Ken Gunderson wrote:
On Sun, 2010-02-14 at 13:27 -0500, Dennis Clarke wrote:
Preferentially, I think a reasonable thing for Oracle to do with Solaris
is the following:
(1) Quit giving away Solaris 10. Instead, provide several different
Support Contract levels for Solaris 10, with a very basic one providing
/solely/ security patches for some nominal fee (<$100/yr/server). Other
gradiations as desired, of course.
Sun.COM types just don't get it - Sun _had_ to start giving away Solaris
10 because outside of Suna and old silver back Unix geeks in the
financial service sector datacenters Solaris was becoming increasingly
irrelevant. Those who didn't grow up on Unix in the 80's grew up on MS
in the 90's and they're the ones running the show in much of the
business world now. Some of this latter group discovered Linux and the
*BSD's when they were in college. A subset of those started up what
became the largest sites on the Internet; Yahoo (FreeBSD) and Google
(Linux) using stacks of white box pc grade hardware. Later some big iron
did find it's way into those shops but those aren't the stories you hear
about.
And, from a software-vendor standpoint, does anyone make money on
Yahoo/Google/Hotmail/etc. using Linux or *BSD? Oops. Nope. They're a
huge boon to the end-user, but neither of them helped any outside vendor
to any extent. At best, you get a "reference customer" to say that X is
using a product of yours, but I hardly think that's Solaris' (or AIX, or
HPUX) problem at this point.
In order to regain much lost traction Solaris NEEDS mindshare. Where
there are LOTS of other *nices available for free you'd better make
your's away for free as well if you want to have a chance in hell in the
competition for that mind share. It's like when Bill Gates tried to
ignore the reality of the Internet and write it off as a "passing fad"
because it was competing with internal proprietary network protocols
that he was hoping would sway the day. Well.. that didn't work so he
was smart enough to read the hand writing on the wall.
I agree - giving away Solaris 10 was a great way to gain market traction
again.
That said, and seeing how Oracle seems to have a more
product/revenue-stream focus, I think the "try-it-to-get-hooked" version
is now OpenSolaris, NOT Solaris 10.
I like to run the analogy between RedHat Enterprise Linux and the Fedora
project, as far as support goes. You can't get support for Fedora from
RedHat, but it's a fabulous way to get intro'd to the "RedHat"-way of
doing things. For commercial use, you go buy RHEL.
What I'm saying is that I think the most likely scenario is that Solaris
10 remains a commercially-supported and distributed product with likely
no unpaid/free support or distribution, while a freely-available but
unsupported OpenSolaris provides the introductory hook for the whole
set. Because, let's face it, introductory stuff had better have the
nicest, newest, neatest things in it to get people to try it. And that's
OpenSolaris, NOT Solaris 10.
Here I agree. The Sparc systems can have an RTU on the hostid again. Just
like the old days. I have no idea how you would track x86 systems given
that the iso images are in the wild and you just can not stop people from
passing torrents.
The old days are gone. We can't return. Solaris needs to move forward
or die. x86 is here to stay. Embrace that and use it as a pathway to
big iron SPARC machines for businesses that grow to need them.
There is no indication that x64 is being abandoned, or even slighted. I
think his comment was solely about the inability to hardware-tie an RTU
to a x64 system, as is possible in a SPARC system. I don't see this as
a big issue - support contract numbers have always had a bit of "fudge"
and "honor-system" leeway in them, and I'd really not like to see this
try to be eliminated.
(2) Continue to do (most) development work out in the open in
OpenSolaris, and provide FREE access to everything in the OpenSolaris
repos. Use this as the "first-one's-free" hook to get people
introduced to Solaris as an OS. And, of course, get all of us to do
beta-testing for it. :-) Honestly, I think it's entirely reasonable
for Oracle to declare that There Shall Be No Support Contract for
OpenSolaris - it's a development platform, and I think efforts are
better spent in moving along the development effort as a whole than
having to dedicate some folks to support services.
I have a problem with software where there is no support contract of any
kind. There are too many IT environments that will simply not accept
software which does not have a paper trial and a support contract. That is
still firm policy in some places regardless of the noises made by the
masses with their hands out.
