I have a theory....( Whither Informix?)

I have a theory....( Whither Informix?)

Post by mark.scran » Wed, 01 Nov 2006 10:02:55


ll -

After working "there" for 6 years and at Informix for 6 years,
traveling the world for 1.5 years of the IDS 10.0 launch, I have what I
believe is a plausible theory...but first, some comments....

Continued Development
-------------------------------
IDS 10.5 does have some great new features coming out as many saw at
conference. So truly if IBM didn't care about "something" with regards
to IDS (more in a minute), there would be no need for these features.
There are 3 different beta training classes going on for IDS 10.5 in
the next few months - one is already completed. So this is good. There
are some key players from the "original crew" left at IBM - John,
Johnathon, Madison, Guy.....this is important to all us.

Advertising
--------------
Can we please put this one to bed? That dead horse is really, really
starting to "reek"!

Those of us "out there"
-----------------------------
We are staying busy with a great deal of consulting and other related
work. One "of us" is looking at a very large, multi-year deal on yes,
Informix. I just left a site that is on Informix, and has been since
before SE...with no plans of moving to anyone else.

3rd Party Support
-----------------------
Yes - this is still ugly. Not sure what to do here. And no - I do not
expect IBM to pick up the slack here. If I'm wrong, I'm happy.

A Theory
-----------
Took me awhile to figure it out, but I think i'm onto
something...consider what percentage of IBM employees have made a
career off of, yes, DB2. DB2 on "big iron" is a huge $$$-maker as we
all know. So yes, even though someone said "a company doesn't care
about one specific product....", that couldn't be further from the
truth in the case of DB2 and IBM.

Now - what was the first effort made after the aquisition? To help
Informix users migrate to DB2. I know - I spoke at the first conference
and at 18 cities that year comparing IDS and DB2. The feeling was that
since I was so well connected to Informix clients that I could help
with the transition. As most of you know, my main message was "learn
about DB2 now so you can make an educated decision." BUT - during that
year and in the following years, many, many clients have not migrated
to DB2 but as you all know, to other players. So IBM had a choice -
stop the bleeding "somehow", or let the customers continue to walk
away. To help stop the bleeding - IDS 10.0 was delivered. AND we did a
World Tour. And there actually was some press and some advertisements
(yes - there was), which was great for all of us. But - never was IDS
positioned as "better" than DB2....multi-threaded or not. (Yes - the
base architectures are not even close....trust me). So what did 10.0
and continued development do for IDS? Kept the customers connected - to
IDS. That is good. But let's look at it differently...it also kept IDS
"afloat" and the customers excited, until...the next attempt to get
clients to .... you guessed it....DB2. That's my theory. I've seen
many, many slides on where DB2 is going....even one that said it will
be "multi-threaded" in 2008. One has to appreciate the irony of
that....I don't know any DB2/distributed customers that have asked for
multi-threading - they have chosen DB2 without this architectural
benefit. And the Informix folks already have the product they want.

So it's a theory.....and more food for fodder.....

Mark Scranton
Informix 1995-retirement



Obnoxio The Clown wrote:
 
 
 

I have a theory....( Whither Informix?)

Post by Double Ech » Wed, 01 Nov 2006 15:05:43

ark,

I always enjoy reading your comments.

I remain truly skeptical about the success or failure of Informix no matter
what you or anyone says about Informix. The so-called "customers continued
to walk away" theory that is supposed to have mitigated the need to upgrade
Informix always fascinates me. The idea that Informix or IBM are actually
in touch with their customers is on a par with learning what the Da Vinci
code was all about. It wasn't important to know, but the buzz surrounding
it made it seem like the Da Vinci code is important, the same for the churn
factor for Informix customers. Both the Da Vinci code and Informix seem to
be more about mythology than fact, either for the good things or the bad
things. What will be helpful for the future will be some real factual
information about Informix 10.x, what it is, what it can do, and how to get
developers interested in it enough to want to put applications on it. That
will involve more than just word-of-mouth here in c.d.i.. INFORMIX NEEDS
NEW CUSTOMERS. Nothing wrong with expanding the base. Get it out there,
then build some depth of applications. It's worthless if nobody is using
it.

Regards,

-DE-

XXXX@XXXXX.COM wrote:


 
 
 

I have a theory....( Whither Informix?)

