Monitoring alerts, warnings and alarms is a difficult problem. Alarms
are not always seen, and when a serious problem occurs the number of alarms
is so great that the alarms themselves are of little help and are often
ignored. Our project deals with network alarms (e.g. a node is down) and
alarms generated by sensors installed in the business applications (inventory
too low or too high, excessive number of orders etc.). Our Enterprise Monitoring
prototype provides tightly coupled filtered views of the network current
status and past history, timelines of alarm overviews and traditional textual
details. Filtering attributes are modified interactively to deal with temporary
needs or alarm overflows.
Catherine Plaisant, Assistant Research Scientist
Wei Zhao, Graduate Student Computer Science
Rina Levy, Graduate Student College of Business
Downloadable MPG (83 MB)