Advertisement

Models and internals of the IANOS resource broker

  • Vincent Keller
  • Hassan Rasheed
  • Oliver Wäldrich
  • Wolfgang Ziegler
  • Ralf Gruber
  • Marie-Christine Sawley
  • Philipp Wieder
Special Issue Paper

Abstract

We present the Intelligent Application Oriented System (IANOS) resource broker models and internals. The aim of IANOS is to provide an understanding of and a solution to the problem of how to find the best resource for a given submitted application in order to optimally use a set of HPC resources (an HPCN Grid). Best in the sense “at a given moment and for the requirements of that specific application”. The heart of IANOS is a cost model that minimizes the overall submission cost such as execution time cost, waiting time cost, licenses cost, energy cost, etc. To estimate the execution time cost, a model predicting the performance of an application on a resource where INAOS is installed is provided. This model is based on a parameterization of the application and the resources. IANOS has been deployed and tested successfully on an international testbed across Switzerland and Germany.

Keywords

HPCN Application  Resource Brokering   Cost Model  Grid Efficiency 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Berman F, Wolski R, Figueira S, Schopf J, Shao G (1996) Application-level scheduling on distributed heterogeneous networks, Proceedings of the 1996 ACM/IEEE conference on Supercomputing, pp 39–45Google Scholar
  2. 2.
    Braun TD, Siegel HJ, Beck N, Bölöni LL, Maheswaran M, Reuther AI, Robertson JP, Theys MD, Yao B, Hensgen D, Freund RF (2001) A comparison of eleven static heuristics for mapping a class of independent tasks onto heterogeneous distributed computing systems. J Parall Distrib Comput 61(6):810–837CrossRefGoogle Scholar
  3. 3.
    CSCS: Swiss National Supercomputing Centre (2006) Annual Report, CSCS publication, Manno (Switzerland)Google Scholar
  4. 4.
    Dubois-Pélerin Y, Van Kemenade V, Deville MO (1999) An object-oriented toolbox for spectral element analysis. J Sci Comput 14(1):1–29zbMATHCrossRefMathSciNetGoogle Scholar
  5. 5.
    Gruber R, Volgers P, De Vita A, Stengel M, Tran T-M (2003) Parameterisation to tailor commodity clusters to applications. Future Gen Comput Syst 19:111–120zbMATHCrossRefGoogle Scholar
  6. 6.
    Keller V (2008) Optimal Application-Oriented Resource Brokering in a High Performance Computing Grid (PhD thesis), 4221, EPFL, Lausanne (Switzerland)Google Scholar
  7. 7.
    Menu J, Jaunin M (2008) EPFL DIT HPC Clusters – batch statistics, website, http://dithpcbatch.epfl.ch/BatchStatistics.html, (accessed on 31 July 2008)
  8. 8.
    Rasheed H, Gruber R, Keller V, Ziegler W, Wäldrich O, Wieder P, Kuonen P, Sawley M-C, Maffioletti S, Kunszt P (2008) IANOS: An Intelligent Application Oriented Scheduling Middleware For An HPCN Grid, CoreGRID serie Grid Computing, Springer, pp. 237–248, doi: 10.1007/978-0-387-09457-1_20, http://www.springerlink.com/content/rm0p376x7g02683h
  9. 9.
    Schwiegelshohn U, Tchernykh A, Yahyapour R (2008) Online Scheduling in Grids, International Symposium in Parallel and Distributed Processing, 14–18 April 2008, pp. 1–10, IPDPS 2008, doi: 10.1109/IPDPS.2008.4536273, http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=4536273
  10. 10.
    Ullman JD (1975) NP-complete scheduling problems. J Comput Syst Sci 10(3):384–393zbMATHMathSciNetCrossRefGoogle Scholar

Copyright information

© Springer-Verlag 2009

Authors and Affiliations

  • Vincent Keller
    • 1
  • Hassan Rasheed
    • 1
  • Oliver Wäldrich
    • 1
  • Wolfgang Ziegler
    • 1
  • Ralf Gruber
    • 2
  • Marie-Christine Sawley
    • 3
  • Philipp Wieder
    • 4
  1. 1.FhG SCAISankt-AugustinGermany
  2. 2.EPFL – Laboratory of Computational EngineeringLausanneSwitzerland
  3. 3.IPP – ETH ZurichGeneva 23Switzerland
  4. 4.TU Dortmund – ITMCDortmundGermany

Personalised recommendations