Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.os.vms > #378789

Re: Gartner now recommending the mainframe as a possible option

Show all headers | View raw

In article <10thd9j$1qo65$1@dont-email.me>,
David Wade  <dave@g4ugm.invalid> wrote:
>On 07/05/2026 00:26, Dan Cross wrote:
>> [snip]
>> Telum II is a pretty impressive chip: the virtual cache thing is
>> neat.  And as you point out, it _can_ be cost-competitive with
>> x86 servers, but I bet our stuff at work would give it a run for
>> its money.
>> 
>> The problem with Z beyond that is two-fold:
>> 
>> First, you lose binary compatibility with x86; Linux works
>> because there is a port to Z, but if you need a Windows VM (or
>> VMS?), or need to host a binary-only application, you're out of
>> luck.  And there is still a lot of Windows out there running
>> under ESXi.
>
>of course, that is a big draw back..
>.. and if you have a lot of Windows perhaps hyper-v is a better match..

Likely.  Or something KVM-based if you don't want the vendor
lock-in that comes with Hyper-V.

>> Second, you get locked into IBM's system management/RAS
>> ecosystem.  Are you running Linux on bare metal? 
>
>as far as I know not possible on a modern z box..

Not surprising.  I know there's been work to make KVM work with
Z, but I if you can't run Linux on the bare metal, that doesn't
help you much for the top-level hypervisor.  I suppose you'd go
with z/VM, but then you're paying against IBM's historically
crazy mainframe software licensing terms.  I suppose that's good
for IBM.

>> In an LPAR?
>> Under VM?  IBM may have invented virtualization, but if you go
>> with their solution, you are entering their unique world; caveat
>> emptor.
>
>yes very different in many ways..

Oh yes.  That's the big unknown with going the mainframe route;
it's a real investment a complete technology sea change for a
lot of organizations.  If you've got mainframe expertise already
then I can see the attraction; if you don't, then you've got to
be _really committed_ to the direction.

>> A lot of customers are probably better served with something
>> like Oxide for an integrated solution, or Proxmox for an
>> open-source route.
>
>The big problem with Z for many is the size of the box. Its been 10 
>years since I worked but I really liked our VMware system build with 
>multiple mid-range 2-cpu so 16 core servers. With more 64-bit OSs these 
>days you probably want bigger servers..
>.. but VMWARE does give you the ability to sacale up or out...

That's a good point.  I bet they're betting on a lot of
consolidation; the workloads of a gaggle of small and mid-range
servers moving onto a single high-end machine.

An IBM z17 is not your father's mainframe; it doesn't need the
x-many thousands of square feet of floor space that a 3090 did.
IBM's promotional materials show them in office environments, in
a single rack enclosure.  But still, you're absolutely right
that you're starting at the high-end with one.  I suspect that
the numbers vis x86 really don't make sense until you're talking
about significant workloads.

For a lot of shops, it'd be much cheaper to go with a Linux/KVM
based solution, but they may not have the in-house expertise to
stand that up themselves.  Building out something with
functionality equivalent to ESXi is a heavy lift.

	- Dan C.

Back to comp.os.vms | Previous | Next — Previous in thread | Next in thread | Find similar

Thread

Gartner now recommending the mainframe as a possible option Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP> - 2026-05-05 18:30 +0000
  Re: Gartner now recommending the mainframe as a possible option Arne Vajhøj <arne@vajhoej.dk> - 2026-05-05 16:39 -0400
    Re: Gartner now recommending the mainframe as a possible option Lawrence D’Oliveiro <ldo@nz.invalid> - 2026-05-05 22:00 +0000
  Re: Gartner now recommending the mainframe as a possible option jgd@cix.co.uk (John Dallman) - 2026-05-06 18:58 +0100
    Re: Gartner now recommending the mainframe as a possible option Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP> - 2026-05-07 12:39 +0000
      Re: Gartner now recommending the mainframe as a possible option Dave Froble <davef@tsoft-inc.com> - 2026-05-07 16:32 -0400
        Re: Gartner now recommending the mainframe as a possible option Arne Vajhøj <arne@vajhoej.dk> - 2026-05-07 16:45 -0400
        Re: Gartner now recommending the mainframe as a possible option cross@spitfire.i.gajendra.net (Dan Cross) - 2026-05-08 10:44 +0000
        Re: Gartner now recommending the mainframe as a possible option jgd@cix.co.uk (John Dallman) - 2026-05-08 12:57 +0100
  Re: Gartner now recommending the mainframe as a possible option Alexander Schreiber <als@usenet.thangorodrim.de> - 2026-05-06 22:02 +0200
    Re: Gartner now recommending the mainframe as a possible option kludge@panix.com (Scott Dorsey) - 2026-05-06 17:09 -0400
      Re: Gartner now recommending the mainframe as a possible option cross@spitfire.i.gajendra.net (Dan Cross) - 2026-05-06 22:10 +0000
    Re: Gartner now recommending the mainframe as a possible option David Wade <dave@g4ugm.invalid> - 2026-05-06 23:15 +0200
      Re: Gartner now recommending the mainframe as a possible option cross@spitfire.i.gajendra.net (Dan Cross) - 2026-05-06 22:26 +0000
        Re: Gartner now recommending the mainframe as a possible option David Wade <dave@g4ugm.invalid> - 2026-05-07 09:00 +0200
          Re: Gartner now recommending the mainframe as a possible option cross@spitfire.i.gajendra.net (Dan Cross) - 2026-05-08 11:19 +0000
            Re: Gartner now recommending the mainframe as a possible option David Wade <dave@g4ugm.invalid> - 2026-05-08 15:01 +0200
              Re: Gartner now recommending the mainframe as a possible option Lawrence D’Oliveiro <ldo@nz.invalid> - 2026-05-08 21:12 +0000
              Re: Gartner now recommending the mainframe as a possible option cross@spitfire.i.gajendra.net (Dan Cross) - 2026-05-10 19:56 +0000

csiph-web