SIFMA Exhibitors Announce Databases That Enable High-Speed Analytics

For analysts who need to perform quick yet sophisticated queries on large databases, two SIFMA Technology Management Conference exhibitors are announcing speedy underlying databases.

Relative newcomer QD Technology (booth #3010), based in Rutherford, N.J., is rolling out version 4.0 of its Quick Response Database. QD's technology compresses any ODBC-compliant database into a data mart small enough to be put on a laptop or desktop computer for conducting efficient queries without having to hit a server, the vendor claims.

According to Pete Krainik, QD's chief marketing and sales officer, financial services companies are using the Quick Response Database to analyze risk or fraud potential. For instance, a risk manager might view a dashboard report, detect potential fraud or other risk, and have a window of two hours in which to analyze the report data. "One of the business problems we address is this issue of high-end, heavy-lifting analytics users whose sophisticated ad-hoc queries tie the server up, which slows response time to everyone else, and they get kicked off after 10 hours," he says.

"We solve that problem by giving the individual a compressed version of a database on their desktop that allows them, without being encumbered by the server, to do ad-hoc queries against large amounts of data," Krainik continues. "Users can access this compressed database using SQL query tools and open only the specific rows and columns requested in the query." The product is not designed for analysis of real-time data, however, because the database only allows incremental updates every few hours or once a day, Krainik notes, adding that a 2-gigabyte update would take about 20 minutes.

Putting the compressed database on laptops lets salespeople and others conduct queries remotely without suffering delays. According to Krainik, the data is encrypted for security and offers a "lock and wipe" feature — after a certain amount of time, the database can be locked. The proprietary compression also acts as a form of security; without the right query tools, an outsider can't access the data.

New features in version 4.0 of Quick Response include faster compression and query speeds, adaptors for integrating the database with analytics tools from SAS, Cognos and Business Objects. Pricing for the Response Database starts at $300 per user.

Palo Alto, Calif.-based Kx Systems (booth #1302) also is announcing a new database, the kdb+ v2.4 low-latency database. According to Kx, the database is used in large financial firms to handle as many as 1 billion time-ordered events a day.

The new version is said to double application speeds for filtering and querying data streams, such as market data feeds. Other speed-enhancing features support real-time corrections on order book data as it is received and dynamic indices for nested data — enabling multiple streaming sorts for fast, sophisticated analytics. Users can partition one day's data across multiple local disk drives, preserving rapid parallel disk access for single-day queries, the vendor says.

"Kx customers today capture and execute trades on a billion streaming events each day," states Arthur Whitney, Kx CEO. "We enable them to filter and act on data in real time so that they derive immediate intelligence from the messages streaming in. But we don't stop at streaming data, because firms also need a seamless transition between streaming and historical data. The uniform architecture of kdb+ provides a single end-to-end data environment. Write queries once and use them on any type of data."

Kdb+ v2.4 runs on 64-bit Linux, Solaris, Windows and Mac OS X, and supports multicore chips.