The Outbox pattern is the industry-standard for deduplicating messages. In its most basic form it is about storing the messages you are about to send before proceeding to dispatch them. This approach allows the messages to be generated atomically with the business state transition saving us from multitude of anomalies with creepy names such as Zombie Records and Ghost Messages.

This talk is a deep dive into the Outbox exploring less known aspects of the pattern. In the first part we are going to discuss what drives the dispatching of the Outbox messages with pros and cons of popular drivers such as client, retries and polling/timer.

In the second part the focus is going to be on the deduplication function that is often part of the Outbox implementation. We'll look at the challenges related to managing the ever-growing set of deduplication information and discuss alternatives that do not require keeping track of every single processed message.

After the session you will be able to assess the quality of Outbox pattern implementations you find in available messaging frameworks, as well as build one yourself to best match your environment (language, runtime, database, hosting etc.)