This page has updates for the following paper, including an up-to-date classification of grid monitoring systems, and feedback updates. You may consult the actual paper for a description of the grid monitoring taxonomy, as well as an introduction to grid monitoring concepts, GGF's Grid Monitoring Architecture, and a survey of relevant systems.
According to the defined taxonomy, some of the existing monitoring systems (or systems in which monitoring is important) are classified as follows.
You may want to consult the paper for an overview and an extensive characterisation of the systems.
top15 Feb 2005. GridICE was originally classified as a zero-level system but has evolved considerably since the paper was written. It now provides a HTTP-based API for accessing events as XML documents, and thus qualifies as a second level system with a centralised republisher. Also, note that GridICE is a new tool, not an evolution of InterGrid Monitor Map or EDT_Monitor, and Nagios is merely used for scheduling the monitoring observations.
15 Feb 2005. The published paper classifies Mercury as a second level system with distributed republishers, whereas it actually is a third level system, since it's republishers can be configured to form an arbitrary hierarchy.
If you are involved in the development of a grid monitoring system (or one that implements monitoring as a core function) that is not listed above, please let me know to add your system in the classification section. Also, feedback is most welcome if you think that your system is not accurately described or classified. Thanks to Zoltan Balaton (Mercury) and Sergio Andreozzi (GridICE) for providing feedback.
top