Post by Serge Riel » Wed, 01 Nov 2006 22:34:57


DB2 for LUW has been running multi-threaded on Windows since its
inception on that platform (DB2 V2 in 1993 I believe).
Given that the "process-model" is encapsulated we can flip between
threaded and multi-process "at will" (meaning at a moderate cost).

No customer ever asks for infrastructure features. That's why they are
infrastructure. I doubt Informix introduced threading due to customer
demand.

W.r.t getting users of IDS onto DB2
I stood up at the IIUG meeting in Anaheim and I will do it again here:

There is no such thing as a project to make DB2 and IDS the same.
All DB2 for LUW SQL features current and planned cross my cluttered
desk. I have no SQL item (that's where the *** hits the road for an
app after all) active under development or planned which purpose it is
to increase compatibility with IDS.
No case sensitive table names, no double quoted strings, no SERIAL
column, no ::-casting, SPL, etc, etc.... in DB2.

Cheers
Serge
--
Serge Rielau
DB2 Solutions Development
IBM Toronto Lab

WAIUG Conference
http://www.yqcomputer.com/
 
 
 

I have a theory....( Whither Informix?)

Post by Obnoxio Th » Wed, 01 Nov 2006 22:40:21


Serge Rielau said:

OK, you can sit down now, Serge. We all know what you look like.

--
Bye now,
Obnoxio

"... no bill is required as no value was provided."
-- Christine Normile

--
This message has been scanned for viruses and
dangerous content by OpenProtect( http://www.yqcomputer.com/ ), and is
believed to be clean.
 
 
 

I have a theory....( Whither Informix?)

Post by Double Ech » Wed, 01 Nov 2006 23:03:59

Serge,

It's simple to choose DB2 or IDS.

DB2 is for customers who don't really know a good database from a bad one
and want something besides SQL-Server. They probably like Oracle. They
probably run apps and just want a good nuts-n-bolts database that is on
the enterprise-db list. They couldn't tell the difference between a schema
and a good mystery novel. DB2 runs good, has chimpanzee-simple HA, and
is pretty reliable. It's got IBM's name on it, and can do a lot of things
without threading. It looks like a distant cousin of flat files. Despite
DB2's bland construction it has a lot of good features, and is a nice choice
if you have to choose between SQL-Server, Oracle, or Sybase.

IDS is for customers who really do know a good database from a band one
and want something besides Oracle or SQL-Server or DB2 or Sybase. IDS is for
customers that really do things with data, and don't depend on more and
more powerpoint slides or future hope from the vendor. IDS works great
today, not tomorrow, and has the most flexibility and options of any
database to choose from. IDS not only has multi-threading it also has
virtual processors, something no other database has, and amazingly enough
is still one of the best ideas in database engines.

-DE-
 
 
 

I have a theory....( Whither Informix?)

Post by Serge Riel » Wed, 01 Nov 2006 23:36:12


DE,

So what is XPS all about? Could it be there is some (multi billion $
market) IDS does not cover? Is it inconceivable that in this space DB2
has something to offer to companies like VISA?

Cheers
Serge
--
Serge Rielau
DB2 Solutions Development
IBM Toronto Lab

WAIUG Conference
http://www.yqcomputer.com/
 
 
 

I have a theory....( Whither Informix?)

Post by Double Ech » Wed, 01 Nov 2006 23:49:36


The problem for XPS is more about marketing than features. Since XPS is
practically unknown it's suffering the same fate as IDS. DB2 at the
clustered level works basically the same as XPS. I've installed both from
scratch, done the DBA work for both XPS and DB2 and XPS is still my preference.
Monitoring DB2 is painful, and difficult to write scripts for, with many of
the monitoring commands split up between different user ids. Very confusing.

If there is "something to offer" companies, such as either offering XPS or DB2,
there should be enough material available to a consultant or sales professional
to demonstrate the features of both, and let the customer decide. Both are
good shared-nothing products. I do like the config for DB2, it is well thought
out for clustering, but the poor monitoring features of DB2 and its lack of
multi-threading leaves me with concern that it won't be able to handle big
big data. I'm not saying those are real problems, but they _are_ real
perceptions. Informix professionals should be proud of the products they
represent, and there should be some real marketing of those products. XPS has
a great history and is a great product, but until it too has someone blowing
its horn it will end up in the same boneyard of great technology that didn't
make it in the market. It deserves a decent chance to compete, but I don't
see that happening anytime soon.
 
 
 

I have a theory....( Whither Informix?)

