Supercomputing and business intelligence collide

Superfast number-crunching can help businesses respond faster, says Jim Goodnight of SAS
Written by Nick Heath, Contributor

Superfast number-crunching can help businesses respond faster, says Jim Goodnight of SAS

A blend of supercomputing and business analytics is allowing businesses to identify new ways of maximising their returns, SAS CEO Jim Goodnight said yesterday.

High performance computing clusters are being used to analyse large amounts of corporate data, allowing businesses to identify opportunities and risks far more quickly than has been possible in the past, Goodnight said at the company's Premier Business Leadership Series conference in Las Vegas.

"There are so many future applications being isolated by this new computational methodology we are developing," he said, adding that it would be well suited to future tasks such as performing "progression analysis of millions of records with thousands of variables" and delivering an answer in seconds.

SAS' high performance computing service spreads the analytical work between multiple computer processors, and can offer the ability to use computer memory rather than a hard disk as storage - speeding up the rate at which data can be processed.

This close to real-time analysis could generate fresh savings for businesses, Goodnight said, by allowing business analytics software to be used to inform business decisions, where previously the software would have been limited by the time taken to analyse data.

"We are doing things like calculating the value of risk to a bank. We are working with a bank in Singapore right now and we have got their times for risk computations down from 18 hours to about 12 minutes," he said.

In the retail sector Goodnight said SAS is working on "mark down optimisation".

"Items that aren't moving are marked down [in price] by retailers but it has always been done a lot by guesswork or spreadsheet. We're taking all the data on how products are moving out the stores and recommending mark downs for specific items and stores, so that retailers can make sure that item isn't on sale at the end of the season and that they maximise mark down revenue."

Editorial standards