Event management

IBM® Intelligent Operations Center focuses on the integration and optimization of information within and across multiple domains in a central operations hub, in real time and over long periods. Event data management enables IBM Intelligent Operations Center to assimilate data from multiple systems to constantly predict and react to significant events and trends.

Event messages are self-contained data items that contain basic but complete information to which recipients can respond. The IBM Intelligent Operations Center data receiver component can pull data items from CSV files, database tables or Esri servers, or by using the CAP. In addition, data items can be sent to IBM Intelligent Operations Center through a REST API.

Events come into IBM Intelligent Operations Center in different forms, which are based on the nature of the operations and domains in the central operations hub. Some examples of the forms of event are triggers, thresholds, complex events, and manually generated events.

Triggers are events that are generated by something that happens and usually require an action to be taken by the recipient. The following list contains some examples of triggers:
  • Fire or smoke alarms that are set off
  • Information technology systems that go down
  • Intrusion detectors that are tripped
  • Natural events that are picked up by sensors, such as earth tremors
IBM Intelligent Operations Center can receive information about such events from external systems and route it to the appropriate action. For example, the appropriate action might be to trigger a procedure, or to route the information to an integration point. In general, it is likely that lower-level indicators would be summarized and only passed to IBM Intelligent Operations Center if they merited wider attention. For example, all fires might not be reported as events. However, a fire that involves multiple divisions of the fire service and environmental protection expertise, because of hazardous material, would merit reporting to the operations center.
Threshold events help you determine when the measurements that are obtained from a sensor or other source have moved outside the normal range. Basic threshold events are comparisons that compare two or more measures and report a trend. More sophisticated threshold events can compare measures against a threshold that is created by historical information. The following events are examples of threshold events:
  • Over and under temperature alarms
  • High and low water levels
  • Air quality and water purity that breaches environmental standards
  • Excessive power consumption
IBM Intelligent Operations Center can manage such events in the form of key performance indicators (KPIs).

Complex events bring together information from multiple systems to determine whether a group of related events should be reported. For example, the toll road authority receives a trigger event from its monitoring system that indicates that the computer link for credit card authorization is down, followed shortly by a threshold event from the financial system warning that they are close to their credit limit for unauthorized payments. The combination of the two issues is much more serious than either in isolation, so a complex event is generated to raise awareness and coordinate a resolution.

Events that are entered manually are especially important to cities. Some of the events are observed incidents, such as crimes and traffic accidents. Other examples of events that are entered manually are those events that are generated from emergency calls from citizens, from reports that are made by city officials, or from management systems that report on city status. The following events are the most common types of event that are entered manually:
  • Severe weather warnings
  • Crime reports
  • Fires
  • Road traffic incidents – accidents, congestion, unusual loads
  • Upcoming events – rock concerts, road races, parades

Complex event processing allows a city to easily identify exceptions to city systems, occasionally to identify trends from unrelated data, and to predict future issues.