Case Study

RE-PLATFORMING FROM A PROPRIETARY DATABASE TO AN OPEN-SOURCE DATABASE

BUSINESS NEED

The client ’s core operational system was on Oracle database. The client wanted to move over to PostgreSQL open-source database, to reduce total cost of ownership.

'Data Integrity' of the re-platformed database and Information security during the replatforming project execution were of paramount importance for the client.

PROPOSED SOLUTION

Milinia proposed a solution for both needs

  • Technical solution for re-platforming the operational system to PostgreSQL
  • A customized ‘Project execution approach’ to meet the information security needs of the client during project execution, transition and for the re-platformed database.

The core operational system was on 3-tier architecture - Presentation tier, Business logic tier and Data tier. The data tier consisted of ‘Data Access Layer’ and ‘Database’. Database includes Tables, Views, Procedures, Functions and other database level objects.

Milinia proposed to modify the ‘Data Access Layer’ and ‘Database’ to achieve the business need of the client with the Presentation tier and Business logic tier left intact.

PROJECT EXECUTION

Key steps in project execution:

  • Replace Oracle with PostgreSQL as the database
    • Create Tables and set integrity constraints in PostgreSQL
    • Convert all Oracle stored procedures and functions to their equivalent in PostgreSQL
  • Modify ‘Data Access Layer’ source code for connecting to PostgreSQL database and performing the required database operations
  • Migrate data from Oracle database to PostgreSQL.
Database replatforming and data migration from Oracle to PostgreSQL
Database replatforming and data migration from Oracle to PostgreSQL

The data migration was performed with PeriWin ETL. Data from 16 Oracle schemas were migrated to one PostgreSQL schema enabling the client Head Quarters to get a consolidated view of data. Primary keys and foreign keys were automatically transformed by ‘PeriWin ETL’. Data migration was performed with 2 trial rounds and a final migration over a weekend under the supervision of the client . Data migration scripts were created during the trial rounds and the final data migration was run as an automated process.

The size of operational system was 2572 function points. The entire project was completed in 4 calendar months.