Comparison of DB2 and Oracle?

[Rhino]

One of my friends, Scott, is a consultant who doesn't currently have
newsgroup access so I am asking these questions for him. I'll be telling him
how to monitor the answers via Google Newsgroup searches.

Scott has heard a lot of hype about DB2 and Oracle and is trying to
understand the pros and cons of each product. I'm quite familiar with DB2
but have never used Oracle so I can't make any meaningful comparisons for
him. He does not have a lot of database background but sometimes has to
choose or recommend a database to his clients.

Scott has enough life-experience to take the marketing information produced
by IBM and Oracle with a grain of salt and would like to hear from real
DBAs, especially ones who are fluent with both products, for their views on
two questions:

1. What are the pros and cons of the current releases of DB2 and Oracle?

2. What other sources of *independent* information are available to help
someone new to databases choose between DB2 and Oracle?

This is *not* a troll and we don't want to start a flame war! Scott just
want some honest facts to help him decide which product is best at which
jobs.

--
Rhino

[Mark Townsend]

michael newport wrote:

Linux / Apache / PHP have taken off because they are reliable and OpenSource.
I predict the same for Ingres.

Linux, Apache and PHP are succesful because there is a strong developer
and user community. Ingres doesn't have this, and making something
OpenSource doesn't cause this community to automatically build.

Linux in particular benefited from the focus companies like Oracle, IBM
and others placed on it. The same level of focus is unlikely to happen
for Ingres.

[michael newport]

> Linux, Apache and PHP are succesful because there is a strong developer

and user community. Ingres doesn't have this, and making something
OpenSource doesn't cause this community to automatically build.
Linux, Apache and PHP did not start off successful. They grew.

Ingres has existed for a long time, the base IS there
comp.databases.ingres
Linux in particular benefited from the focus companies like Oracle, IBM
and others placed on it. The same level of focus is unlikely to happen
for Ingres.

Companies focus on Linux because it is free. A huge advantage.

Oh and by the way Ingres is also free.

[michael newport]

> How many OS versions of DB2 are on the market?

How many OS versions of Oracle?
How many OS versions of Ingres ?
For DB2 you find different databases for quite every platform (OS 390,
UNIX, AIX, mainframe...) - name it. For every problem they have a
database - incompatible between each other...
In Oracle you deal with the same architecture on every OS platform
they support.
In Ingres you deal with the same architecture on every OS platform
they support.
Some of the things I like in Oracle

* a lot of features to select from (Oracles index types i.e.)
* the shared sql approach
* multi-versioning and read consistency implementation (SELECT without
being blocked by writes i.e.)
you would like Ingres then.
at least, all databases return the data that you store,

depends on the human factor.

[Serge Rielau]

Michael,

This thread is now very much off topic.
Comparison between Oracle and Ingres should be in these respective
newsgroups if you insist on having these debates.
I don't see where comp.databases.ibm-db2 is relevant here.

Cheers
Serge

[Jean-David Beyer]

michael newport wrote:

Well, you need to get more experience with new stuff. Doing the same
thing over in a different environment should give you an increased
appreciation of what you are doing, and what you could be doing.

It did, and the similarities were all too obvious.

That's not the fault of the product. That direct and proximate
responsibility falls on you for being a dinosaur. How much code have
you implemented with bulk binding? How much with the model clause?
How much with analytic functions? How many materialized views with
refresh logs?

its answers the users needs.
and it was written by the dealine.
which meant my company got paid.
although some of this money was then sent to Oracle to pay for the
licence.
if we had used Ingres we could have done the same job for less, or
increased our profits.

I used to work for a vendor of a product that worked on multiple
databases, including Ingres. They dropped Ingres support due to lack
of interest from potential customers. Are you sure whoever paid your
company would have been interested with Ingres? Many products are
considered more desireable simply because they are more expensive.
Stupid, true, but the way of the world.

I agree that CA sales and marketing were bad. But Ingres the product is not.
CA also wasted time and money on speculative products like Jasmine and Opal.
Linux / Apache / PHP have taken off because they are reliable and OpenSource.
I predict the same for Ingres.

I would be curious what the advantages of Ingres might be over other free
(depending on exact usage) dbms's such as postgreSQL and MySQL. I know
that Ingres has been around since even before Oracle existed (late
1970s?). I suppose postgreSQL is a descendant of Ingres.

For desktop use, it probably matters little, though after fussing around
with a bunch of them, I chose to pay IBM for their DB2 UDB because it just
plain worked better and they seemed to follow standards (such as for
Embedded SQL) better than did Informix or postgreSQL did at the time I
tried them (mid to late 1990s).

--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ Registered Machine 241939.
/( )\ Shrewsbury, New Jersey http://counter.li.org
^^-^^ 13:05:00 up 4 days, 14:57, 3 users, load average: 5.37, 5.04, 4.59

[Joel Garry]

Mark Townsend <ma***********@comcast.net> wrote in message news:<41**************@comcast.net>...

michael newport wrote:

Linux / Apache / PHP have taken off because they are reliable and OpenSource.
I predict the same for Ingres.
>
Linux, Apache and PHP are succesful because there is a strong developer
and user community. Ingres doesn't have this, and making something
OpenSource doesn't cause this community to automatically build.

That's an interesting perspective, Mark. I agree for linux (certainly
my own interest in linux stemmed from getting away from windows
together with unix bigotry - and now that new versions don't run
[without a whole lot of work] on any of my half-dozen computers and
are suffering from bloatware, I'm not so interested for my own sake).

Apache and PHP, on the other hand, I would argue, benefitted from
having a "killer app," and some cluelessness from MS. Of course, now
that I've argued it, I have to say linux also benefitted from the
same, but subsequent to the developer interest. They all started
because of the strong developer community, but success came later.
For Apache, only a year later, taking over the high growth market that
NCSA created, but later nonetheless.

Linux in particular benefited from the focus companies like Oracle, IBM
and others placed on it. The same level of focus is unlikely to happen
for Ingres.

And so my interest in linux is now professional, rather than personal.
This is a good thing. But linux is only just now becoming successful
in the db world. Some people might still consider it the next big
thing. It's here, bleeding edge places are using it, but
mainstream... maybe all those hp/solaris job ads reflect the
forward-looking people leaving :-)

From my perspective, the only thing that can save Ingres is a killer
app. And I think anyone writing anything new would choose something
else (or db independence), unless some underemployed Ingres people
brainstormed and got lucky.

jg
--
@home.com is bogus.
I'll gladly agree to take only 25%.
http://www.signonsandiego.com/uniont...b27lerach.html

[DA Morgan]

michael newport wrote:

Hans,

development and maintenance costs are human factors.

Yes! And they are onging.

If the analysis and /or programming is bad then these costs will be higher.
And do not forget, Oracle will charge you for a licence, this is ongoing.

To reduce the total cost of a project over several years,

Reduce development and maintenance costs,

Human.

By writing and maintaining LESS code,

Human.

By having more capability in the vendor's product,

The database market is saturated with capable products.
What does RMAN do that the OS does not ?

By using that capability.

Human.

Many many things so rather than spouting off about things of which
you have no knowledge why don't you invest some effort and learn
about them. All you've done so far is advertise your lack of knowledge
about both Oracle and DB2.

But for starters ... incremental backup of changed blocks.
--
Daniel A. Morgan
University of Washington
da******@x.washington.edu
(replace 'x' with 'u' to respond)

[DA Morgan]

michael newport wrote:

So why buy Oracle when Ingres is free.
If you use Oracle like you use Ingres, you are absolutely correct.

Its just a database. You use it as you need to, to do your job.
See previous post for comparisons on how to do this.

Having stuck your foot into your mouth I was assuming you wouldn't
follow up by trying to swallow it. You claim to be a database
developer but your posts read like a first year Java newbie.
Yes Ingres has changed significantly in the last 7 years.
It is also free.
That is $400 dollars saved.
Now imagine if you were a large company.

$400 is less than we spend in a week for free softdrinks for our
employees. Get a life.

--
Daniel A. Morgan
University of Washington
da******@x.washington.edu
(replace 'x' with 'u' to respond)

[DA Morgan]

michael newport wrote:

DA Morgan <da******@x.washington.edu> wrote in message news:<1098752193.942650@yasure>...

michael newport wrote:

I am now (1 year) working with Oracle and my work involves doing the
same stuff that I did with Ingres (see previous post).

That's not the fault of the product. That direct and proximate
responsibility falls on you for being a dinosaur. How much code have
you implemented with bulk binding? How much with the model clause?
How much with analytic functions? How many materialized views with
refresh logs?

its answers the users needs.
and it was written by the dealine.
which meant my company got paid.
although some of this money was then sent to Oracle to pay for the
licence.
if we had used Ingres we could have done the same job for less, or
increased our profits.

Had you used Ingres the customer likely would have hired a firm that
knew how to work in a real database. Your firm would have been paid $0
and you'd be unemployed.

Why not just admit that you have reached the point in your life
where you want technology to stop and let you keep doing what you
did in neolithic times.

If you are talking about Oracle report server, I will agree with you.
if you are talking about JAVA then you need to thank SUN not Oracle.
But technology for its own sake is a waste of money.

Not once in this entire thread have Oracle Report Server (it doesn't
even exist any more) or JAVA been part of any post. Don't try to change
the subject. That is a activity best left to small children.
--
Daniel A. Morgan
University of Washington
da******@x.washington.edu
(replace 'x' with 'u' to respond)

[DA Morgan]

michael newport wrote:

DA Morgan <da******@x.washington.edu> wrote in message news:<1098752226.378168@yasure>...

michael newport wrote:

Daniel,

what do you do at the University of Washington ?

nothing to do with education ?

Regards
Michael Newport

Teach databases something that might have interested you
once in your life.

I am still interested, which is why we are having this discussion.
But rather than back a product because it has a particular brand,
I prefer a more realistic discussion of experience.

Have you ever used Ingres ?

I don't "back" a product. I work routinely in multiple products. That
I teach one relates to what the university's students want ... not what
I do.

But no one wants to learn Ingres. It is a decaying corpse that CA has
attempted to bury at sea. If you want to work with a real open-source
database the clear choice is MySQL.

--
Daniel A. Morgan
University of Washington
da******@x.washington.edu
(replace 'x' with 'u' to respond)

[DA Morgan]

HansF wrote:

michael newport wrote:

And do not forget, Oracle will charge you for a licence, this is ongoing.

Aaah ... the ongoing license fee myth! (#83?) Sadly incorrect, just as your
other myths.
Oracle charges once for a perpetual license. An Oracle license allows you to
use it forever if you wish (which is why there are still sites using
Oracle6).

