Re: Oracle or Informix? (Insights and experiences wanted - NO RELIGIOUS WARS OR FLAMES PLEASE!!!)

"David L Jarvis" <djarvis_at_relay.cablexpress.com> wrote:

:Greetings Oracle and Informix users, developers, and administrators.
 

:I have been charged with comparing these two companies and their
:products and making a decision on which one my employer should choose.
...snip...

Here is a post from the comp.databases.olap group that may be relevant. Although this refers to DW/OLAP use and your requirements where for OLTP it should still be interesting:

Neil Raden <nraden_at_ix.netcom.com> wrote:

:In article <326D3E38.11A7_at_us.oracle.com>,
: Larry Stevens <lstevens_at_us.oracle.com> wrote:
 

:>
:>Some thoughts from Oracle:
:>
:>Oracle7 7.3 has huge improvements for data warehousing. Bitmap indexes,
:>hash joins, parallel-aware optimizer, initial partitioning support,
:>improved star query support, histograms and more. Along with 7.3 comes
:>an updated tuning methodology for taking advantage of the new features.
:>With a methodical approach to optimizing, you may be able to achieve
:>stunning performance gains (or not - it's not a panacea.)
:>
 

:Well, let's set the record straight. Oracle 7.3 has some small, incremental improvements for
:data warehousing. Bitmaps and hash-joins are useful tools, but it's the integration of all of
:these features, including fast loader WITH indexing and full referential integrity, replacement
:of records in the database (rather than just appends) supported in parallel mode, etc. etc. The
:query optimizer is still far too stupid to support data warehousing, expecially in the absence
:of critical capabilities such as multi-table join indexes, DSS-tuned algorithims for GROUP BY,
:DISTINCT, etc.
 

:The 7.3 story is somewhat better performance on some issues, but poor integration of new
:features with each other. The query optimizer is still based on small-table optimization, an
:almost trivial issue in real-life data warehousing applications, the fast loader is almost never
:adequate for real DW situations, requiring poorly informed customers to need embarrassing and
:expernive hardware upgrades mid-project to meet update windows, and, worst of all, the much
:trumpeted parallel query option is only useful for large table scans, a situation that favors
:batch-oriented backroom analytical applications, not on-line analytical processing.
 

:For my money, Informix is way ahead at this point, DB2/PE has a much better query optimizer, and
:the specialized tools, like Red Brick, still enjoy a substantial advantage overall.

Otherwise you should also look into the Informix Universal Server that Informix have annonced for availability by the end of this year. It's in a completely different ballpark from the O' "Universal Server" which is only a strung together set of disparate servers.

Informix has integrated support for many new datatypes and methods (functions) that can be applied to these. Although most of the talk about Universal Server has sentered around new sofisticated datatypes it is also extremely interesting in an OLTP setting with mostly traditional data. There will be so many things you can do so much easier it will be a dream come true for database designers and developers. It will to some extent be hard as well with new concepts and methods to learn, but it will be fruitfull - very much so.

Technologically Informix is clearly in the lead.

Nils.Myklebust_at_idg.no
NM Data AS, P.O.Box 9090 Gronland, N-0133 Oslo, Norway My opinions are those of my company
The Informix FAQ is at http://www.iiug.org Received on Mon Oct 28 1996 - 00:00:00 CET