Event-Driven Architectures (EDA) are very much in vogue, driven by the Internet of Things, Big Data and Event-Sourced Microservices. However, most of the articles on-line refer to the technical implementation and, more narrowly focused, IT/OT system events or GUI interaction events such as a ‘Click Event”.
What is an Event?
“The term “event” generally has two meanings. On one hand, it denotes a perceptible, real-world, occurrence, and on the other, its representation of “state” in a computer system. The latter is also called an “event object.” The term itself is therefore ambiguous.”
The focus of this article is the “business event,” namely the status change within an organisation or the environment (e.g. Supply Chain or Customer Behaviour).
A Business Event Taxonomy – Why it’s needed.
Taxonomy refers to how information is grouped, classified and labelled within a shared information environment. ,The term is derived from the Greek taxis (“arrangement”) and nomos (“law”).
1. Events (Event Objects) need to satisfy a number of generic requirements which require consistent meaning within the Enterprise (or extended-enterprise such that they can be: classified, searched, filtered, aggregated etc..
2. A clear and consistent definition of business events is required to take advantage of digital opportunities. Events underpin digital design patterns.
3. The focus of this article is business events. However, system events (including signals from Operational Technology such as SCADA systems) are often translated to business events. Such events are often filtered or aggregated to create a business event. For example, a temperature sensor on an engine component may send a signal every few minutes, however, it’s only when the temperature tolerance is exceeded over a defined time window that a business event is published.
4. Both business events and system events can be grouped by the three ‘A’s:
- Advice – An informational event that doesn’t require immediate action.
- Alert – A warning event, usually a trigger for preemptive action.
- Alarm – Usually an instruction to take immediate action or to re-prioritise.
5. & 6. The three ‘A’s: Advice, Alert & Alarm, are a useful lens to use with business stakeholders to start a discussion about data-driven solutions.
7. A taxonomy of events is simply a family tree that groups events by type. At the top of the tree, we have events, then we group by business events and system events. Under that, we have event categories (the three ‘A’s). And below that, we have the event sub-types (which tell you exactly what the event was – for example a Facebook ‘like’ versus a ‘Proof of Delivery’.
8. & 9. The metadata of the event (message). If the business requirement includes the need to make provision for events yet to be defined (this usually the case), this needs to be defined in a meta-meta model. In other words, it’s a ‘template’ for event messages known or not. An event message is typically broken down to two sections:
- Event Metadata – the fundamental data the identifies the event.
- Payload Metadata – the variable information tightly associated with the event.
Events are the currency of ‘Situational Awareness‘ which for many organisations is the abstracted strategic goal yet often coached as a particular ‘Digital’ solution. How a business take advantage of this ‘awareness’ will vary from business to business. However, understanding the meaningful events, within their context, might be the most important aspect of their Digital journey.
#eda #datadriven #soa #bigdata #iot #ai #acd #ml #eip