Parallel execution of sequential code

Jan Gönnecke

Parallel execution of sequential code

Based on a sequential PL/SQL code, a runtime environment was created that was not allowed to change the logic of the existing code, but had to allow a secure parallelization of the same. The deployment process was initially switched from a centrally used database to a methodology that allowed development based on tickets with a complete code base in Git repositories. Changes to the created branches led to the automatic execution of a CI/CD pipeline, for which happy tests were created in order to directly detect behavioral changes caused by changes to the code.

The project used PL/SQL in Oracle databases as the main programming language. Scripts were created in shell scripts and logic was taken from Perl scripts. Data was stored in json files, which helped with comparability over time in the repository. The first parallel parts of the code were tested early on in the project to keep the end customers encuraged to further support the project implementation.