 |
Recent
Articles |
Make A Linux Home Data Server Of An Old PC
So, you've got an old computer, and you don't know what to do with it. Sure, it can't play new video games, maybe can't run the latest and greatest software...
Linux Flash Player Saga Continues
Adobe's Mike Melanson has posted some more information about the lengthy delays in releasing a new version of Flash Player for Linux. "What could possibly be so difficult about..."
Linux Options Omitted From Vista Survey
It's either the Microsoft way or the highway for computer users according to a survey about the economic impact the Vista operating system could have in Europe. The potential for Linux making an impact...
Eclipse Touting Application Development
The Eclipse Rich Client Platform (RCP) has seen significant growth as a choice
for desktop application builders. Despite all the chatter about web services and browser-based applications, there is still...
Why Was Linux Successful (and Will It Stay So)?
Internet News reported on a panel discussion of the reasons why Linux was successful.
I had been thinking about that myself recently, and had earlier found an...
Python
Steps Up In Django And TurboGears
IBM presented a two-part series on developing applications with Python, using
Django as the framework in one example and TurboGears in another. I've been an IBM fan for quite some time...
|
|
|
10.25.06
Oracle Linux
 By A.P. Lawrence
At "Larry Ellison's Linux nears release", Ashlee Vance is unsure why Oracle would want its own Linux distro.
Actually, it makes a lot of sense. It's not the direct revenue of selling the OS with the database: that's almost completely unimportant. What is important is having control of the OS - tuning and tweaking it to be a home for their database. That's where the advantage for Oracle lives. By tightly controlling the OS environment, they can gain performance, increase reliability, and decrease their support costs.
Support costs go down because Oracle Linux code would always think about what Oracle database needs. It would meet those needs precisely, with no surprises. There would be no installation issues, no tuning issues, no OS upgrade issues: it would all be nicely packaged under one umbrella.
They could even stop supporting every other OS. That's possible today because of virtualization: it doesn't matter if someone wants to have a Microsoft centric network, they can still run a Linux based app under VMWare or Virtual PC. Certainly dumping all other OSes would be radical, but think how much more support and programming overhead would go away.
Virtualization also means that Oracle Linux wouldn't have to concern itself with anything but Oracle Database. If designing and tuning for that makes it less desirable for some other purpose, so what? Whatever it is, it can be run somewhere else.
Given the performance and reliability gains they'd get from tightly coupling the OS and the database, the other OS versions would probably fade away through natural selection anyway. Customers would have the choice of buying Oracle for Microsoft and getting a product that doesn't run as fast or as reliably as the"Oracle Linux" based version would. If they had any esoteric problems, they'd get to bounce like a ping pong ball between Microsoft and Oracle support. Buy the Linux/Oracle package, and all the support is at one place, and at a place that can actually do more to solve problems.
Controlling the OS makes perfect sense for Oracle. It would make even more sense for their smaller competitors (Progress is a good example). It makes sense for anyone with a special purpose application to peddle.
*Originally published at APLawrnece.com
About the Author:
A.P. Lawrence provides SCO Unix and Linux consulting services http://www.pcunix.com
|
