Metrics for Grid Applicability: A Distributed Elliptic Curve Platform Assessment
The distribution of computational load among several nodes is an important step for problems requiring High Performance Computing (HPC). This phase is critical because bad decisions could cost time and money.
The emergence of heterogeneous networks with resources available for a wide group of users as provided by Grid Computing  creates the need for formal processes and metrics to evaluate if the application of the problem to the Grid is feasible.
In this article we introduce some auxiliary indicators for measuring the potential applicability of a parallel application to a Grid. Using the measures defined in the internal draft (GWD-I) produced by the Network Measurement WG  from the Global Grid Forum and RFC 2330 , the authors present some aggregate metrics that are useful in the characterization of the parallel applications that are well adapted to existing grids.
The defined auxiliary metrics were useful in the validation of a concrete application that factorizes numbers using the elliptic curves method (ECM)  over a testbed Grid. The results of the application of the metrics to this specific mathematical algorithm are presented.
Unable to display preview. Download preview PDF.
- 1.Bradner, S.: RFC 1242: Benchmarking terminology for network interconnection devices, Status: INFORMATIONAL (July 1991)Google Scholar
- 2.Ferrari, T., Giacomini, F.: Network monitoring for grid performance optimization. Computer Communications (2002) (submitted for publication)Google Scholar
- 3.Foster, I., Kesselman, C.: The Grid: Blueprint for a New Computing Infrastructure. Morgan Kaufmann, San Francisco (1999)Google Scholar
- 4.José Guimarães, P.T.: Spino: A distributed architecture for massive text storage. ICEIS 1, 244–248 (2001)Google Scholar
- 6.Lowekamp, B., Tierney, B., Cotrell, L., Hughes-Jones, R., Kielmann, T., Swany, M.: A hierarchy of network measurements for grid applications and services - draft. Technical report, Global Grid Forum - Network Measurements Working Group (July 2002)Google Scholar
- 7.Paxson, V., Almes, G., Mahdavi, J., Mathis, M.: RFC 2330: Framework for IP performance metrics, Status: INFORMATIONAL (May 1998)Google Scholar
- 8.Radajewski, J., Eadline, D.: Beowulf howto. Technical report, Linux Documentation Project (November 1998)Google Scholar