Re: FW: Children of the 1970's - Great Operating Systems

Paul,

Five years is a long time in this industry.
Five years ago Google was only five years old, and
five years before that it didn't exist.
Today Google's market capitalization is $184 billion
for a 10 year old company, vs IBM's $169 billion for a
120 year old company (the name IBM dates back to 1924
in the US, 1917 in Canada, but it's history goes back
to around 1888).

I see you mentioned Bill Zeitler's interview from
around 5 years ago where he stated i and p series
would not merge. You can be sure the plan for the
merger has been in the works for over five years.
That proves one thing - you can't always believe what
IBM executives state publicly.

And whatever became of that $500m that IBM was going
to commit to "our" system over the next two years -
five years ago? That certainly never happened. As
Frank said "Why bother to invest lots of money if the
platform is going to die anyway?"

The end isn't going to happen overnight. There will
be no sudden moves or dramatic announcements. Even
IBM isn't cruel enough (or stupid enough) to put a
bullet in it's head. It will be put out to pasture to
live out it's natural life, the same way it happened
with VSE & OS/2. The only "investment" they are
making in the product is to keep the current customer
base happy with something they can continue to run on
until we all eventually fade away.

Yes, a lot can change in five years. :-(

...Neil

--- Paul Tykodi <ptykodi@xxxxxxxxx> wrote:

Dear List,
Five years ago, Dr. Soltis and I had an e-mail
exchange about the future of "i". With the current
discussion about Neil's recent article, I thought
you might like to see another historical perspective
on where we might be going.
Best Regards,
/Paul

-----Original Message-----
From: Frank Soltis [mailto:fsoltis@xxxxxxxxxx]
Sent: Wednesday, January 22, 2003 9:06 AM
To: Paul Tykodi
Subject: Re: Children of the 1970's - Great

Operating Systems

Paul,

You hit it right on the head when you said:

Ultimately I believe the continued success of the

iSeries platform will be

driven by its relevance to the continuing business

needs of IBM customers

and not by hardware performance gains or

simplified software licensing.

With the new announcements this week and the

statement that it will commit

$500M to the platform over the next two years, IBM

is both making a

statement of support for the iSeries and taking a

big gamble. Why bother

to invest lots of money if the platform is going

to die anyway? The reasons

for making this investment in an integrated

platform are somewhat similar

to what happened in the 1970s.

During the 1970s only the biggest companies could

afford to hire people

with the skills needed to support a mainframe-type

of computer, such as

the S/370. This gave rise to the integrated

platforms from DEC, HP and IBM

that didn't need the specialized staffs. Midrange

customers fell in love with

these machines, because they were more accessible

and far more

understandable to the people running the business.

They also met the needs

of these businesses.

Integration was the order of the day. I remember

when DEC introduced their

"All-in-One" offering in the middle 1980s that did

both data processing

and office functions. IBM at the time said you

needed separate systems for

office and data processing. DEC integrated the

two, and their VAX systems

rapidly became the best selling midrange platform

as a result. It took the

introduction of the AS/400 to bring us back into

the game.

During this same time, IBM was also trying to sell

S/370s into midrange

companies with some fairly low-cost hardware

models. The 9370 model, for

example, came out in 1987. It was a great piece of

hardware for mid-sized

companies at a very competitive price. Yet it

failed in the marketplace.

Hardware cost was not a problem for many smaller

companies; it was the

support costs.

Then in the late 1980s came the PC servers, and we

all became computer

operators. Suddenly, anyone could manage a PC. Why

do we need the

integrated servers when PCs are so easy to manage?

This was the beginning

of the decline of the integrated midrange servers

from DEC, HP and to some

extent IBM. There was another factor, however,

that led to the decline,

and that was the rise of Unix.

Prior to 1990, Unix was only found on engineering

workstations. Not to be

outdone by the PCs, the Unix proponents argued

that their systems could

also be managed by anyone (a claim that is still

not true for most

businesses). They also argued that Unix systems

are higher performance

than PC systems, because they ran on RISC

hardware. Both DEC and HP decided

that it was cheaper to use Unix on their computers

than to continue to invest

in their own operating systems. It was the

customers who forced DEC and HP to

keep some level of investment in VMS and MPE,

otherwise both of these

operating systems would have died much sooner.

IBM, on the other hand,

decided to continue investing in both AIX and

OS/400. Today, LINUX is seen

by some inside and outside of IBM as the ultimate

operating system. Only

time will tell.

An interesting thing happened on the way to the

21st century. Customers

discovered that the proliferation of the PC and

Unix servers during the

1990s have created some very complex

"infrastructures" (a word hardly

anyone used in the 1990s). Managing all the

servers is complex and costly.

Does this sound like the 1970s? Maybe we need some

integrated servers?

Maybe they could even manage themselves?

IBM now says the next generation of server for

e-business on demand must

have four specific attributes. It must be

integrated, open, virtualized

and autonomic. Without getting into too much

detail, IBM has only two servers

that can even claim to be close to having these

attributes. They are the

zSeries and the iSeries. Not too many people

believe that the zSeries is

the best choice for the midrange, so that leaves

the iSeries.

