Re: Choosing a database for a DSS system (40-100G range)

    Eyal Kattan wrote in message <3642AD78.8694C3F_at_infomall.co.il>...
>Yes, I am based partially on IBM information, only because it always
 turn out to

>be very accurate and not exaggerated. It took 3 days for Oracle to load
 a 300G

> database while DB2 needed 16 hours only.

    The last thing I am willing to trust are these so-called "shootouts" and crap. I can load 300GB data with the right platform and tools within hours on Oracle. And I'm sure the same can be said of Informix and DB2. This marketing bs only confuses the issues. It's like watching two pre-schoolers argue about whose dad has the longest penis (actually saw it - but that's another story... :-).

    TPC is useless when it comes to the Real World (tm) where you do not have a "DBA expert" that wrote parts of the DB kernel to tune the database. Anyone who believe in this fantasy of TPC benchmarking is probably in the academic environment and never had to deal with the real business issues like lack of skills, looming deadlines, restricted budgets, and justification to the board for every single piece of ISware you want to buy, etc. etc.

>Now, you probably want to tell me again that IBM is wrong, well,
 unfortunately, IBM doesn't work

>hard on marketing, and therefor they are much more conservative with
 what they announce

>regarding their product's performance.

    Does this really matter? Do you honestly expect a corporate to buy product xyz simply on the merits of that product's TPC rating!!! For heavens sake man, look at Windows'95. The biggest abortion of an operating system ever, that satisfies almost no technical or performance requirements and yet it has 90+ % of the desktop operating system market!

    It will be great if software can be chosen and used on technical merits alone. Unfortunately, that is not how things work out there in the big bad business world...

    regards,
    Billy Received on Fri Nov 06 1998 - 00:00:00 CST