Uhm... apparently big corporate marketing types didn't notice but we're
in the midst of the worst depression we've seen since the BIG
Depression? Moreover, I don't see it as a matter of standing with my
hand out. I see it as objectively evaluating the options. On one side
we have a LOT of freely available *nix like offerings, a subset of which
offers commercial support. So I can hone expertise in one such OS and
opt for commercial support as suits individual client needs.
Once again, see above about the a good way to get mindshare and revenue
at the same time, using different policies for OpenSolaris vs Solaris 10.
In an ideal world, maybe Solaris 10 would remain free with free security
patches. Then again, maybe the lost revenue stream from that model
doesn't make up for the "mindshare" gain. I can't judge (and to my
knowledge, everyone else is just guessing, too).
In my opinion Sun was on the right path towards regaining lost mindshare
with 1) porting Solaris to x86, 2) making it free to use, and 3)
providing security patches. If they'd only done it 5 years sooner then
I suspect they'd still be Sun and not Oracle.
Nah. I've used Solaris x86 since way back (it was available since at
least 1995 in Solaris 2.4, IIRC), and frankly, thinking that Linux would
explode the way it did in the late 90s is ex-post-facto logic - it was
plainly on the road to be a player, but its complete explosion was due
to a lot of different harmonious factors, and not really predictable.
Solaris 10 hit just at the right time (introduction of
mass-consumer-level 64-bit computing with the Opteron) and the two fed
off eachother. You can argue that maybe right then was when it would
have been smart to open the development project (that is, start
OpenSolaris 5 years earlier), but making Solaris 9 or before available
openly really is just wild-ass speculation as to efficacy.
And, frankly, Sun being bought had everything to do with things other
than the Solaris ecosystem.
Give Solaris 10, 11, etc. away for free. Provide free security patches.
The important thing is to get Solaris back out there. Make your money
elsewhere.
We'll just have to agree to disagree.
We really do need to realize that this is 2010. Not 1994. There are vast
talented organizations that have a business objective to crack and hack
and attack networks and information access points. Internally also. All
operating systems today and forever in the future must give serious
thought to security and quality engineering. That can not be done without
an established revenue stream. Simply put, any business minded individual
in a customer IT division would ( and should ) look away from software
which does not have a support contract. The absence of that support and
revenue stream is a clear indication of lack of quality. Right or wrong,
true or false, people make decisions on purchases and IT policy with
arguments like this. I am sure you have experienced the "real world" and
it is very far from the ivory tower. It it simply full of politics,
baseless opinion and fighting middle management attempting to establish
their own world view within some corporation somewhere. Its amazing to me
that some places ( half of Fortune 100 and ALL of government agencies )
create a product or service and can function at all.
+100! (Had to clone a bunch of me years ago to keep up with the ever
increasing work load...;)
Once again, I think it's fallacious to argue that a piece of software in
ACTIVE DEVELOPMENT should be used in production. You're seriously
negligent as an IT professional to do so. OpenSolaris is in active
development. No support contract is going to change that. I'm simply
pointing out that any (relatively small) revenue stream that Oracle
/might/ make from selling OpenSolaris support contract will hurt the
timetable for delivery of the PRODUCTION-level Solaris Next.
If you are interested in OpenSolaris for production use, then you'd be
also interested in Fedora 12, Debian Squeeze, or similar products. None
of which have commercial support available from their developer
organization(s).
Sorry for the digression but the point I am trying to make is that
software without a support contract is simply unacceptable and the RFP
gets pushed onto the floor before you get past the table of contents.
That is the "real world". Want a good product with a future? Ensure it
makes money as its first feature and everything else is secondary.
Which is exactly my point with OpenSolaris. It's /not/ production-ready
yet for the vast majority of users, so naturally, why on earth would you
sell a support contract with it? Production-ready code is in Solaris
10, which will likely /require/ a service contract for support. QED.
--
Erik Trimble
Java System Support
Mailstop: usca22-123
Phone: x17195
Santa Clara, CA
_______________________________________________
opensolaris-discuss mailing list
[email protected]
