Modernisation of an inventory and cost valuation system

Context

A large, complex enterprise system supporting financial and operational calculations for stock valuation, inventory movements and cost accounting. It needed to be modernised towards more maintainable cloud-native architectures.

Challenges

1. 01Complex business logic spread across multiple modules and applications.
2. 02Dependencies between legacy services.
3. 03Large volumes of financial and inventory data in batch workloads.
4. 04Maintaining data consistency during the migration to Snowflake.
5. 05Country-specific accounting rules (hyperinflation, taxes, countries that distribute to others) that required extending the infrastructure.

Solution & architecture

Involvement in structuring the data re-architecture in Snowflake and in bringing new cost accounting features into production for the entire multinational: handling hyperinflation for Turkey and Argentina, adapting taxes for all countries, and valuation in the hub countries (such as Spain and China), which act as distributors for the rest of the countries and required many more factors to be taken into account, which meant adding components to the infrastructure. In addition, development and maintenance of backend services in Java (17–23) with Spring Boot and REST APIs; dependency analysis and refactoring of legacy components; migration of processing workloads; batch data pipelines; and adaptation of Oracle and DB2 SQL logic to be Snowflake-compatible, optimising SQL and data access. All of this with Kafka, containers (Docker, Kubernetes), CI/CD, automated tests and monitoring.

Key decisions

Not documented yet

Trade-offs

Not documented yet

Results

Lessons learned

• ✦Understanding the whole enterprise architecture, and not just isolated data tasks, is what makes it possible to modernise without breaking the business.

Ask about this case study