Treehouse Software Customer Case Study:
B-Source SA
The following is a recent discussion between Livio Varisco and Roberto Licandro at B-Source SA in Lugano, Switzerland and Treehouse Software Project Managers.
B-Source (formerly Boss Lab) is a Swiss company specializing in providing IT services and Bank Back - Office services to banks and insurance companies by means of an "outsourcing business model". B-Source provides these services to clients in Switzerland and abroad.
Value proposition, our services
Our target market
Certifications and credentials
Our Banking Operations Support System - an integrated banking system that supports most of a bank's front- and back-office activities, from loan and securities processing to marketing and internal audit in a multi-language, multi-currency, multi-regulation, and multi-legislation environment.
Our core system is integrated in a core database where all operations are executed and the data is stored. We developed interfaces to connect our system to third party solutions or external financial data providers (i.e., SWIFT, SIC, SIS, Reuters).
To better support data integration and the analytical platform, we designed and implemented an architecture based on a Operational Staging Area (OSA) and an Operational Data Store (ODS). These solutions are using an ORACLE 9i RDBMS.
The following diagram shows how our services portfolio is composed:
tRelational and DPS are the tools that allowed us to reach our goals in a pragmatic and non-invasive way.
The concept and architecture underlying tRelational and DPS are very smart because they do not work directly on the database nucleus, but use the standard database recovery support instead. This avoids burdening the nucleus, and thus not to affect the performance of the operating database, as shown in the diagram above. For example, a job running every two hours for a few minutes to propagate the updates would never burden the system in a significant way.
Currently we offer our services in the different mentioned models to fifteen customers.
The data integration from the core banking engine into other data management systems (e.g., RDBMS, flat files) was done by using scheduled programs that run during the batch. This solution, which eventually grew over time, became difficult to maintain in terms of the required human and infrastructure resources. Moreover, we are convinced that an RDBMS and the SQL language are better solutions to support DB inspections and extractions with common tools. Therefore, we implemented the following architecture as shown in the following diagram:
The two main challenges were:
Yes. Thanks to the professional support of TSI, in particular Mr. Hans-Peter Will, we have been able to perform the Proof of Concept and the installation into the production environment in a very short time.
But also afterwards we appreciated TSI's support for the education and for the solution definition. The product was tuned during the set-up phase with quick adjustment interventions. Since we went into production (September 2004) we have not to face problems due to the products, whose stability and performance allow extracting a 2-hour online PLOG, and upload it to ORACLE in a few minutes.