Data Stream Analysis (DSAL), a U.K.-based provider of low latency solutions for processing real-time streaming data, recently completed performance testing of its processing speeds, reaching rates of 200,000 calculations and 1.5 million updates per second.

DSAL and StreamBase, which provides its Complex Event Processing engine on which DSAL’s applications sit, doubled processing rates achieved as recently as the second quarter of 2006. The tests were conducted at Intel’s Low Latency Laboratory at Winnersh in Berkshire, England.

The algorithmic trading strategy performance was gauged based on a typical aggregation calculation over time windows of data. “We were basically running 20-minute VWAP calculations continuously on 20,000 instruments and were seeing rates of 200,000 calculations per second,” says Ian Hillier-Brook, CEO of DSAL, which was formed about nine months ago. He adds that the testing was done on out of the box Intel Hardware and StreamBase product without any modifications.

Although he can’t name the client, Hillier-Brook says DSAL is also currently working with a large investment bank to manage and measure their latency around market data. “We’re looking to build into the trading model a factor for latency,” he says. “So say a trade takes more than X amount of time to execute, then let’s not do it. Build in that type of concept.”