Actually Boeing still has two Apollo server with Oracle 5. And still on
the original license.
You can, if you wish, get support for a license. That is annual, and
provides unlimited support calls. Quite different from licensing.

<heavy sigh>

When the facts don't support a position it is not uncommon to have
pseudofacts invented.
--
Daniel A. Morgan
University of Washington
da******@x.washington.edu
(replace 'x' with 'u' to respond)

[DA Morgan]

michael newport wrote:

I agree that CA sales and marketing were bad. But Ingres the product is not.
CA also wasted time and money on speculative products like Jasmine and Opal.
Linux / Apache / PHP have taken off because they are reliable and OpenSource.
I predict the same for Ingres.

And exactly why doesn't your "critical" understanding of the marketplace
indicate anything about MySQL, PostgreSQL, Firebird, and all of the
other open-source code bases? You are a religious zealot pushing
neolithic technology likely because it is all you know.

There is not a chance in 1000 that it will ever have 1/2 the marketshare
of MySQL which is the only serious open-source competitor to Oracle and
DB2.
--
Daniel A. Morgan
University of Washington
da******@x.washington.edu
(replace 'x' with 'u' to respond)

[DA Morgan]

michael newport wrote:

Linux, Apache and PHP are succesful because there is a strong developer
and user community. Ingres doesn't have this, and making something
OpenSource doesn't cause this community to automatically build.

Linux, Apache and PHP did not start off successful. They grew.

Ingres has existed for a long time, the base IS there
comp.databases.ingres

The "base" is database developers not people that write kernel code in
C. They will all die of old age before they figure out how to give the
Ingres kernel capabilities that were in Oracle 8i.

Linux in particular benefited from the focus companies like Oracle, IBM
and others placed on it. The same level of focus is unlikely to happen
for Ingres.

Companies focus on Linux because it is free. A huge advantage.

Nonsense. Absolute ignorant nonsense. I consult for a division of The
Boeing company. The cost of an operating system compared to the total
cost of an application is so small as to be invisible. Do you really
think we are going to build a $15,000,000 system and worry about the
lousy few hundred or few thousand dollars for the O/S?

We chose Linux because it gave us better performance, in lab tests with
our application than did Win2K, WinXP, Solaris 2.9 and HP/UX 11i.

You really are out of touch with reality.
Oh and by the way Ingres is also free.

Oh and by the way ... we really don't give a damn. And neither
does anyone else.
--
Daniel A. Morgan
University of Washington
da******@x.washington.edu
(replace 'x' with 'u' to respond)

[DA Morgan]

Serge Rielau wrote:

Michael,

This thread is now very much off topic.
Comparison between Oracle and Ingres should be in these respective
newsgroups if you insist on having these debates.
I don't see where comp.databases.ibm-db2 is relevant here.

Cheers
Serge

I agree. I am kill-filing this thread. Newport is just trolling and I
am tired of responding to a neolithic fanatic. You'll so no more on
this from me.
--
Daniel A. Morgan
University of Washington
da******@x.washington.edu
(replace 'x' with 'u' to respond)

[michael newport]

> Not once in this entire thread have Oracle Report Server (it doesn't

even exist any more) or JAVA been part of any post. Don't try to change
the subject. That is a activity best left to small children.
Daniel,

in Ingres I wrote 4GL, in Oracle I write PL/SQL
Because you don't know how to write Java in the database?
in Ingres I wrote SQL, in Oracle I write SQL

Because you don't know how to write Java in the database?
Daniel,
it seems that you have short term memory loss,

it also seems that you have never used Ingres.

Regards
Michael Newport

[michael newport]

Serge,

I already did an oracle vs ingres thread but was disappointed with the
answers that I received.

Please notice that I only commented on databases that I have
experience of,
unlike our favourite troll Mr.Morgan who has a problem with anything
not spelt Oracle.

Regards
Michael Newport

[Darin McBride]

DA Morgan wrote:

You can, if you wish, get support for a license. That is annual, and
provides unlimited support calls. Quite different from licensing.

<heavy sigh>

When the facts don't support a position it is not uncommon to have
pseudofacts invented.

To be really fair, this is just nitpicking on the semantics. To most
people, "licensing" really means "total cost payable to the vendor."
If I must buy an annual support contract to ensure the product's
success in my environment, that's the same thing, at the end of the
day - money going from my company to Oracle, IBM, MS, CA, whatever.

Most of the thread arguing against Ingres has talked vaguely about a
"total cost" of ownership - and the one time Michael has a valid point
about the vendor portion of that TCO (which, of course, does not
validate anything else he says), you go and nitpick his terminology.

Then again, perhaps it's not uncommon that when your opponent is
generally making no sense, that you stop reading his posts objectively,
and just assume that the whole argument is absurd, rather than just the
individual (and overwhelming) portions of it that really are absurd?

[HansF]

Darin McBride wrote:

To be really fair, this is just nitpicking on the semantics. To most
people, "licensing" really means "total cost payable to the vendor."
If I must buy an annual support contract to ensure the product's
success in my environment, that's the same thing, at the end of the
day - money going from my company to Oracle, IBM, MS, CA, whatever.

You seem to separate vendor cost from internal cost.

The implication is that any development costs or maintenance costs resulting
from inventing software to compensate for capabilities not in Ingres are
not to be counted, but paying Oracle or IBM for those same capabilities are
to be counted. I'd be concerned about that style of accounting - it's
quite reminiscient of the CapEx vs OpEx accounting invented to get around
regulations in some industries.

