Dann Sheridan writes about the options:
Is it possible that Open Source packages exist which rival what has traditionally been the highly proprietary arena of data center operations? The “standard” operations platforms for data center traditionally includes something like Remedy, HP Openview, CiscoWorks, and a handful of other tools which need to be integrated in order to work together. There is OpenNMS which provides service polling, performance data collections, and event management and notifications. What are missing from this feature set are event correlation, escalation, process definition, and some reporting. Double Choco Latte, or DCL, provides escalation, process definition, and some reporting. The missing piece is a correlation engine. Qualys provides an Open Source correlation engine for IDS. I wonder if this could be customized to correlate general network events. What has been missing from all operations platforms is a predictive failure engine which analyses a complete history of events for a given device or set of devices and predicts the next failure. I know no Open Source engines like this. Finally, a robust reporting interface is needed which will need to be built from scratch once integration of the packages has been designed. I know from experience that the traditional approach costs between $750K and $1MM to get up and running and approximately $150K to maintain annual software licenses and another $80K to $100K maintenance labor and ongoing development. At the very least, Open Source would save you the $150K in annual software maintenance.