history is hard

[Dave Wade]

> -----Original Message-----
> From: cctalk <cctalk-bounces at classiccmp.org> On Behalf Of Peter Coghlan
> via cctalk
> Sent: 25 May 2020 11:24
> To: 'General Discussion: On-Topic and Off-Topic Posts'
> <cctalk at classiccmp.org>
> Subject: RE: history is hard
> 
> On Mon, 25 May 2020 at 09:49:08 +0100, Dave Wade wrote:
> 
> >
> > In these systems the end user operating system, CMS, would also IPL
> > (load) on bare hardware.
> >
> 
> I wouldn't regard CMS as being the definitive end user operating system.
> Often, the reason for having VM is not to support multiple instances of CMS.
> The end user operating system can be anything that can be ipled in a virtual
> machine which is pretty much anything that can be ipled on the real
> hardware (except for CMS of course which can be an end user operating
> system without complying with this rule).

No, when they had VM/ESA in Stockport Council it was mostly for VM/ESA with CICS and CMS just for management and config.
Before that in NERC it was all interactive workload.


> 
> >
> > I would disagree with Peter over the "same under the covers".
> >
> 
> Yes, that was the wrong way to put it.  I was trying to say that these variously
> named entities produced over many decades all do pretty much the same
> thing and have a common ancestry.
> 
> Perhaps something like they "deliver the same functionality" would have
> been better?
> 

Yes, and of course whilst in some ways the whilst the change to SIE was a step change in the way CP delivers VMs in other ways it was a just one point on the evolution of VM.
It started with the ECPS assists (I think Extended Control Program Services) which had more features added over time so the Hardware handled more tasks,  but VM could cope if they were not present.
Then as things switched from 24 bit to 31 bit hardware and XA and ESA architecture appeared CP had to have SIE to work, and the type of VM it could deliver depended on what SIE on the box could support.   
Over time SIE itself has evolved to offload more work from CP (or HCP as its now called) and the architecture has moved from 31 bit to 64 bit. 
On the other hand I expect some parts of CP will be traceable back to VM/370. For example the spool which allows virtual card decks and print files to be stored and manipulated is still similar.
CP still has to manage control blocks for each VM that exists, its just that it uses SIE to execute them. It still manages a directory of users and passwords and controls who can logon.
It has to keep a track of which real devices map to which virtual devices of each VM.

Dave

> >
> > > Regards,
> > > Peter Coghlan.
> > >
> >
> > Dave
> >
> 
> Regards,
> Peter Coghlan.
> 
> > > >
> > > > bill
> > > >