Enterprise Application Integration for A European Investment Bank
The client began the implementation of a new back-office system and needed an interim solution for integration with financial systems. Information about trades arrived from the back-office system as business entities with attributes and transactions that were conducted. By contrast, information about finance systems was stored as General Ledger postings. Therefore, the main challenge was to build a flexible solution that would use certain rules to transform business entities into General Ledger postings. These rules had to be dynamic, changing to be aligned with business needs. We discovered and applied approximately one thousand rules and templates. Another challenge was to build a highly loaded (able to handle 1.5 million integration messages per day) and scalable (both horizontally and vertically) system.
Considering the complex and frequent changes of the transformation rules, we built a flexible solution with a conve nient user interface for non-technical staff. Additionally, we broke down solutions into modules; each of them could run on multiple instances on different servers, providing the required scalability. The solution consisted of three parts:
- Importer Scanner. Receives messages from the back office through the bus, processes and stores them in an internal database. In addition, it generates internal events for posting information into accounting systems.
- Trading Accounting Engine. Enriches the business entity using static data, for example recalculation of trade’s sums into accounting currency.
- Scanning and Transformation Engine. This application transforms and posts data into a target system.
- The scalable architecture supported 1.5 million integration messages per day
- Our flexible solution supported approximately one thousand transformation templates, which may be extended by non-technical users
- A significant reduction in the number of errors during integration, in comparison with the previous system
- Reducing the support team’s workload by up to 40%