A new approach to upgrading legacy systems
A premier financial services company serving 200 million customers in 100 countries around the world daily
The project was easier said than done. The older system software used a proprietary method to communicate with the VAX computers; an intervening layer included all the business rules. A standard upgrade to the newest software version would require developers to reverse-engineer and recode all those rules—a daunting and expensive task for a global company.
The company had entered 21st-century banking with a 20th-century system. A proprietary data store operated on an early version of industry-leading software designed for VAX computers. As such, it lacked the capacity to manage the changing regulations, currency exchanges, and robust growth that were part of the bank’s operations.
The client partnered with EXL to import the existing functionality—data and business rules—directly into the software’s latest version. The situation called for more, however: a relational database that would work in the current business environment and scale up to handle the growing number of transactions.
- $4 million reductions in maintenance costs
- Decoupled data from system for maximum usage
- Provided framework for growth
- Modernized systems
The client needed help upgrading its legacy system and bring it into the 21st Century
- Upgrade legacy systems
- Reduce costs
From its in-house staff, EXL assembled a custom team of experts in both software versions as well as Java developers, web professionals, and database specialists. It soon came up with a better idea: a wrapper that translated requests out of the earlier software version and retrieved the data. To avoid the response time that such a layer could add, team members incorporated significant caching of repeated requests. The wrapper was developed as a Web-compatible service, enabling other systems to integrate with the legacy system as well.
As part of the solution, the team made substantial updates to 15 data-feed formats so it could interface seamlessly with the relational database. This was no easy task: some of the formats dated back to the 1970s, and no documentation was available.
All this development took place quickly. The wrapper concept was completed in a week, and the solution built and tested in a month.
Key benefits were a $4MM / annum maintenance costs eliminated and decoupling data from system for maximum usage and provided a framework for growth.