Oracle Overkill

I just read Justin Kestelyn's article "First Priorities" (August 18, 2000). He has hit the nail on the head!

I have been a database administrator/designer for the last 10 years (mostly with Oracle) and have seen shop after shop abandoning or not even considering good RDBMS solutions because of Oracle's aggressive sales staff (AKA "marketecture").

There is even one venture capital firm here in Denver that recommends Oracle to all its startups without regard to requirements!

When I worked at a cable television company here in Denver we had around a $500K budget for our warehouse database software alone. We contacted Red Brick, Sybase, and Oracle and told them we wanted to conduct a benchmark with their databases. Red Brick and Sybase sent two associates each to perform the benchmark, which lasted a week per vendor. Oracle had up to nine people on site at the same time, including a couple of sales guys running around trying to sell Designer 2000, Forms, and Express. They even went straight to my company's CTO and flew her to Oracle's corporate headquarters for a wine-and-dine session. Needless to say, after that, top brass put a lot of pressure on my manager and his director to select Oracle. But we were determined to select the best performer, and Red Brick was a clear winner for us. Our manager had to go through numerous high-level presentations and really put his neck on the line to convince the company to go with a small RDBMS vendor that no one had heard of. The project turned out to be a huge success mainly because of Red Brick's ability to check referential integrity and still load quickly. Overall, Red Brick excelled in all the areas Oracle did not.

I just had the opportunity to select an OLTP RDBMS for a new company here in Boulder and could have gone with Oracle, but I selected Sybase instead. Maybe it's not as feature packed as Oracle 8i, but it's a decent performer and the people at the company are much better to work with. I hope to see more articles of this nature from you about Oracle.

Joe Buhl
Boulder, Colo.

Embracing Complexity

I really enjoyed Michael Gonzales's "Alleviating Spatial Constraints" (August 18, 2000) and his point about the critical nature of process models. When I look around the industry, I see some technologies that have process models and some that don't. Even where modeling is possible, the results are proprietary, not shared, so we end up with a very fragmented view of the business processes and the data they depend upon. Sound familiar?

Is it likely that the exchange and integration of process models will become increasingly important as we try to automate more of the total data flow? How do you move toward automated impact analysis and change management of data from transactional sources to the end user if you understand only parts of the complete process?

The more I talk to database administrators, the more I hear about impact analysis and change management. They seem to spend a large part of their lives reacting to the downstream effects of change, and they have no tools to help. Sounds like an interesting area for some research.

Gary Robinson
San Jose, Calif.

Michael Gonzales responds: The dependency complex systems have on integrated views of processes and data is becoming critical for many organizations. The challenge, of course, is how we share and view this integration given that it reflects potentially disparate sources, both technically and architecturally. Although not a total solution, the unified modeling language (UML) serves as an excellent canvas from which to model, examine, and present these integrated processes and data. I can only hope that as leading firms begin to emphasize process modeling, that they do so with an eye toward standards such as UML.

Correction

In the article "Finding Common Ground" (IntelligentEAI, August 1, 2000), we erroneously reported that Forrester Research estimates that $8.5 billion was spent on hand-coding interfaces in 1998. The correct amount is $82.5 billion.