No matter which way we try to wiggle, companies need to manage total cost,
not just "total cost payable to the vendor." The proof is in the attempts
at outsourcing - whether it works well (or works at all) is irrelevant, the
relevance is that companies are doing this (in desparation?) to get total
costs under control.

However, as Serge says, this thread is WAY off topic and no longer relevant.
If you want to continue this discussion, I suggest we go to some Ingres or
open source advocacy group. I hereby stop responding to the Ingres and
Open Source discussion in this thread and apologize to all for not having
stopped sooner.

/Hans

[michael newport]

> Then again, perhaps it's not uncommon that when your opponent is

generally making no sense, that you stop reading his posts objectively,
and just assume that the whole argument is absurd, rather than just the
individual (and overwhelming) portions of it that really are absurd?

which bit did you have trouble with ?

[michael newport]

Jean-David Beyer <jd*****@exit109.com> wrote in message news:<10*************@corp.supernews.com>...

michael newport wrote:

Well, you need to get more experience with new stuff. Doing the same
thing over in a different environment should give you an increased
appreciation of what you are doing, and what you could be doing.

It did, and the similarities were all too obvious.

>That's not the fault of the product. That direct and proximate
>responsibility falls on you for being a dinosaur. How much code have
>you implemented with bulk binding? How much with the model clause?
>How much with analytic functions? How many materialized views with
>refresh logs?

its answers the users needs.
and it was written by the dealine.
which meant my company got paid.
although some of this money was then sent to Oracle to pay for the
licence.
if we had used Ingres we could have done the same job for less, or
increased our profits.

I used to work for a vendor of a product that worked on multiple
databases, including Ingres. They dropped Ingres support due to lack
of interest from potential customers. Are you sure whoever paid your
company would have been interested with Ingres? Many products are
considered more desireable simply because they are more expensive.
Stupid, true, but the way of the world.

I agree that CA sales and marketing were bad. But Ingres the product is not.
CA also wasted time and money on speculative products like Jasmine and Opal.
Linux / Apache / PHP have taken off because they are reliable and OpenSource.
I predict the same for Ingres.

I would be curious what the advantages of Ingres might be over other free
(depending on exact usage) dbms's such as postgreSQL and MySQL. I know
that Ingres has been around since even before Oracle existed (late
1970s?). I suppose postgreSQL is a descendant of Ingres.

For desktop use, it probably matters little, though after fussing around
with a bunch of them, I chose to pay IBM for their DB2 UDB because it just
plain worked better and they seemed to follow standards (such as for
Embedded SQL) better than did Informix or postgreSQL did at the time I
tried them (mid to late 1990s).

Open Source is good, and not just Ingres.
But I used Ingres for a long time, and I know it works.
Oracle also works but costs a lot of money.

I also read that IBM and Sybase appear to be going opensource.

[Darin McBride]

michael newport wrote:

Then again, perhaps it's not uncommon that when your opponent is
generally making no sense, that you stop reading his posts objectively,
and just assume that the whole argument is absurd, rather than just the
individual (and overwhelming) portions of it that really are absurd?

which bit did you have trouble with ?

The lack of "T" in your "TCO".

I mean, there are many different costs in owning software. Not just
the initial cost.

1. Purchase cost.

This is what you seem to be focusing on. Unfortunately, it's not the
total cost. For most larger databases, it isn't even always a
significant portion of the total cost.

2. Support costs.

This, with #1, is what you pay to the vendor, and often significantly
outweighs the purchase cost. Sure, Ingres may be free to "purchase",
but what about support costs if/when something goes wrong?

At one time, support came free with purchase. Nowadays, it is swinging
heavily in the other direction, especially with commodity (read: open
source) software. The cost of 24/7 within-the-hour support is
significant, but so is its peace of mind.

3. Development costs.

This is what the purchaser spends to integrate the software into their
infrastructure. This may be a lonely IT tech in a closet somewhere
figuring out how to get the software installed, or it may be an entire
software development engineering team with a few DBAs trying to
architect their business model inside the database. Generally
speaking, this outweighs both #1 and #2 together.

If, then, the database product provides functions, stored procedures,
and other database-isms ("Oracle-isms" or "DB2-isms" for the newsgroups
getting this cross-posted) which save you 2 weeks of development time
in the pursuit of your business goals, right there you've saved a
significant portion of your purchase cost of any of the "expensive"
database vendors. I know that 2 weeks of my time is worth way more
than $400 - although I suspect most DB2 or Oracle deployments cost more
than $400 in purchase costs. Even with $20,000 in purchase costs, if
it saves me 4 weeks in development time, and a corresponding 1-2 weeks
in testing time (since I shouldn't need to debug that function - IBM or
Oracle have already done that for me), I've saved a significant portion
of that purchase cost... at least if I'm contracting. And we get to
market (deployment) 5-6 weeks earlier. If this new database
application is supposed to save the whole corporation 1 hour of work
per person per month, and there are 1000 employees, that's 1250-1500
hours saved in those extra 5-6 weeks, and it only takes an average of
$10/hour to pay for the rest of the purchase price of $20,000. In
other words, the "purchase price" is FREE at the point where the
application would be deployed if I didn't have those extra built-in
functions.

