U.K. asset management firm Insight Investment is about to go live with a reference data management infrastructure from Citadel Associates. The implementation will make Citadel’s Cadis the data management infrastructure underpinning Insight’s growing derivatives operations.
Insight – formed through a merger of Halifax and Bank of Scotland, and the subsequent acquisition of Rothschild Asset Management – completed a successful pilot of the Citadel system late last year. Development has been ongoing since January of this year and it is expected to go live at the beginning of October.
Cadis is being used at Insight to analyse, cleanse and enrich incoming and derived data, then structure and ensure consistency before feeding it into a range of applications. The full implementation will work across 16 systems and have a total of 100 interfaces. In particular it will interface to Algorithmics, Bloomberg, DST’s HiPort, LatentZero’s Sentinel, Minerva and eventually Tesseract, Rimes and Xenomorph. This implementation is based on the Microsoft SQL and .Net framework.
Troy Travlos, IT strategist at Insight, said, “We looked at other products in the data management space including data warehouses and EAI tools, but we selected Cadis for its flexibility and fit. We liked the component approach and its … GUI.”
One of the issues that had to be solved was how to manage the different way each application system identified instruments. Says Travlos, “We needed to ensure we could identify discrepancies in each system in terms and conditions and derived data.”
Cadis does this, says Stuart Plane, director of sales and marketing at Citadel, using data matching capabilities based on the securities identifiers with exceptions management tools. There is a construction GUI that enables the business user to define the rules for data handling, their preferences for data sources, exceptions tolerance levels and management, essentially providing multiple golden copies. “This sophisticated GUI empowers the business analyst and reduces their reliance on programming and development support.” The system also maintains a full audit trail for compliance and risk assessment.
Plane says Citadel differentiates its service from other data management providers targeting buy side firms through its method of separating data storage from processing, and its speed in processing.
The company has secured four paying buy-side clients for Cadis and is involved in other pilots. Plane cites another client example, although he cannot name the client, who is using Cadis to integrate data from sources such as Factset, Russell Mellon and MSCI as well as derived data from their own internal models to feed into the Charles River Development order management system. This is based on an Oracle and Unix framework.
Citadel– not to be confused with the Chicago-based hedge fund of the same name – was founded five years ago in Luton, U.K. as a bespoke financial technology consultancy with primarily asset management firms as clients. It decided to productize its offerings two years ago and has since developed its component-based Cadis system.
The firm has just launched two new components bringing the total to seven within the Cadis suite. The component-based approach, says Plane, allows clients to buy only what they need and “there’s no need to replace a client’s existing infrastructure.”
The new components – DataBlocks and DataGenerator – can be used standalone or as part of an integrated application. DataBlocks provides a set of schemas for storing reference data, transaction history and performance data. The DataGenerator is designed to create and manage reference data, adding new data items or to supplement internal data with external data, through a single point of entry, and communicating these changes to third-party systems as well as other Cadis components.
The other Cadis components are: DataConstructor, DataMatcher, DataPorter, DataInspector and DataEditor. The U.K firm has a couple of consulting partnerships in the U.S. and it plans to increase its presence in the region next year.