Data Flow

Teminology

In the context of a Function Flow, exchange data through a specific Interface is implemented by a DataFlow

Typically, a DataFlow is implemented by one of the many data carriers, including:

• Files,
• Event streams
• API calls
• Message queues
• CORBA calls
• Etc.

A DataFlow belongs to a distinct Interface, hence Interfaces and related DataFlows share the same owner.

Usually, a DataFlow corresponds directly to one technical element like e.g. a particular event stream. This element can be modeled with a Data Flow Item. However, it may happen that a second level of description is needed for your data exchange. An example would be if it happens that different types of events are carried in a single event stream. For this cases, a one-to-many relationship has been provided to Data Flow Item.

Example

FunctionalFlow “Customer Synchronization” is implemented through two Interfaces :

• TRAD.006, an Event synchronization between Core Banking Sysetm and Trading Platform
• TRAD.007, a Batch file synchronization between Core Banking Sysetm and an Account Microservice

Interface TRAD.006 is implemented with a DataFlowItem (typically a topic in a streaming platform like Kafka or others) : /EVT/CUSTOMER

Mapping with Archimate specification

A Data Flow is modeled with a Archimate Data Object.

Archimate 3.1

A data object represents data structured for automated processing.

Data Flow could appear on Application Cooperation Viewpoint :