Gissing Software Rearchitects Market Data Contribution System
With market data volumes increasing, Gissing Software (booth #1601) is unveiling a major upgrade of its multivendor market-data contribution system at the SIFMA Technology Management Conference. The company has revamped the architecture of Gissing ConteX 4.0 , resulting in higher performance rates and lower latency for contributing market data and executable prices, says Richard Gissing, chief technology officer of the London-based company.
Banks and brokers use a contribution system to distribute their quotes — including indicative data to third-party vendors such as Telekurs and Reuters, and executable quotes to Bloomberg. While banks contribute their data across all asset classes, the contribution system is heavily used in fixed income and equity derivatives. It is used less in equities because data for that market tends to go through an exchange trading system.
"Over the last five years that the existing system, ConteX 3.0, has been in production, we've seen a dramatic rise both in the number of instruments that people are contributing and also in the rate at which they update," relates Gissing. As a result, the vendor revamped all the inter-process communications in version 4.0 "so the message routing is more efficient as well as the data architecture to cope with higher volume of data and greater update rates," says the CTO.
According to Gissing, latency contributed by the system has been reduced by at least 300 percent to 400 percent. In addition, there's a significant reduction in CPU usage on the service to allow it to cope with more data, he adds.
Low latency "is going to be a lot more important under MiFID," observes Larry Tabb, CEO of TABB Group, referring to the European Union's Markets in Financial Instruments Directive, which begins to take effect in November. A low-latency market-data contribution system "helps brokers quote and distribute their data out to the marketplace," he adds.
"Pre- and post-trade transparency rules are a significant aspect of MiFID," Tabb continues. "This is tied into liquidity management — basically, ... how do you automate the provision of liquidity across products? All of this is to determine what price do you want to quote for a product," he explains.
"You're going to wind up having a wider array of products being contributed from a wider array of dealers and banks," says Tabb.
According to Gissing, the biggest increases in market data have been in equity derivatives. One client publishes data on 15,000 instruments and updates those at about 10,000 per second, he relates. By the end of 2007, the client expects to publish quotes on 50,000 instruments, with 30,000 to 40,000 updates per second across that instrument set, says Gissing.
Another important new feature included in Gissing ConteX 4.0 is dynamic configuration, which means users now can add in a new record set, add new fields or change the format of the data that is being published without restarting the system, according to Gissing. In the previous system, there were restrictions on making changes to the configuration, he notes.
"That wasn't a big deal five years ago because the number of changes to configuration were infrequent," Gissing says. But now users are creating derivatives, which is leading to dramatic surges in the amounts of data.
"Dynamic configuration is the major thing that our clients have been asking for to cope with increases in data volumes, and it's required the major architectural changes to support that," Gissing says. With version 4.0, firms can add new records, change fields or change formatting in real-time or schedule them to happen at a specific time intraday, he claims.
In addition, the product has a new derived data model that allows users to perform calculations (such as sum, mid, max and mean) based on the contents of the fields the system is receiving. "This was primarily driven by our clients who use spreadsheets to do these calculations," says Gissing. By removing the spreadsheets, they're doing the calculations on their servers, which is much faster, he adds.
Ivy is Editor-at-Large for Advanced Trading and Wall Street & Technology. Ivy is responsible for writing in-depth feature articles, daily blogs and news articles with a focus on automated trading in the capital markets. As an industry expert, Ivy has reported on a myriad ... View Full Bio
Tweet This