And it's this last area that you seem to keep ignoring. I don't think
it's me who is having trouble with the thread...

[Serge Rielau]

michael newport wrote:

I also read that IBM and Sybase appear to be going opensource.

IBM is a company, not a product.
IBM Cloudscape has been open sourced as "Derby".
There aren't even rumours that IBM may open source one of it's
mainstream commercial RDBMS (DB2, IDS, XPS and RedBrick)

Cheers
Serge

[DA Morgan]

Darin McBride wrote:

DA Morgan wrote:

You can, if you wish, get support for a license. That is annual, and
provides unlimited support calls. Quite different from licensing.

<heavy sigh>

When the facts don't support a position it is not uncommon to have
pseudofacts invented.

To be really fair, this is just nitpicking on the semantics. To most
people, "licensing" really means "total cost payable to the vendor."
If I must buy an annual support contract to ensure the product's
success in my environment, that's the same thing, at the end of the
day - money going from my company to Oracle, IBM, MS, CA, whatever.

Most of the thread arguing against Ingres has talked vaguely about a
"total cost" of ownership - and the one time Michael has a valid point
about the vendor portion of that TCO (which, of course, does not
validate anything else he says), you go and nitpick his terminology.

Then again, perhaps it's not uncommon that when your opponent is
generally making no sense, that you stop reading his posts objectively,
and just assume that the whole argument is absurd, rather than just the
individual (and overwhelming) portions of it that really are absurd?

Buy a single copy of Personal Oracle. Get a support contract for less
than $200. Give me a break! I paid more for dinner and a bottle of wine
last night.
--
Daniel A. Morgan
University of Washington
da******@x.washington.edu
(replace 'x' with 'u' to respond)

[michael newport]

> 2. Support costs.

This, with #1, is what you pay to the vendor, and often significantly
outweighs the purchase cost. Sure, Ingres may be free to "purchase",
but what about support costs if/when something goes wrong?

At one time, support came free with purchase. Nowadays, it is swinging
heavily in the other direction, especially with commodity (read: open
source) software. The cost of 24/7 within-the-hour support is
significant, but so is its peace of mind.
I agree support costs money.
But this is not product dependent.
It depends on the support you need.

Mature products give you peace of mind, and Ingres has a long history.
3. Development costs.

This is what the purchaser spends to integrate the software into their
infrastructure. This may be a lonely IT tech in a closet somewhere
figuring out how to get the software installed, or it may be an entire
software development engineering team with a few DBAs trying to
architect their business model inside the database. Generally
speaking, this outweighs both #1 and #2 together.

If, then, the database product provides functions, stored procedures,
and other database-isms ("Oracle-isms" or "DB2-isms" for the newsgroups
getting this cross-posted) which save you 2 weeks of development time
in the pursuit of your business goals, right there you've saved a
significant portion of your purchase cost of any of the "expensive"
database vendors. I know that 2 weeks of my time is worth way more
than $400 - although I suspect most DB2 or Oracle deployments cost more
than $400 in purchase costs. Even with $20,000 in purchase costs, if
it saves me 4 weeks in development time, and a corresponding 1-2 weeks
in testing time (since I shouldn't need to debug that function - IBM or
Oracle have already done that for me), I've saved a significant portion
of that purchase cost... at least if I'm contracting. And we get to
market (deployment) 5-6 weeks earlier. If this new database
application is supposed to save the whole corporation 1 hour of work
per person per month, and there are 1000 employees, that's 1250-1500
hours saved in those extra 5-6 weeks, and it only takes an average of
$10/hour to pay for the rest of the purchase price of $20,000. In
other words, the "purchase price" is FREE at the point where the
application would be deployed if I didn't have those extra built-in
functions.

And it's this last area that you seem to keep ignoring. I don't think
it's me who is having trouble with the thread...

Again, these costs are entirely dependent on people, not product.

More importantly OpenSource software is yours to change.

[michael newport]

Serge,

would you like to see these other IBM products OpenSourced ?

Regards
Michael Newport

[michael newport]

> >>michael newport wrote:

Daniel,

what do you do at the University of Washington ?

nothing to do with education ?

Regards
Michael Newport

Teach databases something that might have interested you
once in your life.

I am still interested, which is why we are having this discussion.
But rather than back a product because it has a particular brand,
I prefer a more realistic discussion of experience.

Have you ever used Ingres ?

I don't "back" a product. I work routinely in multiple products. That
I teach one relates to what the university's students want ... not what
I do.

But no one wants to learn Ingres. It is a decaying corpse that CA has
attempted to bury at sea. If you want to work with a real open-source
database the clear choice is MySQL.

So you have never used Ingres.

I guess the OpenSource tide brought Ingres back to shore.

Why is the clear choice MySQL ?

[Noons]

"Jim Kennedy" <ke****************************@attbi.net> wrote in message news:<Tzifd.6802$HA.6215@attbi_s01>...

If this is the same Ingres I used awhile ago I wouldn't touch it with a ten
foot pole even if you paid me. The concurrency model sucks, start a
transaction, insert a record, lock 95% of the table if it has a primary
key - because the page locks on the index locks most of the pages. NO ONE
ELSE COULD GET ANY WORK DONE, unless you threw out the transaction model and
went to auto commit. POS.

The very same POS. That got shafted out of the market
PRECISELY because of the crap it always was.
Did you ever try crashing the server? Best way to ensure
you lost all your work, with Ingres.

