Source Analysis Guidelines
All Fabric projects include data sources which transfer data through interfaces. Therefore, when creating a Fabric project, the first step is to analyze the data sources needed for your project.
Do the following:
- Identify existing data sources and analyze the data they store.
- Define the interfaces for the data sources required for the specific implementation.
What Do I Need to Know Before Creating a New Interface?
- Identify which source systems store Digital Entities information in your project.
- Check the following for each identified legacy system:
- System name: For example, CRM.
- Description: For example, Customer Relationship Management.
- Main business entities. For example, Customer, Contract or Address.
- Nature of data in each source. For example, Customer Data or Financial Data.
- Data sources technology: For example, a database like Oracle or SQL Server, or a non-database like files or a message queue.
- Check the following for the database (DB) data source:
- DB type of the system. For example, Oracle, MySql or PostgreSQL.
- DB version of the system. For example, Oracle 11g.
- Check if the DB has a built-in connector in Fabric. If it does not, check if the DB has a JDBC driver.
- Check the connection details to the DB.
- Check the following for a non-database (DB) data source (depending on the data source’s technology):
- Server connection details.
- Protocol. For example, SFTP or HTTP.
- Format. For example, JSON.
- For file-based integration, check the location, size and frequency of the files.
- Evaluate the requirements for the project’s hardware sizing based on the expected data volumes.
- Identify the relationships in each system:
- Which source tables are related to the Digital Entities in your project?
- Check for ERD / data model documents for a better understanding of the relationships between the source tables.
- Relationship of the source system to other systems. For example, check if the Billing system holds CRM Customer ID.
- Map duplicated information across several legacy systems and check which system is the data master.