Re: Informix limitations, should we be using Oracle?

From: Obnoxio The Clown <obnoxio_at_hotmail.com>
Date: Wed, 16 Oct 2002 07:05:43 +0100
Message-ID: <3DAD01B7.2070904_at_hotmail.com>


Simon M. wrote:
>
> We are trying to implement what will become a multi-terrabyte data
> warehouse on Informix XPS but have hit a number of significant
> problems. We are at the point where we are considering switching
> database providers to Oracle but want to be sure that the problems we
> are encountering are indeed valid issues. We have prepared a document
> of which I have included an extract that details the issues we are
> having. If anyone can provide me with feedback on these issues to let
> me know if I'm barking up the wrong tree or if indeed they are issues,
> it would be greatly appreciated. We are currently running Informix XPS
> 8.3.1 on a 4CPU, 4Gb HP N4000 with HP-UX 11.11.

What, a single node box? Jeez... knock yourself out! You're largely wasting your time running XPS on a single node. XPS is really built for great big clusters. Try clustering sixteen of your current machine.

> Informix XPS Issues and Comparison to Oracle
>
> Performance
>
> As detailed on TPC websites
> http://www.tpc.org/tpch/results/tpch_results.asp?orderby=dbms
> http://www.tpc.org/information/benchmarks.asp
>
> Looking at the TPC-H results for a 1000Gb database running on an
> HP9000 Superdome
> runs 2.65 times faster than XPS. The pricing of the two databases as
> reflected in the Price/QphH also shows that Oracle represents 3.5
> times more 'bang for the buck' than XPS.

The XPS benchmark is what, over a year old? Nearly two? That's a long time in benchmarking. Personally, I'm amazed that XPS is still in the top 10. Go ask Oracle for a list of multi-terabyte reference sites.

To be honest, it really doesn't look like anyone in your team knows XPS, can I recommend that you employ someone who does, even if it's just for an initial learning curve. Try jparker at artentech dot com or tim at datad dot com . Received on Wed Oct 16 2002 - 08:05:43 CEST

Original text of this message