Sitemap

Event systems in the business domain

2 min readNov 25, 2014

Event systems are receiving a lot of attention these days. The groundwork for event systems, event based programming and reactive programming has been laid in the past decades. The Reactive Manifesto, written by a enthusiastic group of computer science experts, describes the reasons why event-based or message-based systems are necessary to deal with the complexity of today’s and tomorrow’s environments. Large software companies like Amazon Web Services and Microsoft are adding event services to their cloud service portfolio. AWS complemented their Kinesis streaming data processing service with Lambda, a compute service capable of performing transformations and computations on events delivered. Other elements of the AWS event processing story include DynamoDB for storing large quantities of events, RedShift as a big data analytics store and SNS for notifying people and applications about events that occurred and that were processed. Microsoft offers a stack of event services on Azure including Azure Event Hub Service and Azure Stream Analytics Service. Event processing is closely related to big data analytics, with important open source initiatives like UCB Amplab BADS and Apache Spark offering cloud services for analyzing event streams.

Get Rine le Comte’s stories in your inbox

Join Medium for free to get updates from this writer.

But what does this mean for business applications? Traditional applications have been primarily systems of records, but new applications will become systems of engagement. In order to do so, they need to be reactive and event-driven. Let’s look at a number of examples. A sales rep uses his sales app to register interest in the sales of a particular product. When a customer purchases that product in the web shop, the sales rep immediately receives a notification that the product has been sold, and he can directly contact that customer. An operator monitors a bank of machines. The machines continuously emit sensor and measurement data to an event queue, which is analyzed in real time for anomalies. When a certain threshold is exceeded, or when the statistical analysis reveals that one of the machines is not performing that well, the operator is immediately alerted. And what’s more, the customer service representative at the manufacturer is also informed. The operator can directly contact the manufacturer’s representative and discuss the situation as both have access to the actual machine data.

The architecture of event-based business applications will be quite different. Instead of the traditional three tier applications with a user interface, logic and a database, the application consists of a number of event handling and event publishing components, connected through event services, message queues and push notifications. The challenge is to develop and maintain those applications in a reliable, performant and scalable way. The application of engineering methods such as model-driven-architecture (MDA), test-driven-development (TDD) and continuous integration (CI) is crucial.

Rine le Comte
Rine le Comte

Written by Rine le Comte

Personal stories of a software guy