Open source database company Ingres has taken the wraps off its VectorWise technology and is looking for partners and customers to test its performance capabilities.

VectorWise, which was announced last summer, is a technology that has been designed to call on the power of Intel processors to improve database performance. Ingres has been claiming increases in performance of ten-fold or more and the preview programme is a chance to put these claims to the test.

The trial will work by allowing the participants in the programme to download VectorWise, with Ingres providing help on testing procedure and benchmarks.

According to Emma McGrattan, senior vice president of engineering at Ingres, VectorWise is going to present a great value proposition for companies with large databases who are looking to ramp up performance considerably. "I can see companies like telco or financial services companies benefiting from the technology," she said. "Telcos, for example, have a lot of analytics and business intelligence to perform." She admitted that VectorWise, as it stands, won't be for everyone: "Some people will wait for version 2," she said. But she added that the opportunities the technology presents is huge.
"We're offering an increase in performance that previously, you'd only have got with an appliance likeTeradata - the cost of which can be prohibitive."

VectorWise was developed in conjunction with the database research team at CWI of Amsterdam, a research institute in mathematics and computer science. Headed by Peter Boncz, the team has worked with Ingres in marrying chip and database technology.

One of VectorWise's early triallists is already clear of its worth. Rapid-I, a data mining specialist, has been testing Ingres VectorWise. Rapid-I has already seen dramatic results in its initial testing of Ingres VectorWise.

"We are overwhelmed by the performance gains we have seen from Ingres VectorWise," said Dr. Ingo Mierswa, CEO, Rapid-I. "The new implementation of sophisticated algorithms for data analysis was easily done due to the clean and standard SQL programming interface. And thanks to Ingres VectorWise the creation of prediction models was even faster than for in-memory calculation - without any restriction for data set size or without need of loading the data into the client at all."