Archives of the TeradataForum
Message Posted: Thu, 24 Jan 2002 @ 08:27:08 GMT
There is only one answer and it's not Teradata. Before you get excited, it's not Oracle either. Most/all of the contributors to a Teradata forum such as this will tell you Teradata is better. Guess what the contributors to an Oracle forum would say?
The only sensible answer is "it depends". None of us can advise which database is best suited to the needs of the company in question. It depends on many, many factors which we can't answer without the relevant information.
Several years ago I worked for a major UK bank. The manager in charge of data warehousing was constantly on the end of sale pitches by suppliers. He rated NCR's product the highest, but said the sales staff made a bad job of selling the product to him. The opposite was true of Oracle - he rated the sales staff but not the product. The lesson? Ignore the sales sheen, they'll tell you what they think you want to hear.
At another client I worked on the migration of a very large Oracle database off a Sequent platform onto a Teradata box. The Oracle DB originally cost tens of millions of pounds and was never satisfactory. Joins could not be performed, the large tables had to be physically partitioned and the user:DBA ratio was almost 1:1. I learned a lot more about Oracle during this project than I expected, largely due to the severe difficulty in getting the unloads to work. Almost all of my experiences were negative, I'm afraid. Once on Teradata the data was used to extend the existing data model. It worked a treat and the users loved it.
The old Oracle/Sequent platform was de-commissioned after 7-8 very unsuccessful years service. As part of the sales pitch the business were promised the earth by Oracle and failed to deliver. IMHO, NCR undersell their product but it still delivers. I have heard of recent cases where NCR technical staff ramble on about the bynet at sales pitches. This is the wrong approach, surely?
Part of the resistance to the use of Teradata comes from the massed ranks of Oracle/Unix administration and technical support staff who look after the data capture systems. These folks love the complexity, and their mastery of, both unix and Oracle and can't understand the need to buy into a.n.other dbms when they already have a perfectly good OLTP dbms setup.
Ask for client references, read the analysts reports and even check out the benchmark tests. Don't base decisions on the sales pitches alone.
In summary, my experience is that NCR undersell a great product and Oracle oversell a great OLTP but poor DW product.
Hope this helps,
|Copyright 2016 - All Rights Reserved|
|Last Modified: 28 Jun 2020|