Create Data Streams with DirectConnect

DirectConnect is a premium feature that allows you to retrieve data directly from an external database, without having to store the data in Marketing Cloud Intelligence’s database. DirectConnect is used in cases where you have data sensitivity issues, or you want to display and interact with your data from an external source instead.

Whenever an instance relating to this source is initiated, data in DirectConnect is queried and retrieved live from the external database. For example, when loading a widget that uses at least one DirectConnect dimension or measurement, the data is visible and accessible for the duration of that instance only. Unlike TotalConnect or LiteConnect connections, no data is retrieved with DirectConnect, therefore no rows are added. Although data isn’t stored in Marketing Cloud Intelligence’s database, you can still use data from DirectConnect in calculations and in data classifications. These external databases are supported in DirectConnect:

• PostgreSQL
• Redshift
• Vertica
• Presto
• AWS
• Athena
• MySQL
• SQL
• Server Snowflake

• Create a Data Stream with DirectConnect
  Use DirectConnect to create a data stream and retrieve data directly from an external database. You can either configure a database, or connect to a database manually with a user and password. After connecting to the database and testing the connection, you can create a query to determine how data from the database is represented in Marketing Cloud Intelligence. Make sure the scope of the query is as close to the consumption requirements as possible for faster processing and performance.
• Connection and Performance Optimization in DirectConnect Data Streams
  Setting up a DirectConnect data stream establishes a real-time connection to your database. Each time an entity that uses dimensions or measurements sourced from DirectConnect is loaded, your database is queried in real time. A real-time connection can potentially result in many calls being made to your database concurrently, leading to slower performance and exceeding rate limits.