William H. Magill on Wed, 9 Apr 2003 16:15:16 -0400


[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: [PLUG] Can Open Source Replace Oracle?


On Wednesday, April 9, 2003, at 10:46 AM, Edmund Goppelt wrote:
Is she correct, in your view?  Would PostgreSQL or other Open Source
database not be up to the job?  In your opinion, what hardware
configuration does this application require?

This is basically a size and complexity issue.

In general, any OpenSource software usage at the Enterprise Level requires a serious buy-in by Executive Management and commitment of substantial dollars to provide personnel for support, since you can't buy the kind of Enterprise Level support or Open Source that folks accustomed to "glass house" computing expect.

Based on the disclaimer on the web site, I would guess that the database is pretty trivial. Not much more information than was contained on tab cards a few years ago... but the last time I used it (about 5 years ago) the database was on an IBM mainframe. I assume that it still is.

Could it be moved to another platform? Probably.
Could it be moved to another platform easily? I doubt it. There is as much cultural baggage here as political.


The size of the hardware platform necessary is dependent upon the access needs.
You can run any database on a machine with 64K if you are willing to wait.


Hiding behind the 50 user Oracle license issue, then you could run it on a home PC class machine (ie a Dell). Because there are zero resource demands made by that number of users.

However, assuming that number is only a smoke screen, and the number of users accessing the data increases substantially, you can quickly get into a situation where you need two machines -- a front end to handle the web stuff and a back-end to handle the database. And then people start to complain about response time.

Then if the database is actually searchable, CPU demands will skyrocket depending upon the number of records and the size of records.

This size computation becomes directly related to the amount of hardware you need to throw at the problem.

If the database begins to approach gig-a-bytes, you are talking about a VLM machine (Very Large Memory) if you have any interest at all in performance.
Disk I/O can become deadly very quickly with a large database. But if you can throw it all into memory, then access is much quicker... but that probably means you need a 64 bit machine, and Linux support for 64 bits is anything but "solid." (I've been running Alpha Linux for years, remember.)


Or, if you don't go the VLM approach, there is always "big-iron" (which really translates into VLM.)

Another way to approach this issue of course is to have her contact her IBM sales rep and ask THEM to convert things over to Linux for her. It won't be a "home brew" solution, but it will be Linux and Open Source. [Cheap, I never said anything about cheap. :)] And Linux on the Power4 IS a 64 bit solution. Who knows, IBM might even use PostGres instead of DB.

By the way, PostGres is basically a successor to Ingress which can't even come close to Oracle in terms of the size and complexity of databases it supports. ... although 90+% of the world doesn't need that capability, they only think they do. Oracle reps love to play the card: "you're too big and important to run on that dinky thing." CIO Ego's are easy to subvert, and nobody ever chose Oracle at the technical level.

T.T.F.N.
William H. Magill
# Beige G3 - Rev A motherboard - 768 Meg
# Flat-panel iMac (2.1) 800MHz - Super Drive - 768 Meg
# PWS433a [Alpha 21164 Rev 7.2 (EV56)- 64 Meg]- Tru64 5.1a
magill@mcgillsociety.org
magill@acm.org
magill@mac.com

_________________________________________________________________________
Philadelphia Linux Users Group        --       http://www.phillylinux.org
Announcements - http://lists.netisland.net/mailman/listinfo/plug-announce
General Discussion  --   http://lists.netisland.net/mailman/listinfo/plug