This project is one of the largest federal OLTP systems supporting 100k transactions each day and processing over EUR 25 billion per year. With the increasing complexity and the rapidly changing requirements of highly specific functionalities, the number of unit and integration test cases has doubled over the last 3 years, which resulted in increased test runtime. In this contribution, we describe how the architecture of a large test environment, based on Oracle, has been transformed into a flexible and fast embedded in-memory architecture.
Target Audience: Architects, Developers, Testers, DevOps Manager
Prerequisites: Advanced knowledge in Java, Embedded software, Large-scale systems, HSQL
Level: Practicing
Extended Abstract
The subject of this case study is one of the largest federal OLTP software systems, which supports one hundred-thousand transactions each day and ensures the safe processing of over EUR 25 billion per year. With the increasing complexity and the rapidly changing requirements of highly specific functionalities, the number of unit and integration test cases has doubled over the last 3 years. This has resulted in increased test runtime. Since system performance is highly important, it is unadvisable to cut corners in tests. In this contribution, we will describe how the architecture of a large test environment, based on Oracle databases, has been transformed into a flexible and fast embedded in-memory architecture. The main challenges involved overcoming the differences between the SQL specifications supported by the base and target systems and increasing the test setup performance.