[Serge Rielau]

You are asking me whether I want to be fired from my current job.
I currenly own a good portion of DB2 for LUW's SQL Compiler code.
For sure I'm not in it for the charitable work although it sometimes
feels like it.

Cheers
Serge

[Darin McBride]

michael newport wrote:

2. Support costs.

This, with #1, is what you pay to the vendor, and often significantly
outweighs the purchase cost. Sure, Ingres may be free to "purchase",
but what about support costs if/when something goes wrong?

At one time, support came free with purchase. Nowadays, it is swinging
heavily in the other direction, especially with commodity (read: open
source) software. The cost of 24/7 within-the-hour support is
significant, but so is its peace of mind.

I agree support costs money.
But this is not product dependent.
It depends on the support you need.

Mature products give you peace of mind, and Ingres has a long history.

Not quite - unchanging products give you peace of mind - as long as new
features are added, things can break.

3. Development costs.

This is what the purchaser spends to integrate the software into their
infrastructure. This may be a lonely IT tech in a closet somewhere
figuring out how to get the software installed, or it may be an entire
software development engineering team with a few DBAs trying to
architect their business model inside the database. Generally
speaking, this outweighs both #1 and #2 together.

If, then, the database product provides functions, stored procedures,
and other database-isms ("Oracle-isms" or "DB2-isms" for the newsgroups
getting this cross-posted) which save you 2 weeks of development time
in the pursuit of your business goals, right there you've saved a
significant portion of your purchase cost of any of the "expensive"
database vendors. I know that 2 weeks of my time is worth way more
than $400 - although I suspect most DB2 or Oracle deployments cost more
than $400 in purchase costs. Even with $20,000 in purchase costs, if
it saves me 4 weeks in development time, and a corresponding 1-2 weeks
in testing time (since I shouldn't need to debug that function - IBM or
Oracle have already done that for me), I've saved a significant portion
of that purchase cost... at least if I'm contracting. And we get to
market (deployment) 5-6 weeks earlier. If this new database
application is supposed to save the whole corporation 1 hour of work
per person per month, and there are 1000 employees, that's 1250-1500
hours saved in those extra 5-6 weeks, and it only takes an average of
$10/hour to pay for the rest of the purchase price of $20,000. In
other words, the "purchase price" is FREE at the point where the
application would be deployed if I didn't have those extra built-in
functions.

And it's this last area that you seem to keep ignoring. I don't think
it's me who is having trouble with the thread...

Again, these costs are entirely dependent on people, not product.

More importantly OpenSource software is yours to change.

Ok, I see where you're coming from now. But I think you missed
something. If I use a smaller product, such as Ingres, which doesn't
have a function which takes me 4 weeks to implement, vs using Oracle or
DB2 or MSSQL (big three) which does have that function, saving me, in
effect, 4 weeks of development, then the "pricey" database just cost me
nothing - the costs and the savings cancel each other out.

Small, stable vendor means reinventing the wheel on many projects.

[DA Morgan]

Darin McBride wrote:

michael newport wrote:

2. Support costs.

This, with #1, is what you pay to the vendor, and often significantly
outweighs the purchase cost. Sure, Ingres may be free to "purchase",
but what about support costs if/when something goes wrong?

At one time, support came free with purchase. Nowadays, it is swinging
heavily in the other direction, especially with commodity (read: open
source) software. The cost of 24/7 within-the-hour support is
significant, but so is its peace of mind.

I agree support costs money.
But this is not product dependent.
It depends on the support you need.

Mature products give you peace of mind, and Ingres has a long history.

Not quite - unchanging products give you peace of mind - as long as new
features are added, things can break.

3. Development costs.

This is what the purchaser spends to integrate the software into their
infrastructure. This may be a lonely IT tech in a closet somewhere
figuring out how to get the software installed, or it may be an entire
software development engineering team with a few DBAs trying to
architect their business model inside the database. Generally
speaking, this outweighs both #1 and #2 together.

If, then, the database product provides functions, stored procedures,
and other database-isms ("Oracle-isms" or "DB2-isms" for the newsgroups
getting this cross-posted) which save you 2 weeks of development time
in the pursuit of your business goals, right there you've saved a
significant portion of your purchase cost of any of the "expensive"
database vendors. I know that 2 weeks of my time is worth way more
than $400 - although I suspect most DB2 or Oracle deployments cost more
than $400 in purchase costs. Even with $20,000 in purchase costs, if
it saves me 4 weeks in development time, and a corresponding 1-2 weeks
in testing time (since I shouldn't need to debug that function - IBM or
Oracle have already done that for me), I've saved a significant portion
of that purchase cost... at least if I'm contracting. And we get to
market (deployment) 5-6 weeks earlier. If this new database
application is supposed to save the whole corporation 1 hour of work
per person per month, and there are 1000 employees, that's 1250-1500
hours saved in those extra 5-6 weeks, and it only takes an average of
$10/hour to pay for the rest of the purchase price of $20,000. In
other words, the "purchase price" is FREE at the point where the
application would be deployed if I didn't have those extra built-in
functions.

And it's this last area that you seem to keep ignoring. I don't think
it's me who is having trouble with the thread...

Again, these costs are entirely dependent on people, not product.

More importantly OpenSource software is yours to change.

