BACK OFFICE SECURITIES LENDING SYSTEM

Scott Logic worked with a leading US Investment Bank to develop a strategic new SOA-based system with hugely increased capacity and integration with front and back office systems.

SOA REPLACEMENT OF LEGACY SYSTEM

The client’s legacy securities lending system was reaching capacity, and it was forecast not to be able to meet demand within a year. Already the slow daily batch report generation had begun to impact the bank’s trading activities. The client took the decision to go for a green-field development based around a service-oriented architecture, and turned to Scott Logic as a development partner for this two-year project.

The new system was planned to run in parallel with the legacy system for six months, with outputs from both systems compared to ensure accuracy. The new system would be integrated with multiple management systems and front office applications used by traders.

TECHNOLOGY STACK

  • J2EE (EJB3, JMS)
  • JBoss application server
  • JAXB / XML
  • Hibernate
  • JMX
  • JGroups multicast
  • DB2 database

DEVELOPMENT OF CORE COMPONENTS

We provided half of the project development team: five developers and a project manager joining the client’s developers, architecture team, project manager and other team members based in London, New York and Mumbai. The team was well integrated, and our lead developer was chosen to head up the overall team a short way into the project.

We carried out design and implementation of various crucial components:

Run-time efficiency was a priority at all stages of the project, from the architecture to the code itself. Non-functional requirements were ambitious and strictly defined. Equally, the ability of the client to maintain this new strategic system was paramount, and the design was documented thoroughly using UML.

CAPACITY AND MAINTAINABILITY

The new system provided a solid base for future extension, and the modern high-performance architecture allowed the bank to support an increase in throughput from 800,000 to 10,000,000 transactions per day.

Full support was given to the legacy system during the project. The reconciliation system we produced proved invaluable during parallel running, allowing large end-of-day files from old and new systems to be compared and differences analysed well within the desired timescales.