Jan van Til
This memorandum discusses the concept of Event Management. Event Management is meant to be a broader concept than the known concept of Alarm Management.
Despite the fact that the concept Alarm Management is widely known in the industry, a broader concept is needed: Event Management. Alarm Management is seen as a part of Event Management. The primary focus of Alarm Management is on the gathering of information (Signals) about Events that point at potentially unwanted and therefore alarming Situations. Event Management’s primary focus is on the gathering of information (Signals) of all relevant Events.
Event Management is about the management of Events in general. Events that take place in the real world and manifest themselves through Signals. Signals that can be reported to and validated by Event Management. After proper validation the Signals can be presented to an operator (or other professional) who interprets and processes them.
The terms-in-bold above are used as a structure and are discussed in detail below.
An Event always takes place in a specific Situation in the real world. In fact an Event turns an existing Situation into the next Situation. It’s the Event that connects consecutive Situations. Events happen concurrently, everywhere in space and all the time. In order to be able to link Events and Situations together, time and space turn out to be very important dimensions.
A Situation can be described as a more or less static background – the things in place – against which an Event occurs and can be observed. In that Situation the Event acts as a more or less dynamic foreground. The occurring Event creates – together with the existing Situation a new Situation against which the next Event can take place.
Events manifest themselves through one or more Signals. Signals can arise in the course of time and originate from various sources. Signals need to be distinguished from Events; Signals represent Events (the map is not the territory). For every Signal the time and place of its creation become more and more important (call for transparency). Therefore one needs to be able to register (first) creation-time and creation-place of every reported Signal.
Signals can be observed and subsequently reported by Actors. Two kinds of Actors need to be distinguished: human beings (motivated Actors) and artificial devices (preconditioned Actors that act on behalf of human Actors). Depending on the nature of an Actor, (relevant) Signals are observed and reported.
All reported Signals are registered as ‘raw’ Signals. The (ir)relevancy of reported Signals is controlled by business rules. It is important to register the registration-time and registration-place of each reported Signal (this applies to the used business rules as well).
With respect to Signals that are considered irrelevant, no further processing takes place.
Since relevant raw Signals constitute the foundation for further (re)processing, they need to be properly managed (transparency, traceability, accountability, availability etc).
Reported raw Signals mark a shift from the real world to the data processing world. From now on relevant raw Signals become data and all data manipulations are controlled by business rules that are appropriate to the situation (context) that produced those Signals.
All relevant raw Signals can be validated; invalid or missing Signals can be substituted with calculated ones. Based on this upgraded set of Signals other calculations/operations can be performed (aggregates, trends, gradients, grouping etc). Validations and other performed operations depend on formulated information needs and possibly generate new Signals. It is important to register the creation/registration-time/place of each generated Signal (this applies to the used business rules as well). Generated Signals need to be distinguishable from other (validated) Signals. Both generated Signals and other Signals are referred to as Signals hereafter.
All validated/generated Signals can, based on business rules, be evaluated and re-evaluated whenever Situation changes (manifested by new Signals) give rise to it. Based on evaluation results Signals are presented, not presented, presented again or no longer presented.
To be able to perform actions without human intervention, an artificial actor that acts based on business rules performs automatic actions.
For time travelling purposes it is important to keep track of the timestamp of all Signals as well as the business rules that are used to produce them. For transparency reasons and to be able to reproduce information this also means that validated Signals (and business rules) should not be overwritten because they could have been used in the meantime.
Prior to their presentation, validated Signals need to be categorised and prioritised. The way presented Signals are assigned a category and a priority is determined by business rules. Whenever Situation changes give rise to it, Signals are re-assigned a category and priority. Priority levels (e.g. high, medium, low) and category levels (e.g. do-not-display, notification, alarm), including the number of levels, need to be freely adaptable. Depending on business rules (e.g. the source of the Signals or the state of the system) the appropriate category/priority schedules are selected. Presented Signals need to be freely (re)routable depending on nature, origin, escalation level etc.
The exact way these Signals are presented (grouped, layout, colour, blink, audible etc.) and where and to whom they are presented, depends in considerable measure on their origin, business rules and Human Factors and Limitations (Capitals are used here to Stress the Importance).
Presented Signals mark a shift from the data processing world to the real world again.
Human Actors observing presented Signals will be operators (monitoring, controlling, planning etc.) and systems management. But there are other relevant (internal/external) parties as well. Depending on the specific information needs of various Actors, relevant raw Signals are gathered, processed and presented according to the business rules and schemes applicable to the target group.
A human (motivated) Actor now observes, manipulates (group, sort, override, suppress, shelve etc.) and interprets the presented Signals – thereby constructing a mental image of the original Event-in-Situation. Every presented Signal is always Signal-in-Situation. Whenever necessary any relevant contextual information about that situation needs to be easily callable in order to aid in reconstructing the situation adequately.
Subsequently human Actors individually assign meaning and perform corresponding Actions. These Actions are, of course, Events that will manifest themselves in time through Signals. Signals that may cause changes in the previously presented Signals.
Important aspects of the Event Management concept are:
1. In order to be able to facilitate any relation between any two Signals-in-Situation a strong connection with time/space dimensions for Signals must be supported. These dimensions form an important hinge point that enables one to relate any two Signals-in-Situation. Situations themselves are, of course, time/space related too.
2. In order to be able to facilitate time travelling, not only Signals-in-Situation need to be connected to time/space dimensions, but categorisation/prioritisation schemes, business rules etc. that operate on them as well (these ‘things’ also vary over time).
June 2008, 2008 © Jan van Til