[AUUG-Talk]: Proprietary Unixes (Dead?)
enno.davids at metva.com.au
Thu Oct 4 11:25:11 EST 2007
On Wed, Oct 03, 2007 at 11:48:55PM +0930, David Lloyd wrote:
|Are the Unixes that are proprietary dead or dying:
|* AIX - IBM tout Linux these days
|* HP-UX - HP tout Linux / Windows these days
|* Solaris - Is moving towards going Open Source
|* IRIX - is dead?
|* DEC UNIX - is dead?
|* SCO's offerings - is dead?
|Apart from Sun's Solaris, do any of the proprietary Unixes have enough
|backing to keep them going, in terms of market share? Do they have
|enough capital to remain competitive with each other, with Microsoft and
|with the open source offerings?
|I tender they don't. I could be wrong...
For all their strengths, the various open source offerings have little
to offer in the way of innovation. Their main thrust is low TCO beige
boxes and even there the sums only tend to come out if you can deploy
them and abandon them in place. (None of the open source crowd has a
binary patch regime that is able to respond to security threats or bugs
promptly or is worth a pinch of shit in any commercial support sense.
And yes I've had Red Hat's paid for support and lets not go there lest
I become frank and open...)
The commercial UNIX vendors are where much of the innovation is coming
from (and to be fair, given that they have real development budgets that
can hardly surprise anyone). And you don't even need to go far to see it.
Sun's LDOMs and containers (aka zones) and Dtrace and ZFS and for that
matter the Niagara chip. IBMs subcpu LPARs and LVM (which they ported
to Linux I believe?). And lets face it the best GUI and graphics on the
market are in Apple's OSX, arguably the most widely deployed commercial
UNIX (or is that UNIX-derived?) OS there is now? (I'm not offering HPUX
examples mostly as I haven't been near their kit in a decade...)
Don't get me wrong, I'm not about to trade my FreeBSD desktops here at
home. Open source has made the computing world a much better place IMHO
and was already doing that back when net.sources and comp.sources.*
were where we got our fix. But will Linux be pushing the E25k's off the
datacentre floor any time soon. No. Can you configure me a box with >100
cores on an open source platform today? Can you split it in two or three
for now and re-arrange its resource allocation in the event of a disaster?
And you haven't even scratched the surface of replicated SANs, backup,
odd "only a few in the world" interfaces or any of the other things where
open source competes poorly if at all.
Sure you can build a LAMP server without breaking a sweat, you may be
able to cut over your word processing to Open Office (and we'll note Sun
had a big hand in getting that to where it is today). In general, anything
that a few guys can prototype at home Linux can be made to excel at. But
anything that costs real money to build is a non-starter.
So proprietary UNIX dead? Not even close.
More information about the Talk