Derivatives Exchange Automates Securities Data Care
Eurex, the Swiss derivatives exchange established in 1998 as a joint venture of the Deutsche Börse AG and SIX Swiss Exchange, is today one of the most important derivatives markets in the world.
Eurex also operates electronic markets in the field of collateralized financing, involving more than 150 financial institutions, with around 500 users. When Eurex SecLend was launched, the number of securities traded jumped fourfold inone leap. SecLend has expanded the Eurex market since 2005, serving as an electronicmarketplace for global securities lending of fixedincome securities and stocks.
Initial Phase with 20,000 Securities
While master data concerning 6,000 securities had to be updated daily for the repo markets, SecLend began with 20,000, with a plan to increase up to 60,000. This involves extremely complex master data that we draw from external sources, consolidate and put through extensive quality assurance processes. Data flows in regularly in Excel format via email from suppliers such as Bloomberg, SNB Swiss National Bank, Telekurs and others. Previously, a four-person team was engaged 24/7 with data care.
Data Care Manager Emerges
It was already clear by 2004 that this costly and involved manual data updating could become an impediment to the planned growth of electronic trading. Our team put the upper limit for manual processes at master data for about 10,000 securities. So at the end of 2004 the static data pre-processing project (SDPP) was formed with the task of automating master data maintenance. The solution finally developed from a conversation with Jan Trnka, whose company Stabilit had been providing consulting service to SIX Swiss Exchange for a number of years.With associates of the Eurex Administration & Operation department, Trnka developed a profile of requirements that would serve as the foundation for an appropriate software solution. It quickly became clear to us that a rules engine alone would not suffice. Here we were dealing with processing complex information, starting with raw data from one source that was enriched, compared and then validated against data from other sources. Our concept therefore revolved around a rules engine combined with an object-oriented database designed to process complex data quickly. When Stabilit finally received the definitive award in November 2005, the technology portfolio had already been set—data care manager, as the project was christened, would be written in Java code on the basis of the Caché object database by Intersystems and integrate the Visual Rules rules engine and the Kiwi web extractor. In Trnka's view, a relational database would be hopelessly overstretched. "In designing the software, we took care not to hard code any individual business rule," Trnka commented.
All Lights Are Green
In November 2006 the software went into testing as planned. In a first hard test, 100,000 securities with a number of attributes were loaded and validated in just six hours, aggregated and available at the XML interface for the trading systems. In February 2007 the solution was then finally productive. Since then we have not had to modify the data model a single time. What is more, the extensive automation has freed up capacity in the team and now 100,000 securities pose no challenge at all for Eurex SecLend. As the next stage in our project, we are planning already to include an opening and closing price for the securities trading so that customers can be offered a valuation price.
