Ed,
We're primarily an Oracle shop here. DB2 was given a round
several years ago & tossed for reasons that I'm not privy to. Basically
the decision was made before I even interviewed for this job. Now
Oracle is "expensive" as it's the #3 expense item in our budget behind
PeopleSoft and HP, in that order. So to reduce costs the options were
SQL*Server or Open Source. We chose Open Source for new, non critical
projects and have implemented PostGreSQL as a result. That was probably
a very good choice, especially after the news that Oracle acquired
InnouDB which is MySql's core transactional engine. Not good news for
MySql users. Why not SQL*Server?? Well for one it was the problems
that MS and Java are known for as well as our moving to Apache and
Tomcat on Linux servers. .Net is considered a proprietary MS framework
that is going to be eclipsed, again, by other forces in the market place
and if you really look into open source stuff you find a lot of better
solutions that are a lot more cost effective and easier to learn. From
what I'm reading here it appears your stuck with a political decision
made by some VP who is either enthralled with Microsoft, is believing
Microsoft's marketing hype (drinking the kool-aid), or is simply very
mis-informed. Pity, many a good intention has been trashed badly by
those forces. As an example, a company in my home town of Manchester NH
was building a CRM package with a lot of customer input on what they
needed. Problem was that they decided, for political reasons and
Microsoft promises, to use a complete Microsoft stack to build and
deploy on. Well when version 1 came out they then found that their
"customers" were not interested in a complete Microsoft stack. They
wanted to use Oracle, DB2, PostGreSQL, and MySql instead on Linux, Unix,
or Aix platforms. Good intentions, end result they folded without
shipping one copy of the product. 30 people unemployed, $500 million in
venture capital lost. OUCH!!!!
-----Original Message-----
From: ken farmer
[mailto:db2udbdba-ezmlmshield-x768717.[Email address protected]
Sent: Friday, October 21, 2005 10:15 AM
To: LazyDBA Discussion
Subject: Re: Converting from Oracle and DB2 to SQL Server
Edwards Ed wrote:
> To All,
> Have you given consideration to SQL Server or DB2?
> (1) Oracle and DB2 must go; they are too expensive,
> (2) We are going in the direction that industry is going in,
> (3) We are going to train all of our you know who to use .Net for
> creating future applications and migrate Oracle Forms,
> Oracle Reports, and PL/SQL Server Pages (PSPs), ASP,
> Cold Fusion, COBOL, etc to .Net,
> (4) We also are going to look at java
> (5) Our database platform will be SQL Server (we already have SQL
> Server databases and it is cheaper).
>
> BTW, they didn't include our "Oracle Data Warehouse" group" and
> they are not going to do anything to them! I wonder why!
Well, a few things off the top of my head:
(1) Oracle and DB2 must go; they are too expensive,
- well, db2 & oracle *can* be more expensive than SQL Server,
but that certainly isn't a consistent rule - I've had to
license $100s of thousands of dollars of sql server products
in the past.
- why is MSSQL it cheaper? are you comparing mainframes to
intel boxes? 8-way Rs/6000s purchased for $200k four years
ago that could be replaced by 2-way intel/power servers?
- was this confirmed by people who know the products and
worked directly with the vendors? For example, small
departmental servers are very cheap on db2 if licensed
right (don't go with ese, etc).
- will sql server result in higher costs in other
areas? Note that a migration to microsoft often involves
dependencies on multiple commercial apps - exchange,
citrix, etc, etc.
- note that these days labor is *far* more expensive than
hardware or software. so, if a decision saves $10k on
software but increases labor by 10% - then you've probably
lost money on it. Since, as you say, the oracle data
warehouse will stay - then you're probably stuck with
unix/oracle skills anyway. A migration to microsoft
will diminish your ability to share resources between
teams - which will probably increase your labor costs.
- back to above labor point - I've found server
administrative labor cheaper on aix or solaris than
either linux or windows: the number of security fixpacks
is so much smaller - and this translates directly into
admin time for testing patches, taking servers offline,
researching new bugs, diagnosing incompatibilities, etc.
When you combine the decreased difference in cost between
power & intel architectures (you can get a 2-way power5
server for < $5k) plus the savings on aix admin, then you
end up saving quite a lot on the total cost of the platform.
(2) We are going in the direction that industry is going in,
- maybe if you are in a microsoft shop, but it could be
easily argued that .net is being eclipsed by agile
methods on top of languages like ruby, python, and LAMP.
- although sql server is doing well - it is in a precarious
position with mysql & postgresql coming up on its bottom-
end.
- since oracle/db2/etc can run on either windows or unix
they should provide increased future-proofing of your
architecture vs a windows-only solution.
(3) We are going to train all of our...
- well, sounds like you've got a long list of technologies
to move over.
- how long will it take to convert everything? What's the
likelihood that they'll come up with a new direction
before this one is completely implemented?
(4) We also are going to look at java
- hmmm, going with sql server & java isn't very wonderful -
here you bump into the constraints built into a microsoft-
only architecture.
- you may find that java projects are easier to implement
on db2/oracle.
ken
---------------------------------------------------------------------
PLEASE CLICK REPLY-ALL TO SEND A REPLY TO EVERYONE
website: http://www.LazyDBA.com
To unsubscribe: http://www.lazydba.com/unsubscribe.html
DB2 & UDB email list listserv db2-l LazyDBA home page