, W., Martin, O., Newman, H., Paganini, F., Ravot, S., Singh, S.: Fast TCP: from theory to experiments. IEEE Netw. 19(1), 4–11 (2005)

    Article  Google Scholar 

  18. Kola, G., Kosar, T., Livny, M.: Run-time adaptation of Grid data-placement jobs. SCPE 6(3), 33–43 (2005)

    Google Scholar 

  19. LBNL: The distributed parallel storage system. http://www-didc.lbl.gov/DPSS. Accessed 30 July 2012

  20. Liu, W., Tieman, B., Kettimuthu, R., Foster, I.: A data transfer framework for large-scale science experiments. In: Proc. 19th ACM International Symposium on High-Performance Distributed Computing (HPDC’10) (2010)

  21. Lu, D., Qiao, Y., Dinda, P.A., Bustamante, F.E.: Modeling and taming parallel TCP on the wide area network. In: Proc. IEEE International Symposium on Parallel and Distributed Processing (IPDPS’05), p. 68b (2005)

  22. Prasad, R.S., Jain, M., Davrolis, C.: Socket buffer auto-sizing for high-performance data transfers. J. Grid Computing 1(4), 361–376 (2004)

    Article  Google Scholar 

  23. Pucha, H., Kaminsky, M., Andersen, D.G., Kozuch, M.A.: Adaptive file transfers for diverse environments. In: Proceedings of USENIX’08 (2008)

  24. Schmuck, F., Haskin, R.: Gpfs: a shared-disk file system for large computing clusters. In: Proceedings of the 1st Usenix Conference on File and Storage (FAST’02) (2002)

  25. Semke, J., Madhavi, J., Mathis, M.: Automatic tcp buffer tuning. In: ACM SIGCOMM’98, vol. 28(4), pp. 315–323 (1998)

  26. Stone, N., Gill, B., Kochmar, J., Light, R., Nowoczynski, P., Scott, J.R., Sommerfield, J., Vizino, C.: Dmover: parallel data migration for mainstream users. Tech. Rep., Pittsburgh Supercomputing Center (2010)

  27. Thomas, M.: Ultralight planets tutorial (2008)

  28. Weigle, E., Feng, W.: Dynamic right-sizing: a simulation study. In: Proc. IEEE International Conference on Computer Communications and Networks (ICCCN’01) (2001)

  29. Louisiana optical network initiative. http://www.loni.org. Accessed 30 July 2012

  30. The TeraGrid. http://www.teragrid.org. Accessed 30 July 2012

  31. Yildirim, E., Yin, D., Kosar, T.: Prediction of optimal parallelism level in wide area data transfers. IEEE Trans. Parallel Distrib. Syst. 22(12), 2033–2045 (2011)

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations

  1. State University of New York, 338H Davis Hall, Buffalo, NY, USA

    Esma Yildirim

  2. State University of New York, 338J Bell Hall, Buffalo, NY, USA

    Tevfik Kosar

Authors
  1. Esma Yildirim
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Tevfik Kosar
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Esma Yildirim.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Yildirim, E., Kosar, T. End-to-End Data-Flow Parallelism for Throughput Optimization in High-Speed Networks. J Grid Computing 10, 395–418 (2012). https://doi.org/10.1007/s10723-012-9220-9

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10723-012-9220-9

Keywords

Search

Navigation