Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Data Warehousing

3 views
Skip to first unread message

Michael E Willett

unread,
Jan 10, 1995, 10:30:48 AM1/10/95
to
Re: Data Warehousing

Sue Hansen at Sun writes:

>Any "success stories"??

Hi Sue,

There are a number of data warehousing success stories in a lengthy data
warehousing story in the January 16 issue of Information Week. The
success stories include a couple of Red Brick success stories that
Red Brick emphasizes in its literature.

I think that Sybase has announced either a data warehousing strategy or
an actual product just recently, and they may have some success stories.

There is also a big data warehousing conference coming up in Orlando in a
couple of weeks that may be of interest.

Mike Willett, EE
Storage Computer Corp.
11 Riverside Street
Nashua, NH 03062
Tel. 603-880-3005

::::I/O-accelerated, very fast storage for SCSI systems to 1,000 GB::::

NRaden

unread,
Jan 11, 1995, 2:30:51 AM1/11/95
to
In Message-ID: <3es7bd$d...@cronkite.Central.Sun.COM>,
su...@riviera.central.sun.com asked the following provocative questions:

>Specifically, we are client/server specialists and
>must educate the company regarding data
>warehousing - evaluating its benefits vs. costs -
>in order to select the most appropriate IS strategy
>and eliminate current system degradation and
>expand DSS capabilities.

>Can anyone who has implemented data
>warehousing provide me any information?
>Any "success stories"??

Benefits vs. cost in data warehousing is a very murky area. Very often,
the justification for a DW will be to relieve the mainframe and shift the
load to a client/server environment, where TPA's and *DASD* are generally
less expensive. In this case, it is usually easy to produce a cost/benefit
analysis, but the numbers are always misleading. For example, if this move
will *save* me 1/2 of a CPU and 10 DASD spindles, can I put that money in
the bank? Probably not. Now, the data center folk will go to the mat with
me on that one, but they're nuts.

Often, a DW will be built to satisfy a single application, for example
precision marketing. It is easy to project the benefits of a successful
implementation, but exceedingly difficult to track and measure in
practice. Nevertheless, these are the most plentiful examples, and the
ones with the highest P of success.

Sometimes the sponsor of a DW is an IT organization just looking for a way
to
*get the users off our backs for data.* Needless to say, these are
generally failures.

What is particularly interesting is that the original concept of a data
warehouse is lost in the current shuffle. The Red Brick/Ralph Kimball/Star
Schema data warehouse is a read-mostly database used primarily in sales
and marketing applications, and by the retail and CPG companies
predominantly. This is an approach that is highly useful, and my
organization is happy to prosper providing this type of service. However,
back in the Stone Age, some of us thought that the whole point of a data
warehouse was to provide a suite of enabling technology and practices for
the lofty goal of greatly improved decision-making capabilities.

This suite would include: 1) tools to map disparate, dispersed and often,
erroneous legacy data into databases of clean, actionable data that was
certified and blessed by the organization, 2) enabling infrastructure for
all interested parties to partake of these comistibles through an
intutive, *friendly* and common interface, and 3) provide the framework
(aka, an application architecture) for people to add value to the
warehouse through the creation of applications in their subject area. Very
few enterprises have attempted this, and only a fraction of them are
producing results. Why?

Kay Hammer, the founder and current CEO of Evolutionary Technologies,
Inc., producers of Extract! (DW builder as in #1 above that we are very
fond of), hit the nail on the head when she said *...by their very
definition, a warehouse is a public application crossing operational
boundaries. Consequently, it can be a career-limiting move not to consider
carefully all these issues during the design phase." Turf war,
provincialism, inertia. These are the enemies of a robust data warehouse
initiative.

I would say that the smart move is to start small and get it right the
first time. Data standards and data quality administration are just too
difficult in most organizations. But then, only 50 or so of the current
Fortune 500 will be around 50 years from now. Is it any wonder?

Neil Raden
Decision Support/Data Warehouse Consulting
Envirometrics, Inc. 805.564.8672
133 E. De La Guerra St. 805.962.3895 (fax)
Santa Barbara, California 93101 E-Mail: NRa...@aol.com

Kwok Fung

unread,
Jan 11, 1995, 1:23:34 PM1/11/95
to
In <3es7bd$d...@cronkite.Central.Sun.COM>, su...@riviera.central.sun.com (Sue Hansen) writes:
>
>I am taking an Information Systems class in graduate school and
>our group has been tasked with providing a solution to a company's
>decision making process.
>
>Specifically, we are client/server specialists and must educate the
>company regarding data warehousing - evaluating its benefits vs. costs -
>in order to select the most appropriate IS strategy and eliminate current
>system degradation and expand DSS capabilities.

Hmm... Trying to find a strategy by proposing a solution before
you even understand what the strategy should be ? May
be after that, you can define the business problem, eh ?
I've got it, first come out with a solution, then the strategy,
and then define the problem. But why not implement something
first, before figure out what the solution is ?

Kwok L. Fung (kf...@gov.calgary.ab.ca)
(kf...@kfung.dpsd.gov.calgary.ab.ca)

0 new messages