Skip to main content
Log in

Design issues and performance comparisons in supporting the sockets interface over user-level communication architecture

  • Published:
The Journal of Supercomputing Aims and scope Submit manuscript

Abstract

Since user-level communication (ULC) architecture provides only primitive operations for application programmers, there have been several researches to build a portable and standard communication interface, such as sockets, on top of ULC architecture. Basically there are three different approaches to supporting the sockets interface over ULC architecture: LAN emulation, a user-level sockets, and a kernel-level sockets. The primary objective of this paper is to compare these approaches in terms of their design, implementation, and performance.

We have developed and implemented a kernel-level sockets layer over ULC architecture, since there is currently no available implementation. We also present different design and implementation decisions on data receiving, data sending, connection management, etc. in the three approaches. Through the performance comparison, we show that LAN emulation approach exhibits the worst performance both in latency and bandwidth. Our experiments also show that a user-level sockets is useful for latency-sensitive applications and a kernel-level sockets is effective for applications which require high bandwidth and full compatibility with the legacy sockets interface.

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

Access this article

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

  1. Balaji P, Narravula S, Vaidyanathan K, Krishnamoorthy S, Wu J, Panda DK (2004) Sockets direct protocol over InfiniBand in clusters: is it benefitial? In: Proceedings of the IEEE international symposium on performance analysis of systems and software, IEEE Computer Society Press, Los Alamitos, pp 28–35

    Chapter  Google Scholar 

  2. Balaji P, Shivan P, Wyckoff P, Panda D (2002) High performance user level sockets over gigabit ethernet. In: Proceedings of the IEEE international conference on cluster computing, IEEE Computer Society Press, Los Alamitos, pp 179–186

    Chapter  Google Scholar 

  3. Baker M (ed) (December 2000) Cluster computing white paper (Version 2.0). IEEE Task Force on Cluster Computing

  4. Bozeman P, Saphir B (1999) A modular high performance implementation of the Virtual Interface Architecture. In: Proceedings of extreme linux conference

  5. Buonadonna P, Geweke A, Culler D (1998) An implementation and analysis of the Virtual Interface Architecture. In: Proceedings of ACM/IEEE international conference on supercomputing (CDROM), IEEE Computer Society Press, Los Alamitos, pp 1–15

    Google Scholar 

  6. Cameron D, Regnier G (2002) The virtual interface architecture. Intel Press

  7. Chen Y, Wang X, Jiao Z, Xie J, Du Z, Li S (2002) MyVIA: a design and implementation of the high performance Virtual Interface Architecture. In: Proceedings of the IEEE international conference on cluster computing, IEEE Computer Society Press, Los Alamitos, pp 160–167

    Chapter  Google Scholar 

  8. Compaq Corporation, Intel Corporation, and Microsoft Corporation (1997) Virtual interface architecture specification, Version 1.0

  9. Emulex Corporation-Products. http://www.emulex.com/products/legacy/vi/clan1000.html

  10. Emulex Inc (2001) cLAN for Linux. Software User’s Guide

  11. Fischer M (2001) GMSOCKS—a direct sockets implementations on myrinet. In: Proceedings of the IEEE international conference on cluster computing, IEEE Computer Society Press, Los Alamitos, pp 204–211

    Google Scholar 

  12. InfiniBand Linux SourceForge Project: http://infiniband.sourceforge.net

  13. InfiniBand Trade Association. http://www.infinibandta.org

  14. Itoh M, Ishizaki T, Kishimoto M (2000) Accelerated socket communication in system area network. In: Proceedings of the IEEE international conference on cluster computing, IEEE Computer Society Press, Los Alamitos, pp 357–358

    Google Scholar 

  15. Kim J-S, Kim K, Jung S-I, Ha S. (2003) Design and implementation of a user-level sockets layer over Virtual Interface Architecture. Concurr Comput Pract Exp 15(7–8):727–749

    Article  Google Scholar 

  16. Liu J, Wu J, Panda DK (2003) High performance RDMA-based MPI implementation over InfiniBand. In: Proceedings of the 17th annual international conference on supercomputing, ACM Press, New York, pp 295–304

    Chapter  Google Scholar 

  17. McKusick MK, Bostic K, Karels MJ, Quarterman JS (1996) The design and implementation of the 4.4 BSD operating system. Addison-Wesley, Reading

    Google Scholar 

  18. MVICH—MPI for the Virtual Interface Architecture. http://old-www.nersc.gov/research/FTG/mvich

  19. OpenIB.org. http://www.openib.org

  20. Seifert F, Kohmann H SCI SOCKET—a fast socket implementation over SCI. White paper, Dophin Interconnect Solutions Inc

  21. TOP500 Supercomputer sites. http://www.top500.org

  22. Yu J-L, Lee M-S, Maeng S-R (2001) An efficient implementation of Virtual Interface Architecture using adaptive transfer mechanism on myrinet. In: Proceedings of international conference on parallel and distributed systems, IEEE Computer Society Press, Los Alamitos, pp 741–747

    Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Jae-Wan Jang.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Jang, JW., Kim, JS. Design issues and performance comparisons in supporting the sockets interface over user-level communication architecture. J Supercomput 39, 205–226 (2007). https://doi.org/10.1007/s11227-007-0109-5

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11227-007-0109-5

Keywords

Navigation