Advertisement

Utilizing PVM in a Multidomain Clusters Environment

  • Mario Petrone
  • Roberto Zarrelli
Part of the Lecture Notes in Computer Science book series (LNCS, volume 3666)

Abstract

A cluster is often configured with computational resources where there is only one IP visible front-end machine that hides all its internal machines from the external world. Considering clusters located in different network domains to be used in a computation, it is difficult to exploit all the internal machines of each cluster. This paper presents a PVM extension that enables us to exploit clusters in a multidomain environment so that each clustered machine can take part in a PVM computation. To improve the system performance, the PVM inter-task standard communications has been replaced by a method based on UDP sockets. Moreover, the existing code written for PVM can easily be ported to use these features.

Keywords

Virtual Machine Round Trip Delay Cluster Grid High Performance Computing Application Internal Machine 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Top 500 supercomputer site, http://www.top500.org
  2. 2.
    Gentzsch, W.: Grid Computing, A Vendor’s Vision. In: Procs of the 2nd IEEE/ACM Intl. Symposium on Cluster Computing and the Grid, Berlin, Germany (May 2002)Google Scholar
  3. 3.
    W3C Consortium, Web Services activity, http://www.w3c.org/2002/ws
  4. 4.
    Agrawal, S., Dongarra, J., et al.: NetSolve: Past, Present, and Future - a Look at a Grid Enabled Server. In: Berman, F., Fox, G., Hey, A. (eds.) Grid Computing: Making the Global Infrastructure a Reality. Wiley Publisher, Chichester (2003)Google Scholar
  5. 5.
    Foster, I., Karonis, N.: A grid-enabled MPI: Message passing in heterogeneous distributed computing systems. In: Supercomputing 1998, Orlando, FL (1998)Google Scholar
  6. 6.
    Karonis, N., Toonen, B., Foster, I.: MPICH-G2: A grid-enabled implementation of the Message Passing Interface. Journal of Parallel and Distributed Computing (May 2003)Google Scholar
  7. 7.
    Keller, R., Krammer, B., et al.: MPI development tools and applications for the grid. In: Workshop on Grid Applications and Programming Tools, Seattle, WA (June 2003)Google Scholar
  8. 8.
    Kurzyniec, D., Wrzosek, T., et al.: Towards Self-Organizing Distributed Computing Frameworks: The H2O Approach. In: Parallel Processing Letters (2003)Google Scholar
  9. 9.
    Hwang, P., Kurzyniec, D., Sunderam, V.: Heterogeneous parallel computing across multidomain clusters. In: Procs. of the 11th EuroPVM/MPI Conference, Budapest, Hungary (September 2004)Google Scholar
  10. 10.
    Kurzyniec, D., Hwang, P., Sunderam, V.: Failure Resilient Heterogeneous Parallel Computing Across Multidomain Clusters. International Journal of High Performance Computing Applications (2005); Special Issue: Best Papers of EuroPVM/MPI 2004Google Scholar
  11. 11.
    Geist, A., Beguelin, A., et al.: PVM: Parallel Virtual Machine. In: A Users’ Guide and Tutorial for Networked Parallel Computing. The MIT Press, Cambridge (1994)Google Scholar
  12. 12.
    Beguelin, A., Dongarra, J., et al.: Recent Enhancements to PVM. Intl. Journal for Supercomputer Applications (1995)Google Scholar
  13. 13.
    Springer, P.L.: PVM Support for Clusters. In: IEEE 3rd Intl. Conf. on Cluster Computing, Newport Beach, California, USA (October 2001)Google Scholar
  14. 14.
    Frattolillo, F.: A PVM Extension to Exploit Cluster Grids. In: Procs. of the 11th EuroPVM/MPI Conference, Budapest, Hungary (September 2004)Google Scholar
  15. 15.
    Frattolillo, F.: Exploiting PVM to Run Large-Scale Applications on Cluster Grids. International Journal of High Performance Computing Applications (2005); Special Issue: Best Papers of EuroPVM/MPI 2004Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2005

Authors and Affiliations

  • Mario Petrone
    • 1
  • Roberto Zarrelli
    • 1
  1. 1.University of MoliseItaly

Personalised recommendations