Archives of the TeradataForum

Message Posted: Wed, 23 Jan 2002 @ 11:49:27 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

"Which is better" has to be answered by "for whom" "for what" and "when" (at what point in the evolution of an intelligent organisation). It really is like comparing Chalk and Cheese.

First, a relational DB allows incremental growth of data including dimensions without the need to dissemble and rebuild data cubes - the latter experience is costly, painful and inhibits evolution and, consequently, intelligence. Teradata's (automated) spread of data across the nodes and facility for ad-hoc table joins is invaluable in enabling data that is loaded today to be later analysed by or compared to data that has not yet been defined and may not be added till months or years later. Unless you are an Oracle shop with lots of spare DBAs and programmers, do not build yourselves a bottleneck of scarce and expensive resource.

Second, ease of access to data on Teradata enables far more users to share the data and contribute their brain power to "business intelligence" - user licences and the cost of training and support for other systems quickly become prohibitive.

Thirdly, Teradata's optimiser and parallel processing is truly unique in enabling users to explore the data with ad-hoc queries. Later, the best queries may be refined and run in overnight production but true intelligence depends on facilitating and continually testing new queries.

Also remember Teradata's capability as an @ctive datawarehouse.

For me, the compelling issue is evolution: I have not met a company where users do not want to start analysis before data from ALL sources is loaded and/or where growth and reorganisation is unforeseeable. Regrettably, I have met IT staff who want to control their organisation and feel they should control rather than enable but even they should be convinced by the need for ad-hoc queries by large numbers of business analysts.

I just wish NCR's sales and marketing would major on their unique selling points instead of encouraging potential buyers to compare sample output to that from others.

Ken Hansen
Ivy House Associates
Independent Consultants

	<Prev	Next>		<<First	<Prev	Next>	Last>>