The challenge

A tier one global banking operation embarked on a major project that would create a single central data repository to hold all of its trade and trade-related data consistently and without duplications.

The core project needed to quickly create a large number of mappings that would transform the very wide range of trade data into the repository’s chosen canonical format. The mappings also had to be easy to maintain in the longer term.

Initially the analysis team used spreadsheets to define the required mappings for developers, but this approach proved far too cumbersome.

So the bank faced a challenge …

The project

As with many financial organisations, the global markets operation of the bank consisted of numerous teams, each having their own data requirements.  This resulted in multiple data stores, with a lot of duplication.  The bank therefore decided to create a central data repository or Data Cache – a single location from which all teams could access trade data.  All trade publishing teams would feed data directly and exclusively into this Data Cache, and all interested ‘consumers’ (both human teams and downstream systems) would obtain their data from it.

Aside from creating and maintaining the data store, this vision required thorough understanding of the data published by each source system, so that all trades could be stored in a consistent way.  Although all trades were already published using a common in-house XML format, the actual usage of the schema differed considerably between systems.  To overcome this, the Data Cache was based on a centrally-defined Data Model controlled by a single Data Architecture team.

Now each source feed needed to be transformed into this model.  To facilitate this, a team of analysts was brought in, arduously creating lengthy spread sheets detailing the required transformations for each feed.  These would be passed to the developers to implement.

This process was found to be slow and error prone, as the developer implementing the logic did not have a thorough understanding of the business meaning behind it.  A solution was needed that speeded up the build process, and gave analysts access to the implemented logic.

The solution

In a strategic move, Transformer was introduced into the project, completely replacing the spreadsheet-and-coding approach. Trace Financial consultants quickly trained a team of 10 analysts to use the Transformer DesignTime workstation.

Using Transformer meant that the analysts could specify and build clear, easily understood mappings  in a single step process with no coding stage.  Transformer’s test-as-you-build facilities also meant that the analysts could test each mapping as it was created. Completed work could thus be delivered into formal QA with a high degree of assurance that it would pass first time.

As a result the project moved forward far faster, soon delivering over 70 feeds from 18 different source systems, covering both trade and reference data.

Subsequently the project’s emphasis changed from on-boarding to maintenance.  With a constantly changing team, the ability for new analysts to use Transformer DesignTime to quickly familiarise themselves with the mapping logic proved invaluable.

Overall, using the Transformer approach to building and maintaining complex mappings was recognised to have been a key factor in the success of this strategic project.

About Trace Financial

Trace Financial Limited are leading specialists in financial message transformation and have been designing, implementing and supporting mission critical solutions for the financial industry for over 30 years. Trace Financial is part of the privately-owned Trace Group.

Find out more about Transformer here or download the Transformer Brochure below.