DM: "CHESS": Harnessing Data Warehouse Power for Data Mining

[Joe Evans]

FOR IMMEDIATE RELEASE

Contact:  Joe Evans
 +44 (0)1256 355 899
 joee@isl.co.uk

8th April 1998

        "CHESS": Harnessing Data Warehouse Power for Data Mining

Integral Solutions Ltd has announced its programme for a fully 
scalable,

interactive data mining solution, based on its award-winning 
Clementine
Data Mining System.

CHESS - the Clementine High Efficiency Scalable Server - will allow
Clementine's uniquely interactive and visual style of data mining to 
be
applied to very much larger data sets, at greater speed.

Under the CHESS programme, ISL is inviting vendors of databases,
warehouses and parallel computer platforms to help shape the future of
business-led data mining. Initial partnerships in this area will be
announced shortly.

The CHESS server is a middleware component which is positioned as 
close
as possible to the host database / data warehouse. CHESS will optimise
Clementine visual programs for efficient execution on the database,
taking full advantage of software and hardware support already in 
place
for large scale data warehouses. Initial experiments on serial
processors show typical performance gains of at least 10 to 100 times.
On parallel database architectures, the gains can be much greater.

"Success in data mining is primarily determined by the quality of the
business input. Having established leadership in supporting 
business-led

data mining, we are now taking the lead in scalability and 
performance,"

said Alan Montgomery, head of ISL. "With CHESS, users will be able to
unleash the full power of their investment in database and data
warehouse technology for high value enterprise data mining. From an 
end
user's point of view, the Clementine user interface will remain
unchanged - but with vastly more power behind it."

Partnership with ISL in CHESS is an attractive proposition for 
database
technology suppliers, according to Colin Shearer, ISL's Technical
Director. "Together, we will develop versions of CHESS optimised for
particular database platforms, using the specific performance features
of each DBMS," says Shearer." This will enable our partners to 
position
their database and data warehouse solutions as uniquely powerful data
mining platforms."

"This is a radically different approach to using parallel technologies
simply to make algorithms go faster - we're making the complete data
mining process scalable," adds Montgomery. He cites the experience of
Dan Reiman who, when evaluating Clementine for Fannie Mae, said "I got
just as accurate models, in just as short a time using Clementine, as 
I
did with much more expensive parallel tools."

Reiman stressed: "Other data mining vendors concentrate on optimising
computer time; Clementine optimises MY time." CHESS will preserve
Clementine's interactive style, which has proven so successful for 
data
mining by non-technologist business professionals, and provided a
ten-fold productivity boost for professional analysts. Now they will 
be
able, where appropriate, to mine many times the volume of data.

CHESS is scheduled for trials within the Clementine user community 
from
third quarter 1998, with the full product release expected early in
1999.

ENDS

http://www.isl.co.uk