[Info-vax] OT: Arun Kishan

[JF Mezei]

Neil Rieck wrote:
> environment). Meanwhile, just like the story of the "Tortoise and the
> Hare", the x86-64 architecture (under evolutionary pressure from
> AMD64) as been catching up to, and may be very close to surpassing,
> Itanium.

Isn't IA64 already behind the 8086 ?   It may catch up/surpass the 8086
when Tukwilla is released, but I suspect it will be eclipsed again with
the next 8086.


Lets not forget that for a company such as HP, it isn't just the chip
that is different, but also the systems.


With Windows gaining ability to support more CPUs, I suspect we will
start to see larger "industry standard" systems being built.


As 8086 SYSTEMS scale up, it reduces the range of IA64 systems that
offer unique capabilities. Eventually, the costs of maintaining a
totally separate low volume architecture when the 8086 already covers
almost all of its range of systems will force HP to decided to
standardize on the 8086.

> HP management will reassure you by claiming that OpenVMS can run in a
> virtualized mode hosted by HP-UX but does that OS run on x86-64?
> Perhaps both OSs will need to run from PC emulators :-)

You'll have a 8086  Windows mainframe running an IA64 mainframe emulator
running HP-UX, running VMS :-)

> Alternatively, HP management could make use of the lower labor costs
> at HP-India to quietly begin an OpenVMS port to x86-64.

HP could have done that any time it wanted and make us of the highly
capable team it had in the USA who knew VMS inside out.

If HP ports HP-UX to the 8086,  I suspect we will see something like
Apple's Rosetta which can run Pa-Risc and IA64 binaries on the 8086.
(but only for HP-UX applications).

In terms of emulating VMS, I am not sure this is the route they will
take. Sure, there are 3rd parties that already provide VAX/Alpha
emulation.  And there may be an IA64 emulator.

I think the focus will be on porting middleware to another OS to make
migration from VMS possible/easier. I imagine that OMX has had high
level talks with Livermore/Hurd about the real future of VMS and they
would know exactly what componments would be ported to what platform to
make the porting of their software possible.

Of the large remaining customers, I suspect HP would now have good
feedback on what features/middeware of VMS makes it hard for them to
port to another platform.

Say RTR is absolutely necessary for OMX. And HP ports RTR to work on
Linux. Then it pretty much assures that OMX exchanges will remain HP
customers since they will need to buy RTR from HP nop matter what
platform the exchange software runs on.