IBM's commitment to the iSeries was demonstrated

by this week's

announcements. It is being positioned right in the

middle of IBM's move to

e-business on demand. The price reductions and the

continuing investments

that are being made is a pretty strong statement

that IBM believes this

integrated platform will be relevant for existing

and new customers in the

future. Once again, only time will tell if this is

true.

Frank
_______________________________

Frank G Soltis
IBM iSeries Chief Scientist
Rochester, Minnesota
e-mail: fsoltis@xxxxxxxxxx




                      "Paul Tykodi"
                      <paul@intermate-u       

To:       Frank

Soltis/Rochester/IBM@IBMUS
                      s.com>                   cc:

                                               Subject: 

Children of the

1970's - Great Operating Systems
                      01/21/2003 02:01
                      PM






Dear Dr. Soltis,

Recently with the iSeries division change in

general manager announcement

followed by the announcement of the new iSeries

hardware/software, I have

seen a number of discussions occurring on the

midrange-l mailing list

about the future of the iSeries server platform.

It happens that my experience

with IBM Midrange hosts dates back to 1986 and

includes the S/38, AS/400,

and iSeries. Initially I worked for Fidelity

Investments and was an end

user of IBM Midrange server technology until the

beginning of 1990. Since that

time I have been employed by a few of the third

party vendors that have

offered twinax compatible peripherals. In my case,

as time passed, I ended

up working for vendors that specialized in

printing solutions tailored to

the IBM AS/400 and iSeries environments.

The change from twinax to TCP/IP connectivity that

occurred on IBM

Midrange hosts in the late 1990's caused the

business model of my employer to

change in that the focus of the organization was

broadened from just IBM Midrange

hosts to pretty much any non-Windows or Novell

host that created output to

be printed (The Enterprise Printing niche). As

such, I began gaining

expertise in how Linux and Unix hosts handle

printing as well as how more

specialized hosts such as HP 3000 and HP (formerly

Dec) OpenVMS as well as

VAX systems handled printing tasks as well.

One of the things I noticed as I began gaining

expertise in multiple host

server platforms was that the loyalty of the end

user communities of these

platforms and the capabilities of certain other

host servers mirrored in

many ways the IBM Midrange capabilities and host

community loyalty quite

closely.

When you look at integrated host server platforms

today (hardware and

software sold as a bundle), you will find the

following:

HP (formerly Dec) VAX Servers - now dead. The

operating system lives on as

OpenVMS.

HP 3000 - now dying. New purchases are ended in

Oct. 2003 and end of life

is Dec. 31st, 2006.

 IBM iSeries - still alive but sales and earnings

are down dramatically.

What strikes me is that there must have been some

significant market

forces at work in the early 1970's when all these

systems (or earlier ancestors)

were developed that required a platform with

extremely good reliability

and superior integrated database functionality. I

am a little too young to

make any informed guesses about what these market

forces might have been since

I was still in grade school and high school during

much of the 1970's.

What I am wondering about today is whether new

market forces are now at

work that are changing the requirements of

computing so that an integrated

server such as the Dec VAX, HP 3000, or IBM

iSeries are no longer required to

meet the demands of customers going forward (I

don't claim to know the answer

at this moment).

I think the concern for many IBM Midrange server

customers is that they

know about what happened to both the Dec VAX and

HP 3000 systems. Prior to the

actual end of life announcements being made by

these other vendors many of

the strategies currently being undertaken by the

IBM iSeries management

were tried by these other vendors and were

ultimately unsuccessful. Since very

little on our planet occurs in a vacuum, I believe

I am seeing many

customers asking IBM to explain where they believe

the computing

requirements of the world are headed and why an

integrated server

(iSeries) is still needed in this new reality.

Recently I saw an e(logo) server interview with

Bill Zeitler in which he

was asked whether the i and p series servers would

merge in 2004 when AIX will

be able to run OS/400 in a partition. He said that

wouldn't happen because

IBM needed an integrated server offering in its

e(logo) server product

line. He did not explain why IBM thought this was

necessary though.

My purpose for sending you this e-mail is to say

that I personally think

the iSeries community places a great deal of trust

in you. I also believe that

for the iSeries to remain a viable product for IBM

in the next 5 years and

beyond, IBM will need to visit the deaths of some

of the competitive

products and explain why IBM believes that the

iSeries will remain useful

in the next 5 years while these other competitive

products were retired. If

IBM were interested in creating a webinar or other

type of promotional program

to talk about this issue, I think that you would

be one of the best people

currently involved with the iSeries platform to

lead such a program.

Ultimately I believe the continued success of the

iSeries platform will be

driven by its relevance to the continuing business

needs of IBM customers

and not by hardware performance gains or

simplified software licensing.

I'm not sure whether anyone has looked at where

the iSeries ancestors came

from to determine where the iSeries might be going

so I thought it was worth

writing you this e-mail to pass along this

insight.

Wishing you and IBM continued success with the

e(logo) servers and

especially our buddy the iSeries,

/Paul

 Paul Tykodi
Principal Consultant
TCS - Tykodi Consulting Services LLC

Neil Palmer, Cambridge, Ontario, Canada

(This account not monitored for personal mail,
remove the last two letters before @ for that)


__________________________________________________________________
Connect with friends from any web browser - no download required. Try the new Yahoo! Canada Messenger for the Web BETA at http://ca.messenger.yahoo.com/webmessengerpromo.php