Ok, I see where you're coming from now. But I think you missed
something. If I use a smaller product, such as Ingres, which doesn't
have a function which takes me 4 weeks to implement, vs using Oracle or
DB2 or MSSQL (big three) which does have that function, saving me, in
effect, 4 weeks of development, then the "pricey" database just cost me
nothing - the costs and the savings cancel each other out.

Small, stable vendor means reinventing the wheel on many projects.

Well said. Not once in my experience has the cost of any database
product been 10% of the cost of a system on which it was implemented:
Often substantially less than that. Last project on which I worked we
spent $80,000 for the database licenses. We spent $2.5M on development.
I'm going to sweat 3.2% of the system cost vs. the cost of testing
changes to the product's kernel to implement missing features? Perhaps
in someone's dreams ... most certainly in my nightmares. Mr. Newport has
obviously never done serious development where he was responsible for
the budget too.
--
Daniel A. Morgan
University of Washington
da******@x.washington.edu
(replace 'x' with 'u' to respond)

[michael newport]

> Not quite - unchanging products give you peace of mind - as long as new

features are added, things can break.
Test plans are often overlooked, but this is people dependent.

3. Development costs.

This is what the purchaser spends to integrate the software into their
infrastructure. This may be a lonely IT tech in a closet somewhere
figuring out how to get the software installed, or it may be an entire
software development engineering team with a few DBAs trying to
architect their business model inside the database. Generally
speaking, this outweighs both #1 and #2 together.

If, then, the database product provides functions, stored procedures,
and other database-isms ("Oracle-isms" or "DB2-isms" for the newsgroups
getting this cross-posted) which save you 2 weeks of development time
in the pursuit of your business goals, right there you've saved a
significant portion of your purchase cost of any of the "expensive"
database vendors. I know that 2 weeks of my time is worth way more
than $400 - although I suspect most DB2 or Oracle deployments cost more
than $400 in purchase costs. Even with $20,000 in purchase costs, if
it saves me 4 weeks in development time, and a corresponding 1-2 weeks
in testing time (since I shouldn't need to debug that function - IBM or
Oracle have already done that for me), I've saved a significant portion
of that purchase cost... at least if I'm contracting. And we get to
market (deployment) 5-6 weeks earlier. If this new database
application is supposed to save the whole corporation 1 hour of work
per person per month, and there are 1000 employees, that's 1250-1500
hours saved in those extra 5-6 weeks, and it only takes an average of
$10/hour to pay for the rest of the purchase price of $20,000. In
other words, the "purchase price" is FREE at the point where the
application would be deployed if I didn't have those extra built-in
functions.

And it's this last area that you seem to keep ignoring. I don't think
it's me who is having trouble with the thread...

Again, these costs are entirely dependent on people, not product.

More importantly OpenSource software is yours to change.

Ok, I see where you're coming from now. But I think you missed
something. If I use a smaller product, such as Ingres, which doesn't
have a function which takes me 4 weeks to implement, vs using Oracle or
DB2 or MSSQL (big three) which does have that function, saving me, in
effect, 4 weeks of development, then the "pricey" database just cost me
nothing - the costs and the savings cancel each other out.

Small, stable vendor means reinventing the wheel on many projects.

Are you thinking of a particular function ?

I was forced to use Oracle report server (paid for) but found it to be
very buggy, so I had to 'reinvent' some functionality using utl file
and Unix.

When you say big 3, do you mean by market share ?

[michael newport]

DA Morgan <da******@x.washington.edu> wrote in message news:<1098928130.887686@yasure>...

michael newport wrote:

Linux, Apache and PHP are succesful because there is a strong developer
and user community. Ingres doesn't have this, and making something
OpenSource doesn't cause this community to automatically build.

Linux, Apache and PHP did not start off successful. They grew.

Ingres has existed for a long time, the base IS there
comp.databases.ingres

The "base" is database developers not people that write kernel code in
C. They will all die of old age before they figure out how to give the
Ingres kernel capabilities that were in Oracle 8i.

Linux in particular benefited from the focus companies like Oracle, IBM
and others placed on it. The same level of focus is unlikely to happen
for Ingres.

Companies focus on Linux because it is free. A huge advantage.

Nonsense. Absolute ignorant nonsense. I consult for a division of The
Boeing company. The cost of an operating system compared to the total
cost of an application is so small as to be invisible. Do you really
think we are going to build a $15,000,000 system and worry about the
lousy few hundred or few thousand dollars for the O/S?

We chose Linux because it gave us better performance, in lab tests with
our application than did Win2K, WinXP, Solaris 2.9 and HP/UX 11i.

and the reason that Linux exists is that it answers a market need !
people are fed up of paying licence fee's for bloatware.

and as you say yourself a free product can give better performance
than its expensively licenced rivals !!

[michael newport]

> $400 is less than we spend in a week for free softdrinks for our

employees. Get a life.

see a dentist !

[Billy Verreynne]

Why Oracle and not DB2? There are numerous sound technical reasons.

And this..

==
/home/billy/> sqlplus dataware@whs
SQL*Plus: Release 9.2.0.5.0 - Production on Mon Nov 15 15:27:06 2004
Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved.
Enter password:

Connected to:
Oracle9i Enterprise Edition Release 9.2.0.4.0 - 64bit Production
With the Partitioning, Oracle Label Security, OLAP and Oracle Data
Mining options
JServer Release 9.2.0.4.0 - Production

SQL> set timing on
SQL> select count(*) from x25_calls;

COUNT(*)
----------
672839836

Elapsed: 00:00:35.18

SQL> exit
Disconnected from Oracle9i Enterprise Edition Release 9.2.0.4.0 -
64bit Production
With the Partitioning, Oracle Label Security, OLAP and Oracle Data
Mining options
JServer Release 9.2.0.4.0 - Production
==

Now anyone that have an idea what databases are about, will know what
a SELECT COUNT entails, I/O wise.. and how critical table and index
designs plays in optimising access and lowering I/O.

