Table of Contents Author Guidelines Submit a Manuscript
Scientific Programming
Volume 12, Issue 4, Pages 239-251
http://dx.doi.org/10.1155/2004/896517

Grid Environment for On-line Application Monitoring and Performance Analysis

Bartosz Baliś,1 Marian Bubak,1,2 Włodzimierz Funika,1 Roland Wismüller,3 Marcin Radecki,2 Tomasz Szepieniec,2 Tomasz Arodź,2 and Marcin Kurdziel2

1Institute of Computer Science, AGH, Kraków, Poland
2Academic Computer Centre CYFRONET AGH, Kraków, Poland
3Universität Siegen, Siegen, Germany

Received 24 January 2005; Accepted 24 January 2005

Copyright © 2004 Hindawi Publishing Corporation. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

This paper presents an application monitoring infrastructure developed within the CrossGrid project. The software is aimed at enabling performance measurements for the application developer and in this way facilitating the development of applications in the Grid environment. The application monitoring infrastructure is composed of a distributed monitoring system, the OCM-G, and a performance analysis tool called G-PM. The OCM-G is an on-line, grid-enabled, monitoring system, while G-PM is an advanced graphical tool which allows to evaluate and present the results of performance monitoring, to support optimization of the application execution. G-PM supports build-in standard metrics and user-defined metrics expressed in the Performance Measurement Specification Language (PMSL). Communication between the G-PM and the OCM-G is performed according to a well-defined protocol, OMIS (On-line Monitoring Interface Specification). In this paper, the architecture and features of the OCM-G and G-PM are described as well as an example of use of the monitoring infrastructure to visualize the status and communication in the application, to evaluate the performance, including discovering the reason of the performance flaw.