A fully self-contained message is purely a complete representation of a specific event. And the message can be published and archived as one. The message can either immediately or later be interpreted as the event without relying on additional data stores that would need to be in time-sync with the event during message processing. An example of a self-contained event message is shown in the following diagram:

Figure 9

This is in contrast to only passing references to the data in the event, with the information itself being stored elsewhere. This is similar to the claim check pattern described for the integration solution, and that's why we use the term claim check event message here. The following diagram shows the same example with a claim check event message used:

Figure 10

Depending on the event type, the strategy to choose might be clear. A raw event, for example, will mostly use a self-contained event message. But when modeling a business...