Can any other database, Open Source or commercial, come anywhere close
to this? I doubt it.

And no, this nothing to do with hardware. The above was run against an
old K-class HP-UX platform.

--
Billy

[Pete H]

what a dork...

Pete H
vs****@onwe.co.za (Billy Verreynne) wrote in message news:<1a**************************@posting.google. com>...

Why Oracle and not DB2? There are numerous sound technical reasons.

And this..

==
/home/billy/> sqlplus dataware@whs
SQL*Plus: Release 9.2.0.5.0 - Production on Mon Nov 15 15:27:06 2004
Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved.
Enter password:

Connected to:
Oracle9i Enterprise Edition Release 9.2.0.4.0 - 64bit Production
With the Partitioning, Oracle Label Security, OLAP and Oracle Data
Mining options
JServer Release 9.2.0.4.0 - Production

SQL> set timing on
SQL> select count(*) from x25_calls;

COUNT(*)
----------
672839836

Elapsed: 00:00:35.18

SQL> exit
Disconnected from Oracle9i Enterprise Edition Release 9.2.0.4.0 -
64bit Production
With the Partitioning, Oracle Label Security, OLAP and Oracle Data
Mining options
JServer Release 9.2.0.4.0 - Production
==

Now anyone that have an idea what databases are about, will know what
a SELECT COUNT entails, I/O wise.. and how critical table and index
designs plays in optimising access and lowering I/O.

Can any other database, Open Source or commercial, come anywhere close
to this? I doubt it.

And no, this nothing to do with hardware. The above was run against an
old K-class HP-UX platform.

[Billy Verreynne]

ph******@intellicare.com (Pete H) wrote:

what a dork...

And that is the best you can do Pete in response to a [SELECT COUNT]
on a VLT containing 672,839,836 rows that returns the answer in 35
seconds?

I've read Oracle being slammed for this and that and what not. So
instead of responding in kind, I simply show what Oracle is capable of
in the real world.

It is also not about counting rows in general. It is *what* it entails
(think I/O) and *how* it does it.

And the How It Is Done is what differentiate Oracle from others.
Inovative means of providing accurate and consistent answers - thus
enabling this very visible performance with a [SELECT COUNT]. And
this type of innovation and performance is across the board. Not just
with a [SELECT COUNT]. Though the latter tend to drive home the point
with an extra sharp and shiny edge.
--
Billy

[Lady Chatterly]

In article <63*************************@posting.google.com> mi************@yahoo.com (michael newport) wrote:

Serge,

would you like to see these other IBM products OpenSourced ?
I see what you mean.
Regards
Michael Newport

Why are you so sure?

--
Lady Chatterly

"I don't know who she is. I doubt that its a bot. I have my guess as
to who it is. Regard the frequency of posts. What frequent poster is
missing? That Be Packing, is an old mind trick. Ignore it." -- Pip

[Lady Chatterly]

In article <ed**************************@posting.google.com > in**@Boecker-OCP.com (Yukonkid) wrote:

"Rhino" <rh****@NOSPAM.sympatico.ca> wrote in message news:<33*****************@news20.bellglobal.com>.. .

One of my friends, Scott, is a consultant who doesn't currently have
newsgroup access so I am asking these questions for him. I'll be telling him
how to monitor the answers via Google Newsgroup searches.

Scott has heard a lot of hype about DB2 and Oracle and is trying to
understand the pros and cons of each product. I'm quite familiar with DB2
but have never used Oracle so I can't make any meaningful comparisons for
him. He does not have a lot of database background but sometimes has to
choose or recommend a database to his clients.

Scott has enough life-experience to take the marketing information produced
by IBM and Oracle with a grain of salt and would like to hear from real
DBAs, especially ones who are fluent with both products, for their views on
two questions:

1. What are the pros and cons of the current releases of DB2 and Oracle?

2. What other sources of *independent* information are available to help
someone new to databases choose between DB2 and Oracle?

This is *not* a troll and we don't want to start a flame war! Scott just
want some honest facts to help him decide which product is best at which
jobs.
Hi,

Point made.
without going into much religious talking, ask yourself:
Give neither counsel nor salt till you are asked for it.
How many OS versions of DB2 are on the market?
How many OS versions of Oracle?
How?
For DB2 you find different databases for quite every platform (OS 390,
UNIX, AIX, mainframe...) - name it. For every problem they have a
database - incompatible between each other...
In Oracle you deal with the same architecture on every OS platform
they support.
Are you positive about that?
Some of the things I like in Oracle
Do you wonder if you like in oracle?
* a lot of features to select from (Oracles index types i.e.)
* the shared sql approach
* multi-versioning and read consistency implementation (SELECT without
being blocked by writes i.e.)
Those found in their towards world understanding report.
yk
Oh ...
at least, all databases return the data that you store,

Why are you so positive?

--
Lady Chatterly

"Getting your ass kicked again I see. Lady C is quickly becomeing my
hero." -- Crawdad