I agree with Colin in that the decline in mainframe usage is partly the
"herd wisdom" picked up from the seatback airline magazines starting
around 1990 +/-. The other part is that IBM has, once again, shot
itself in the foot with software. There is a mindset in IBM that goes
back, in my memory, to the bi-annual attempts to stabilize DOS and then
DOS/VS to the early 1970's and convert everyone to "big-OZ". That same
mindset continued with SNA. SNA was and is a great architecture, but it
wasn't and isn't the only way to communicate over a network. Couple
that with the insistence that there there had to be an OS presence
involved with the solution. First it was VSE/VTAM or VS1/VTAM VM/VTAM
was shoved aside even tho there was a VM native solution. Now we are
left with an MVS/XA remnant in GCS. Look, also, at IBM's mis-steps with
OS/2. What a great operating system compared with MS Windows in the
late '80's and thru the '90's! If development had continued, one can
only imagine what it would be today.
We too are on a "get rid of the mainframe" path, altho it has been going
on for at least 10 years and will probably continue after my retirement
which most likely will be in a couple of years.
Jim
Colin Allinson wrote:
I think it is not just beginning but is well under way in most places.
Part of this is a sort of 'religious belief' in the conventional wisdom
that anything that is not a mainframe will be cheaper (more cost
effective) and will run modern systems better. Any arguments to the
contrary, no matter how well supported by facts, are immediately dismissed
as being presented by those stuck in the past with a vested interest in
retaining mainframes. It seems to be a case of 'everybody else is doing it
so it must be right - don't confuse me with the facts'.
Here, like many places, there is a definite plan to eliminate mainframes
from the organisation at almost any cost. Luckily for me it will take
longer than most places, because of heavy reliance on TPF for our core
business, so I should see it to retirement.
Having said all this I do agree with what Steve implies, IBM are not
entirely blameless in this decline.
1. When the Z/series with IFL's was announced a subtle change in the
emphasis would have made a huge difference. If this had been announced as
primarily a shared LINUX server that had the additional benefit of
**also** running legacy mainframe code then there is a chance it would not
have been tarred with the mainframe brush.
2. Software pricing has long been the bane of mainframe economics
(with some justification). It is perfectly true that development and
support costs must be reclaimed but the argument for maintaining high
software costs for products that are now stabilised / out of support is
much harder to justify. We are currently going through a software cost
reduction exercise and are looking at the most expensive products first.
Some that are no longer current fall into this category so we have found
free or reduced cost alternatives. If IBM had reduced the monthly cost for
outdated products to a more reasonable figure then we would probably not
have even questioned them - as it is we eliminate the product and both we
and IBM lose.
Colin Allinson
(speaking for myself)
--
Jim Bohnsack
Cornell University
(607) 255-1760
[EMAIL PROTECTED]