Re: db2 vs oracle

From: Mark Townsend <markbtownsend_at_comcast.net>
Date: Sat, 28 Aug 2004 19:18:05 GMT
Message-ID: <4130DA70.3090404_at_comcast.net>


Data Goob wrote:
> Interesting viewpoints about Oracle vs IBM.
>
> I recently scanned through the autobiography about Larry Ellison called
> "SOFTWAR", at my local bookstore. Very interesting book, and full of a
> lot of interesting information, and lots of innaccuracies. The telling
> point at least to me was about Larry and what the company is all about.
> ( You can find it in the business book section under "arrogance" 8-)

So your comments are based on a brief scan of a 500 page book in a bookstore, written by a journalist about a single person, and on that you determine the usefulness of the technology produced by over 40,000 other people ? Great analysis.
http://www.urbandictionary.com/define.php?term=goob&r=f

>
> In the book they mention that Oracle is more about applications than the
> database. If you have an application that requires Oracle, you will indeed
> have to use Oracle. But I would caution against using Oracle as a database
> choice especially in light of where your people are in skills. In Larry's
> own words they indicate the direction of the company has less to do with
> being a database company and more to do with applications.

Complete and utter bullshit. You have absolutely NO basis for your characterization, yet you make it. Why ?

> Larry is in
> his own words more interesting in winning than providing a product that
> is indicative of being a good technology choice. Certainly the grid is
> interesting, but it is not necessarily clustering, nor is it really even
> applicable to a lot of business requirements.

You obviously have no idea how the grid applies to business. Note that one of the foremost proponents of the grid architecture is indeed Ford.

> Oracle will be a big nut
> to crack in an organization that has never used a relational database.

2 days is all it takes - see
http://download-west.oracle.com/docs/cd/B14117_01/server.101/b10742/toc.htm Scan it online.

>
> IBM has an excellent set of databases, and the company can also sell you
> a complete end-to-end solution, meaning hardware as well as software.

They cannot actually sell you any business applications, can they. A complete end-to-end as long as you take the actual business out of the definition of the solution (you know, HR, General Ledger, Mnaufacturing, Supply Chain etc). You are a silly person.

> I
> have seen the latest results on benchmarks and both Oracle and DB2 are
> at the top. IBM will dominate the landscape with Power5 hardware, and
> Power6, etc. with loads of innovation that I do not see coming from
> any other vendor.

Correct - IBM hardware is very,very good. Both Oracle and DB2 run very well on it.

However, hardware innovation does not = software innovation.

DB2 is NOT particulary innovative software, and indeed, a large part of the DB2 software devleopment plans are to provide features from other databases to aid in migration from SQL Server and Oracle - witness the addition of both indentities and sequences in the last couple of releases. The latest release of DB2 is a prime example of this - there is very little that is actually innovative (or even new) in the 8.2 release.

>
> It is indeed important to make the choice in the right context as others
> have indicated. Should it really be Oracle vs DB2 as a database choice?
> Or in your case DB2 vs SQL-Server or Ingres or MySQL or Informix? These
> are database products that would probably be more suited to a comparison
> today. Incidentally Oracle has not had a major architectural change in
> its engine since V7, and that was what, 10+ yrs ago? SQL-Server hasn't
> had a major upgrade in what 6 yrs? DB2 has been changing rapidly to meet
> the market, and so are a few of the others.

You are confusing two things. All databases meet new market requirements, some faster than others, and none of them should require major architectural changes to do so. If a database company has required a major architectural change in the last 10 years to meet new market requirements then that is an indication that they have screwed up big time in their planning. The only DB that I'm aware of that did a complete rewrite in the last 10-20 years was Informix with the advent of SMP technology, and look what hapopened to them.

I hope you are not implying that DB2 has had a major architectural change in the last 10 years as it simply hasn't (indeed it's still based on techniques pioneered over 40 years ago, such as shared nothing, read locks, aries etc).

> Go with the latest not the
> late.

Go with the proven not the unproved. Leading edge not bleeding edge. You are a silly person.

> Interesting too that DB2 is morphing very rapidly, responding to
> what customers want. This should be a big checkbox no matter what your
> choice. You won't need the grid unless you're part of the SETI project.

By all means please continue to avoid the grid. Do so with a fervour, and by all means base your technical analysis of products based on a scan of a book in a book store, and your own personal feelings about a single person.

Not for you is any interest in low cost computing, or the practical application of the principles of consolidation, commoditization, standardization and automation to Information Technology.

The grid absolutely does not and should not have a place in your future.

BTW, Darwin was absolutely correct.You might want to scan "The Origin of Species" the next time you are in a book store.

>
> Daniel Morgan wrote:
>

>>> You're using a wrong approach to determine the RDBMS that will suit
>>> your need. The kind of application you're thinking of running should
>>> be the first concern. Are you goning to be running OLTP, DDS or
>>> datawarehouse application. I don't think Oracle will bet UDB, Sybase
>>> or SQL Server when it comes to OLTP application. As far as
>>> datawarehouse is concern, Sybase has a specific product that is design
>>> for that specific application called Sybase IQ.
>>
>>
>>
>> Not according to a lot of published benchmarks. And not according to
>> the owners of the biggest OLTP systems on the planet.
>>
>> But then what does Sybase have to do with the OP's question? The
>> OP specifically stated a choice between DB2 and Oracle and most likely
>> either would work just fine. So comments about SQL Server and Sybase
>> are irrelevant.
>>
>>> The mistake organization make when picking database platform is the
>>> same kind I am seeing from the approach you're taken.  If you running
>>> mission critical application, then you should be concern about backup
>>> and recovery. Oracle backup and recovery is too complicated otherwise
>>> they won't need a 4 days training seesion on the topic. It takes a
>>> couple of hours to teach the same function in other platform. My point
>>> is that you have to think of what is important in the application
>>> you're running.
>>
>>
>>
>> BTW: Oracle backup and recovery unless you are still working with some
>> Paleolithic version consists of a few mouse clicks in OEM. So your
>> comments indicate little but ignorance about the product.
>>

>
Received on Sat Aug 28 2004 - 21:18:05 CEST

Original text of this message