Hi,
According to what i know about i5/OS, this should be feasible, but not
trivial.
There are two possibly ways I see:
a)
Create a new platform that i5/OS runs on. That would be PC based
Systems. This would feel like a CISC to RISC migration in terms of
headaches. I think quite a few components will need to be updated. The
end effect would be an i5/OS that can run on PCs. TIMI will enable
legacy programs to be retranslated and running on the PC platform.
IOPs are no longer necessary for i5/OS operation (though still widely
used - so a few adjustments would be necessary). This would probably
support a very limited set of hardware, and could be run barebones on a
PC, or within vmware. Essentially, it would be a full migration, and
systems like a next generation machine could use the x86 platform in
it's lower end (like SUN does with AMD machines, and SPARC in the higher
end).
This would be quite a lot of work, and would only make financial sense
if the lower end platform would get switched to x86.
b)
PearPC for example emulates PPC5 CPUs, for running Mac OS X. With a
large amount of work (still less then a)), one could change that
emulator to a Power5+ emulator, and then create appropriate virtual
devices for use by i5/OS, emulate IOPs for when they're necessary.
a) might happen if it decides that it wants to grow the market for the
System i by Offering more competitive prices. I seriously doubt that.
b) is something that System i evangelists with enough knowledge of the
Power5+ platform could start on today. Support from IBM for this
experiment would still be somewhat necessary for a smooth experience,
but it might be possible to hack this together without support from IBM.
-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of albartell
Sent: Monday, July 16, 2007 2:55 PM
To: 'Midrange Systems Technical Discussion'
Subject: i5OS running as a VM on PC
Hi All,
Just got done reading two stories on ITJungle.com:
http://www.itjungle.com/tfh/tfh071607-story03.html and
http://www.itjungle.com/tfh/tfh071607-story01.html. The two things
talked
about were a shortage of new i5 talent coming from educational
institutions
and market share that IBM is losing to Oracle - nothing surprisingly new
to
any of us.
In the end I think we are seeing these stats simply because it is fairly
difficult to get your hands dirty with an i5 and corresponding
technologies
vs. developing a business around PC based technologies, or more directly
stated: I can't download RPG and i5OS to my PC and start developing
applications. Again, nothing really new here - we have talked about
availability before. This leads me to the question/statement of this
thread.
Based on my limited knowledge, I am wondering what the complexities
would be
to "port" i5OS to something like a VMWare instance running on my PC. I
would only consider this potentially possible based on hearing about the
different hardware changes where statements like "pSeries has the same
guts
as iSeries" and "single level storage, hiding hardware implementation"
are
made here an there. In my mind if IBM could make i5OS more available to
the
general population, it would start catching on more, or rather, much
faster.
For the sake of argument let's leave timeshare OUT of the mix. We all
know
they exist and are relatively inexpensive, but they are very restrictive
in
letting you "mess around" with a machine to learn how it works (and
rightly
so).
To reiterate my question: What are the complexities to "port" i5OS to
something like a VMWare instance running on my PC?
Thoughts?
Aaron Bartell
http://mowyourlawn.com
midrange.com
