21 Feb
2012
21 Feb
'12
4:02 p.m.
-----Original Message-----
From: cctalk-bounces at classiccmp.org
[mailto:cctalk-bounces at classiccmp.org] On Behalf Of Dave McGuire
Sent: 21 February 2012 00:54
To: General Discussion: On-Topic and Off-Topic Posts
Subject: Re: What is a mainframe? (was Re: NASA unplugs last
mainframe)
On 02/20/2012 06:24 PM, Dave wrote:
I think it was
Sadly no. I am pretty sure I could have had the one from work when we binned
it but I didn't take it as I had concerns about keeping it working given the
fun the IBM guy had with it when it went wrong, and the fact the getting a
licence for VM would have been fun and expensive....
-----Original Message-----
At Newcastle there were lots of problems with their IBM
370 eventually tracked down to some circuits with
faulty
bonding of the copper
to the pcbs in some units. Recently when we had
an IBM
MP3000 which some
would call a mainframe (It runs 370/xa and ESA
code) it had a lower
availability that our Windows servers.
That's scary.
I have an MP3000 here that I'm trying to get running. Do
you know a
bunch about them?
I'd definitely call it a mainframe; it uses the
same processor
chipset as an S/390 G5. (for others reading this, ZERO relation to
PowerPC G5)
Oh yes it's a real monster of a beast, I think perhaps it should be called a
"super mainframe" , "Mainframe Complex" or "mainframe
cluster" because what
you get is a box that can run several independent copies of operating
systems. Looking at some articles on the i/o re-enforces this view. In fact
I don't think it would make sense to run a single zOS image on one of these
modern boxes as you probably can't get it any where near flat out.
To me a Mainframe was largely about the i/o, so
perhaps
eparate, dedicated ,
programmable, identifiable i/o processor? I am
note sure if
the z9 has
these, or if like the MP3000 the i/o processors
are
"emulated" by a service
processor...
The z9 has massive, massive I/O capacity that is most
definitely NOT
emulated.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA