Always do the math

By Marty Nuckles
I don’t know who originally said it, but I believe it is still good advice. Do the math and see where it leads. In the case of the AMD Opteron servers, the key data elements I would like to have are:
1. What was the total number of Opteron Server sales versus the total number of x86 servers for the last 3 years, year by year?
2. What are the projected numbers for the next year, two years and three years?
3. What percentage of these servers, both Opteron and Intel, were sold with Oracle software?
4. What are the projected numbers for the next three years?
I no longer have inside contacts at SUN or oracle, so I can’t get the actual numbers, but I’m guessing that the Opteron numbers are fairly small. In all likelihood, the Opteron numbers are too small to justify the costs of development, maintenance and support. So, from the business side, it makes sense to simplify the product options and stick with things that you are sure you can sell, despite the arguments that can be made over technological superiority.
However, given the fact that I’m discussing Oracle here, I’m can’t assume that anything is being done on a rational, dollars and cents approach to business. I’ve always got to give in to the mild paranoia that I developed while working there. The Opteron decisions could be founded in another set of political moves inside Oracle as it digests SUN. Check Paul Ruebens’ article on Oracle and Linux at:

http://www.serverwatch.com/eur/article.php/3873696/Oracles-Linux-Server-Slant

for some interesting insights. Additionally, pay attention to announcements on changes in licensing, support and consulting. Savio Rodrigues’ article at:

http://www.infoworld.com/d/open-source/license-change-leaves-sun-solaris-users-crossroads-858?page=0,0>

is another good read. So, from my paranoid side, I suspect this is not good news in the long run for current Oracle customers or future customers. I’m guessing that everyone should be preparing for fewer choices at higher costs over the long haul.