Post by mark.scran » Thu, 02 Nov 2006 01:18:58

DB2 multi-threaded on Windows"? The answer is in the question. This
one made me chuckle. But this is good for all the Z/OS customers
migrating to Windows, right? (chuckle again).

XPS and DB2? I did the session in Charlotte many years ago when two
very large banks merged, and one had XPS and the other DB2 EEE. One
side of the room had the XPS'ers, the other, DB2'ers. I compared and
contrasted the 2 architectures and feature sets. The new, "larger" bank
chose XPS.

A large company in Phoenix was pushed down the DB2 road a couple of
years ago. Let's just say that they will have a 28TB engine on XPS by
the end of the year.

Large bank in England was once told the XPS 8.51 was the last release.
He said "I want to buy more XPS....regardless."

And "customers don't ask for [multi-threaded]...." - dead wrong there.
The "line-up" of IDS customers at the "first Anaheim" (after the
acquisition) before the panel of execs on the stage....Dan M., a
customer, asked specifically about multi-threaded versus not...Don T.
deflected the question with "who cares whether it's multi-threaded at
the end of the day..." - all the Informix customers groaned. That's
because they DO know their architectures (NOT "infrastructures"), and
understand the technical value of an architecture that will scale. Why
do they know this? Cause our Informix classes over the years, from Sys
Admin down to the Internal Architecture class thoroughly covered the
concept of multi-threaded, and how Informix implemented it back in the
early 90's in v7. And why do the Informix folks know so much about
"process-based" versus "thread-based"? Because of Informix Online
version 5, and it's process-based, single-threaded architecture. When
the Informix developers found v5 would not scale (due to an
ever-expanding process footprint as more and more users connected as is
true in process-based architectures (without a connection concentrator
that is)), they implemented true multi-threading. (Most remember that
the first release, v6, was "interesting", but v7 cleaned up stuff).

XPS is actually a "much sadder" story than IDS in my opinion. I've seen
the power of XPS, and still believe it is the best product ever
released by Informix. Everyone dings XPS saying "nobody has it"...well
- it's market was never meant to be as big as IDS, for the following
reasons: 1) it's target was very large DW environments; 2) it's price
point was much, much higher; 3) it's implementation was very
hardware-driven, and sometimes takes months to implement (and
rightfully so). 4) Early releases did not have OLTP features - it was
meant purely for DW. 8.31 brought many OLTP features, but the market
was still DW.

OK - that's enough for now.

Thanks for listening -
Mark Scranton
Informix 1995 - 1st great-grandchild




Double Echo wrote:

 
 
 

I have a theory....( Whither Informix?)

Post by DA Morga » Thu, 02 Nov 2006 01:56:37


And that is precisely what those of us who get to meet and greet
the salesforce are seeing.

How many people from IBM are knocking on doors selling DB2? How
many Informix? The number tell the tale.
--
Daniel A. Morgan
University of Washington
XXXX@XXXXX.COM
(replace x with u to respond)
 
 
 

I have a theory....( Whither Informix?)

Post by Double Ech » Thu, 02 Nov 2006 02:16:07

ne little note to attach to what you're saying.

Informix _had_ been very restricted in dishing out XPS, somewhat of
an attitude that it was only reserved for the elite. I think the
price tag the last time I priced it was $50,000/processor, but the
price is less of an issue than getting it in the hands of Informix
people so they can play with it, test it, and implement lab environments
where the customer can see and feel XPS. Again, if nobody is using it
then it's basically worthless. There has to be a sea change at IBM
to get these products as available as DB2. I downloaded DB2 EEE,
implemented basically an XPS cluster in DB2 EEE, and tested it for a year
before the customer bought it. But they bought it because they saw
the power of shared-nothing. I could have probably begged IBM to get
me a copy of XPS and do the same thing with XPS, but I wasn't sure if
Informix was going to be around.

In a lot of cases
these kinds of projects have to be done below radar just to get a
proof of concept--not so much for management, but for the guy who
will present the POC results. The DBA should be able to have some playtime
with these products so they don't stand there like a dumbass when all
the Microsoft weenies and others argue against the product. There
is nothing like real, in-house benchmarks to back up a reason for
buying something like XPS. So in short, IBM needs to get XPS out
there for download. You can configure it basically the same as DB2
EEE, only better. Virtual and physical partitioning, it rocks.


XXXX@XXXXX.COM wrote: