Linux Advocates Fear Solaris 10.
Mike Cox
Linux became popular was that they were not as conservative as BSD, and
would add features constantly. That meant if you wanted some enterprise
functionality for free, you'd have to go with linux (SMP, etc.).
The downside of that linux philosophy was that backwards compatiblitiy went
down the drain. glibc interfaces and kernel calls would change
significantly between versions. Sometimes the "enterprise" feature would
be botched, like the scheduler in the 2.4 kernel.
Now that Solaris 10 is going to be open sourced, Linux advocates rightly
fear that Linux will be abandoned in favor of Solaris 10. Similar to how
BSD was left in droves for Linux. That's right, SUN's brilliant strategy of
making Solaris open sourced is using the linux advocates' techniques
against them.
The reason people will move to Solaris in droves is the same that caused
BSD's user flight. It supports more enterprise features. But the
difference this time is that Solaris is cheaper than Redhat, is completely
backwards compatible, and scales much better on many more CPUs.
Now if only SUN would release its compiler under a BSD license....
John Bailo
> Now that Solaris 10 is going to be open sourced, Linux advocates rightly
> fear that Linux will be abandoned in favor of Solaris 10. Similar to how
> BSD was left in droves for Linux. That's right, SUN's brilliant strategy
> of making Solaris open sourced is using the linux advocates' techniques
> against them.
Having used Solaris.
I doubt it.
Peter Köhlmann
< snip typical M Cox idiocy >
Well, you know, it is called "Slowaris" for a reason
And for that same reason you wish it would replace linux. Then Win would not
stand out in sharp contrast as a buggy, slow POS
--
Microsoft software doesn't get released - it escapes, leaving
a trail of destruction behind it.
Nathan Dietsch
Peter Köhlmann wrote:
> begin Mike Cox wrote:
>
> < snip typical M Cox idiocy >
>
> Well, you know, it is called "Slowaris" for a reason
The OP is a troll, don't feed him. Don't belittle yourself by stooping
to his level.
>
> And for that same reason you wish it would replace linux. Then Win would not
> stand out in sharp contrast as a buggy, slow POS
This sentence makes no sense.
MfG,
Nathan Dietsch
Joerg Schilling
Nathan Dietsch <use...@ndietsch.com> wrote:
>Peter,
>
>Peter Köhlmann wrote:
>> begin Mike Cox wrote:
>>
>> < snip typical M Cox idiocy >
>>
>> Well, you know, it is called "Slowaris" for a reason
>
>The OP is a troll, don't feed him. Don't belittle yourself by stooping
>to his level.
Peter Köhlmann also seems to be a troll (as he uses the word
"Slowlaris").
___________________
/| /| | |
||__|| | Please do |
/ O O\__ NOT |
/ \ feed the |
/ \ \ trolls |
/ _ \ \ ______________|
/ |\____\ \ ||
/ | | | |\____/ ||
/ \|_|_|/ \ __||
/ / \ |____| ||
/ | | /| | --|
| | |// |____ --|
* _ | |_|_|_| | \-/
*-- _--\ _ \ // |
/ _ \\ _ // | /
* / \_ /- | - | |
* ___ c_c_c_C/ \C_c_c_c____________
--
EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin
j...@cs.tu-berlin.de (uni) If you don't have iso-8859-1
schi...@fokus.fraunhofer.de (work) chars I am J"org Schilling
URL: http://www.fokus.fraunhofer.de/usr/schilling ftp://ftp.berlios.de/pub/schily
Joerg Schilling
>> begin Mike Cox wrote:
>>
>> < snip typical M Cox idiocy >
>>
>> Well, you know, it is called "Slowaris" for a reason
>
>The OP is a troll, don't feed him. Don't belittle yourself by stooping
>to his level.
Peter Köhlmann uses the word "Slowaris" so he is a troll too :-(
Eltee
> That is why they post so much FUD on sites such as slashdot.org. The reason
> Linux became popular was that they were not as conservative as BSD, and
> would add features constantly. That meant if you wanted some enterprise
> functionality for free, you'd have to go with linux (SMP, etc.).
>
> The downside of that linux philosophy was that backwards compatiblitiy went
> down the drain. glibc interfaces and kernel calls would change
> significantly between versions. Sometimes the "enterprise" feature would
> be botched, like the scheduler in the 2.4 kernel.
>
> Now that Solaris 10 is going to be open sourced, Linux advocates rightly
> fear that Linux will be abandoned in favor of Solaris 10.
On the contrary. My guess is that Linux will adopt the best features of
Solaris and become an even better OS.
Rich Teer
> The reason people will move to Solaris in droves is the same that caused
> BSD's user flight. It supports more enterprise features. But the
> difference this time is that Solaris is cheaper than Redhat, is completely
> backwards compatible, and scales much better on many more CPUs.
I know that MIke Cox is a troll, but I think this is the first sensible
thing he's ever written!
--
Rich Teer, SCNA, SCSA, author of "Solaris Systems Programming"
President,
Rite Online Inc.
Voice: +1 (250) 979-1638
URL: http://www.rite-group.com/rich
Edwin
If Solaris is better than Linux, and it really does become open source,
why would current Linux users have any problem switching to it?
What Eltee wrote is more likely, that Linux will adopt all the best
parts of Solaris, and become an even better OS.
Rich Teer
> Well, you know, it is called "Slowaris" for a reason
Only by people who are as out of touch as you are. As documented
on this page: http://www.sun.com/2004-1012/feature/ Apache 1.3.29
on Solaris 10 can handle substantially more connections per second
than RHEL 3.0.
Richard L. Hamilton
"Edwin" <thor...@juno.com> writes:
[...]
> If Solaris is better than Linux, and it really does become open source,
> why would current Linux users have any problem switching to it?
Because they believe in the following illogic:
M$ is evil
M$ is a corporation
corporations are evil (give or take a blind eye to Red Scat)
> What Eltee wrote is more likely, that Linux will adopt all the best
> parts of Solaris, and become an even better OS.
Yes, and I like my pork chops on the wing, too...
--
mailto:rlh...@smart.net http://www.smart.net/~rlhamil
Lasik/PRK theme music:
"In the Hall of the Mountain King", from "Peer Gynt"
Peter Köhlmann
> On Mon, 13 Dec 2004, Peter [UTF-8] Köhlmann wrote:
>
>> Well, you know, it is called "Slowaris" for a reason
>
> Only by people who are as out of touch as you are. As documented
> on this page: http://www.sun.com/2004-1012/feature/ Apache 1.3.29
> on Solaris 10 can handle substantially more connections per second
> than RHEL 3.0.
>
Oh yes. I see. An independent study from certainly very independent parties
Just like the very independent TCO studies from MS
It certainly serves to remind us that we never should trust statistics we
haven't fabricated ourselves out of whole cloth
--
We may not return the affection of those who like us,
but we always respect their good judgement.
Rich Teer
> On the contrary. My guess is that Linux will adopt the best features of
> Solaris and become an even better OS.
Linux will doubtless "borrow" some features from Solaris, but some
of the best features will be left out because Linus won't want them.
I mean, if he won't accept the ability to produce a core dump on
kernel panic, what chance has something like DTrace got?
Rich Teer
> If Solaris is better than Linux, and it really does become open source,
> why would current Linux users have any problem switching to it?
For many users, the choice of OS is a religeous issue. I mean,
look how many Windoze advocates there are, and I think we can all
agree that their OS of choice is underneath the bottom of the pile
when it comes to OS technology!
> What Eltee wrote is more likely, that Linux will adopt all the best
> parts of Solaris, and become an even better OS.
See my response to Eltee...
Joerg Schilling
Rich Teer <rich...@rite-group.com> wrote:
>I know that MIke Cox is a troll, but I think this is the first sensible
>thing he's ever written!
As this is the first senseful reply on his posting, I need to
send you a ;-)
Joerg Schilling
Rich Teer <rich...@rite-group.com> wrote:
>On Mon, 13 Dec 2004, Peter [UTF-8] Köhlmann wrote:
>
>> Well, you know, it is called "Slowaris" for a reason
>
>Only by people who are as out of touch as you are. As documented
>on this page: http://www.sun.com/2004-1012/feature/ Apache 1.3.29
>on Solaris 10 can handle substantially more connections per second
>than RHEL 3.0.
This is something that is also proven by reallity.
After I did upgrade www.berlios.de from Linus to Solaris, I did
feel that there is a second CPU in the box.
Peter Köhlmann
> In article <Pine.SOL.4.58.0412130930180.8903@zaphod>,
> Rich Teer <rich...@rite-group.com> wrote:
>>On Mon, 13 Dec 2004, Peter [UTF-8] Köhlmann wrote:
>>
>>> Well, you know, it is called "Slowaris" for a reason
>>
>>Only by people who are as out of touch as you are. As documented
>>on this page: http://www.sun.com/2004-1012/feature/ Apache 1.3.29
>>on Solaris 10 can handle substantially more connections per second
>>than RHEL 3.0.
>
> This is something that is also proven by reallity.
> After I did upgrade www.berlios.de from Linus to Solaris, I did
> feel that there is a second CPU in the box.
>
I am sure you did
Just as zhose windows guys in here who never ever received a virus
--
"I don't want to belong to any club that would have me as a member."
-- Groucho Marx
Joerg Schilling
Peter =?UTF-8?B?S8O2aGxtYW5u?= <Peter.K...@t-online.de> wrote:
>> This is something that is also proven by reallity.
>> After I did upgrade www.berlios.de from Linus to Solaris, I did
>> feel that there is a second CPU in the box.
>>
>
>I am sure you did
>Just as zhose windows guys in here who never ever received a virus
Looks like it makes no sense to treat you like an honest person :-(
The speed difference of administrative scripts that run about one our
cannot be just sweeped away by a religion like the one you seem to be
stuck to.
General Protection Fault
On Mon, 13 Dec 2004 19:42:37 +0100, Peter Köhlmann wrote:
> begin Joerg Schilling wrote:
>
>> In article <Pine.SOL.4.58.0412130930180.8903@zaphod>,
>> Rich Teer <rich...@rite-group.com> wrote:
>>>
>>>> Well, you know, it is called "Slowaris" for a reason
>>>
>>>Only by people who are as out of touch as you are. As documented
>>>on this page: http://www.sun.com/2004-1012/feature/ Apache 1.3.29
>>>on Solaris 10 can handle substantially more connections per second
>>>than RHEL 3.0.
>>
>> This is something that is also proven by reallity.
>> After I did upgrade www.berlios.de from Linus to Solaris, I did
>> feel that there is a second CPU in the box.
>>
>
> I am sure you did
> Just as zhose windows guys in here who never ever received a virus
I've *received* plenty, I'm just not dumb enough to actually be infected.
--
FreeBSD 4.8-RELEASE i386
1:10PM up 20 days, 18:32, 0 users, load averages: 0.00, 0.00, 0.00
Thomas Vincent
arena and such. Linux users will win more machines in places were
commodity matters, like server farms and the such.
Solaris 10 has a lot of great points like dtrace and ZFS that make it
the first version of Solaris in many years worth looking at.
Cheers,
Tom
all mail refused
>This is something that is also proven by reallity.
Say what you like, but there's only one 'l' in my version of "reality".
Got the guitar CD ?
--
Elvis Notargiacomo master AT barefaced DOT cheek
http://www.notatla.org.uk/goen/
7.031: OnACPower returned value( 0x1 ) which is Equal To 0x1
tha...@dexter.glaci.com
> On Sun, 12 Dec 2004, Mike Cox wrote:
>
>> The reason people will move to Solaris in droves is the same that caused
>> BSD's user flight. It supports more enterprise features. But the
>> difference this time is that Solaris is cheaper than Redhat, is completely
>> backwards compatible, and scales much better on many more CPUs.
>
> I know that MIke Cox is a troll, but I think this is the first sensible
> thing he's ever written!
His point is valid, but I see it more as the reason Sun will stay
relevant on the high end rather than the reason Linux users will
switch. The fact is, most servers are four processors or less,
and Linux scales very nicely at that level. If Sun tries to 'take
back' the commodity server market with open source Solaris, they
have a hard struggle ahead of them. They might have gotten some
traction five years ago... but the market has changed a lot
since then.
Sun's best strategy is to differentiate themselves on hardware,
fold Solaris's high end features into Linux, and save themselves
a bucketload of development cost in the process. I just don't
expect open source Solaris to gain the developer following
that Linux already has, so Sun will likely continue to carry
most of the Solaris development burden. If they go Linux, donate
some Sun big iron to a few universities and prominent kernel
developers, they begin to reap immediate benefits.
--
Thad Phetteplace
http://www.AliensForBush.com - Geeky Political Humor
Chris Cox
>>I know that MIke Cox is a troll, but I think this is the first sensible
>>thing he's ever written!
>
>
> His point is valid, but I see it more as the reason Sun will stay
> relevant on the high end rather than the reason Linux users will
> switch. The fact is, most servers are four processors or less,
> and Linux scales very nicely at that level. If Sun tries to 'take
> back' the commodity server market with open source Solaris, they
> have a hard struggle ahead of them. They might have gotten some
> traction five years ago... but the market has changed a lot
> since then.
>
> Sun's best strategy is to differentiate themselves on hardware,
> fold Solaris's high end features into Linux, and save themselves
> a bucketload of development cost in the process. I just don't
> expect open source Solaris to gain the developer following
> that Linux already has, so Sun will likely continue to carry
> most of the Solaris development burden. If they go Linux, donate
> some Sun big iron to a few universities and prominent kernel
> developers, they begin to reap immediate benefits.
>
I'm not Sun's greatest advocate, but I do use Sun equipment.
I think Sun has a very agressive strategy. I like what I've seen
in their Opteron designs... I think Sun may well become the leader
in that space. I believe Sun has set an agressive goal to get
20% (?) or so of the x86 server market. To me the biggest
transition is moving from high-margin SPARC (with lower volumes)
to low-margin Opteron (with extremely high volumes). It is
a significant change for Sun... but IMHO, they have about the
best chance of pulling it off vs. any other company. Sun has a lot
of guts and teeth where it counts. Let's see what they can do.
With regards to differentiation, look for Sun to have the first
128-way Opteron (I used to say 64... but that's done now).
Sun is already flooding the market with VERY inexpensive dual
Opteron platforms (I own one, w2100z... easily had for <$2000 from them
now... WOW!). Shoot, just last year I bought a SunBlade 150
and the sucker cost us almost $3000. That $2000 Opteron runs circles
around even a SunBlade 2000 that costs $20000!
Sun obviously wants people to run Solaris... HOWEVER, they WILL
NOT miss a sale if you want to run Linux... or EVEN WINDOWS.
All of the new Opteron platforms are WHQL (sp?) certified to
run Windows. Sun wants to be THE supplier of your hardware
whether you choose Solaris, Linux or Windows.
At the price levels they are offering, I think they have the
best Opteron based platform out there right now.
Hope nobody is too in love with SPARC (please remember that
Sun is Sun.... NOT Sun is SPARC). I wish Sun the very
best on this new direction.
No comment on OpenSolaris.... in my opinion, it's actually
not necessary in Sun's overall new plan. Which actually
makes OpenSolaris more valuable in many respects.
Rich Teer
> switch. The fact is, most servers are four processors or less,
> and Linux scales very nicely at that level. If Sun tries to 'take
Perhaps, but Solaris has other enterprise features that are important
even on low end, 1-4 way systems.
> back' the commodity server market with open source Solaris, they
> have a hard struggle ahead of them. They might have gotten some
I agree that Sun has a challenge ahead of them. But I think their
future is rosy.
> Sun's best strategy is to differentiate themselves on hardware,
> fold Solaris's high end features into Linux, and save themselves
Completely disagree. Why would Sun waste R&D $$$ on Linux? They
already have an enterprise class (soon to be) open source OS, that
is free. If you want Solaris features, why not migrate to Solaris?
> a bucketload of development cost in the process. I just don't
> expect open source Solaris to gain the developer following
> that Linux already has, so Sun will likely continue to carry
Maybe, maybe not. You seem to forget that it's only relatively
recently that Linux was the prime open source platform. It wasn't
that long ago when stuff was first developed on Solaris/SunOS.
It could well be that those that aren't rabid Linux advocates
might see Open Solaris as a new and interesting intellectual
challenge, and tag along for the ride for the sheer hack value.
> most of the Solaris development burden. If they go Linux, donate
> some Sun big iron to a few universities and prominent kernel
> developers, they begin to reap immediate benefits.
What's to stop them from doing that for Open Solaris?
Will Hartung
<tha...@dexter.glaci.com> wrote in message
news:41bf289d$0$46427$4c5e...@news.dias.net...
> The fact is, most servers are four processors or less,
> and Linux scales very nicely at that level. If Sun tries to 'take
> back' the commodity server market with open source Solaris, they
> have a hard struggle ahead of them. They might have gotten some
> traction five years ago... but the market has changed a lot
> since then.
But if the transition to a more stable platform is "mostly" transparent,
then the there isn't that much of a struggle.
Granted, we have no facts yet but consider IF the following are true:
o Janus lets users move "legacy" (funny -- Linux as legacy, oh well) Linux
binaries to Solaris 10
o Solaris 10 has good hardware support for server components
o Solaris 10 support compiling of Linux source code (dunno if it will or
not)
o Solaris 10 has comparable pricing model to Linux
o Solaris 10 administration, while different from Linux, changing to Sol 10
from Linux is less painful than from Windows
So, IF Solaris 10 provides reasonable support for Linux applications, and a
stable platform for vendor binaries in the future, that makes adoption of
Solaris 10 less expensive, and less risky.
Simple case, say your vendor has App 2.0 of his application. But, the old
version runs on Red Hat XYZ vs the new one on Red Hat QED. In order to use
the new version you need to upgrade your Linux box. Now, if that vendor
supports App 2.0 on Solaris, or if they will support the App running in
Janus of Sol 10, then why not upgrade to Sol 10 rather than Red Hat QED? By
doing so you have a pretty darn good chance that when App 2.5 comes out, you
won't need to update your OS to support the application.
If for no other reason than simply being able to keep your OS stable (i.e.
not for any of the other improvements in Sol 10), that seems like a
compelling concern to think about. It's one thing having to update your
application, but updating the OS is a real pain.
And vendors will think about that as well. Janus gives vendors the ability
to migrate their Linux apps to Sol 10. First they can run and support it in
Janus, then they can try to port the source code over and support a native
version (assuming they may not already have a SPARC version). And once they
make that leap, they have a more stable platform to release binary x86
applications.
Now, obviously this is not all going to happen as soon as Sol 10 hits the
market. But I do think it will happen soon, within the year we'll be seeing
movement. I think if Sun manages to push the "port to Sol 10" meme, that as
folks approach the need for either a new system, or having to upgrade their
old system, they may well consider switching over.
While there are some organizations attracted to Red Hat because of its open
source roots, I think most are attracted to its (originally) low price and
functionality. But if I can run the same commodity software (like OSS Web
server tools, etc.), as well as vendor binaries, on a more stable platform,
with similar pricing chracteristics, then all of a sudden Red Hat loses its
luster. The one thing that Red Hat can not guarantee, simply, is the
stability of its underlying platform, because it's not in control of it. Sun
can, and does, control that, and now it's coming out on the same range of
commodity server hardware, with similar performance and pricing, but with
more long term stability (as demonstrated historically -- past performance
does not guarantee future results :-)).
That makes Sol 10 a compelling option in the future for new builds and
upgrades, IMHO, and all of this is regardless of what happens with
OpenSolaris. OpenSolaris can just be icing on the cake.
This is why I think that Sun can compete, and compete well, with Linux on
the low end, but it won't happen overnight.
Regards,
Will Hartung
(wi...@msoft.com)
tha...@dexter.glaci.com
>> some Sun big iron to a few universities and prominent kernel
>> developers, they begin to reap immediate benefits.
>
> What's to stop them from doing that for Open Solaris?
There is nothing to stop them doing it, and if they are serious
about open Solaris they will. My argument is that their return on
investment would be better in trying to tap into the existing Linux
developer pool rather than trying to build a new community around
Solaris. They seem to be clinging to the idea of using Solaris to
differentiate themselves from the other hardware vendors. That
made sense in the old days of the closed source unix wars, but is
much harder to pull off in today's fast paced open source world.
Of course I'm doing some major crystal ball gazing here. If Sun
plays it right and engages the community with the right kind of
challenges, they can certainly build a thriving developer pool
around Solaris. Given some of their past missteps, however, I'm
not so optimistic. Sun has actually done a lot of great stuff
for the open source community, yet their upper managment continues
to throw away goodwill by the truckloads with some of their
boneheaded statements and other ill-thought actions. I'm not
convinced they have the 'community building' knack. Time will
tell, and I wish them well. Sun has done some great things over
the years and continues to churn out some really great tech.
Later,
Jerry McBride
> That is why they post so much FUD on sites such as slashdot.org. The
> reason Linux became popular was that they were not as conservative as BSD,
> and
> would add features constantly. That meant if you wanted some enterprise
> functionality for free, you'd have to go with linux (SMP, etc.).
>
> The downside of that linux philosophy was that backwards compatiblitiy
> went
> down the drain. glibc interfaces and kernel calls would change
> significantly between versions. Sometimes the "enterprise" feature would
> be botched, like the scheduler in the 2.4 kernel.
>
> Now that Solaris 10 is going to be open sourced, Linux advocates rightly
> fear that Linux will be abandoned in favor of Solaris 10. Similar to how
> BSD was left in droves for Linux. That's right, SUN's brilliant strategy
> of making Solaris open sourced is using the linux advocates' techniques
> against them.
>
> The reason people will move to Solaris in droves is the same that caused
> BSD's user flight. It supports more enterprise features. But the
> difference this time is that Solaris is cheaper than Redhat, is completely
> backwards compatible, and scales much better on many more CPUs.
>
> Now if only SUN would release its compiler under a BSD license....
Solaris 10 was supposed to address the poor performance issues with previous
versions. After installing it, testing it I can say that "it doesn't"
address the performance issue at all... It may scale pretty well, but it
pokes around under light load as if it was carry the world...
It may be better than windows, but it pretty much looks like an ugly duck
when compared to Linux.
--
******************************************************************************
Registered Linux User Number 185956
FSF Associate Member number 2340 since 05/20/2004
Join me in chat at #linux-users on irc.freenode.net
Buy an Xbox for $149.00, run linux on it and Microsoft loses $150.00!
5:08pm up 66 days, 54 min, 8 users, load average: 0.10, 0.14, 0.09
Freeride
> On Mon, 13 Dec 2004, Eltee wrote:
>
>> On the contrary. My guess is that Linux will adopt the best features of
>> Solaris and become an even better OS.
>
> Linux will doubtless "borrow" some features from Solaris, but some
> of the best features will be left out because Linus won't want them.
> I mean, if he won't accept the ability to produce a core dump on
> kernel panic, what chance has something like DTrace got?
>
Bullshit! Sun's CDDL conflicts with the GPL and no Open Source developer is
going to touch it with a 20 foot stick. And I am sure that Sun would love
to have some of its code show up in Linux so it can throw a big giant
monkey wrench into the development and adoption of Linux.
Rich Teer
> Bullshit! Sun's CDDL conflicts with the GPL and no Open Source developer is
Who says Open Solaris will use CDDL?
> going to touch it with a 20 foot stick. And I am sure that Sun would love
That's kind of amusing. I can name three open source developers
who are actively involved with the Open Solaris pilot. So much
for your FUD. Repeat after me until it sinks in: Open Source != GPL.
> to have some of its code show up in Linux so it can throw a big giant
> monkey wrench into the development and adoption of Linux.
I suspect that some Sun code (legally) is already in Linux.
BTW, it's bad netiquette to change the newsgroup line without
notice, so I've re-established the original one.
Casper H.S. Dik
>Solaris 10 was supposed to address the poor performance issues with previous
>versions. After installing it, testing it I can say that "it doesn't"
>address the performance issue at all... It may scale pretty well, but it
>pokes around under light load as if it was carry the world...
What poor performance issues? I've always found that Solaris is
faster than Linux for the things I do. But that's perhaps a function
of the things I do; there's still some catching to do when it comes
to Xserver performace; (no apgart driver available yet).
Casper
--
Expressed in this posting are my opinions. They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.
Casper H.S. Dik
>Bullshit! Sun's CDDL conflicts with the GPL and no Open Source developer is
>going to touch it with a 20 foot stick. And I am sure that Sun would love
>to have some of its code show up in Linux so it can throw a big giant
>monkey wrench into the development and adoption of Linux.
Any license other than the GPL conflicts with the GPL. There are a lot
of freeware developers who wouldn't touch the GPL with a 30 foot stick.
(And, as the original author, you can release code under several concurrent
licenses; there exists code which is both under the GPL and the BSD license
and where you as user can chose which license you use for redistribution)
Joerg Schilling
Rich Teer <rich...@rite-group.com> wrote:
>> Sun's best strategy is to differentiate themselves on hardware,
>> fold Solaris's high end features into Linux, and save themselves
>
>Completely disagree. Why would Sun waste R&D $$$ on Linux? They
>already have an enterprise class (soon to be) open source OS, that
>is free. If you want Solaris features, why not migrate to Solaris?
I completly agree, it is easier to add the few extra features from Linux
than adding the missing features to Linux. Writing drivers is not that
hard and Sun did already start with massive driver development and a
new x86 driver group.
>> a bucketload of development cost in the process. I just don't
>> expect open source Solaris to gain the developer following
>> that Linux already has, so Sun will likely continue to carry
>
>Maybe, maybe not. You seem to forget that it's only relatively
>recently that Linux was the prime open source platform. It wasn't
>that long ago when stuff was first developed on Solaris/SunOS.
I also forsee that there are other reasons why a developer would switch
from Linux to Solaris.
One reason is that the Linux development is dominated by Linus.
It is even funny to see that Linus repeatdly publishes statements that
he cannot see how Sun may dominate the OSS market. From what I know
about Sun, I can tell that Sun does not seem to be interested in
dominating the OpenSolaris development. This may be an important factor
for OpenSolaris.
Joerg Schilling
Jerry McBride <mcbr...@comcast.net> wrote:
>Solaris 10 was supposed to address the poor performance issues with previous
>versions. After installing it, testing it I can say that "it doesn't"
>address the performance issue at all... It may scale pretty well, but it
>pokes around under light load as if it was carry the world...
Looks like you never compared Solaris & Linux performance :-(
With my usage, Linux always has been slower or at most only as fast than
Solaris.
Linux has poor memory management and scheduling. How can such an OS
behave better than Solaris?
Joerg Schilling
Casper H.S. Dik <Caspe...@Sun.COM> wrote:
>Jerry McBride <mcbr...@comcast.net> writes:
>
>>Solaris 10 was supposed to address the poor performance issues with previous
>>versions. After installing it, testing it I can say that "it doesn't"
>>address the performance issue at all... It may scale pretty well, but it
>>pokes around under light load as if it was carry the world...
>
>What poor performance issues? I've always found that Solaris is
>faster than Linux for the things I do. But that's perhaps a function
>of the things I do; there's still some catching to do when it comes
>to Xserver performace; (no apgart driver available yet).
I concur.
I don't compare Xserver performance, but for the typical work of a developer
and for Web Servers, Solaris is faster than Linux and already has been
faster than Linux in pre Solaris 10 times.
Joerg Schilling
Freeride <free...@maillinux.org> wrote:
I corrected your violation of the nettiquette (you did change
the Newsgroups: witout notice!)
>Bullshit! Sun's CDDL conflicts with the GPL and no Open Source developer is
>going to touch it with a 20 foot stick. And I am sure that Sun would love
>to have some of its code show up in Linux so it can throw a big giant
>monkey wrench into the development and adoption of Linux.
Well, with CDDL people like Linus have more problems with trying to dominate
a project..... but not all people behave this way.
Note that it is not the CDDL that conflichts with the GPL but it is the
GPL that conflicts with the CDDL.
The problem is that the GPL tries to reduce freedom in a way that is not
compatible with the CDDL.
The GPL does not conflict with only the BSD license because the BSD license
tolerates this freedom reduction.
The problem with the GPL is that it tries to enforce several restrictions
and that it does not allow a mix of "compatible" licenses. The fact that
it tries to impose GPL rules to any piece of software (if taken carefully)
only allows the original Author to add software to a GPL project. This is
all not true for the CDDL. So why should authors not like CDDL?
Peter C. Tribble
j...@cs.tu-berlin.de (Joerg Schilling) writes:
> In article <oc5492x...@spinner.my.domain>,
> Jerry McBride <mcbr...@comcast.net> wrote:
>
>>Solaris 10 was supposed to address the poor performance issues with previous
>>versions. After installing it, testing it I can say that "it doesn't"
>>address the performance issue at all... It may scale pretty well, but it
>>pokes around under light load as if it was carry the world...
>
> Looks like you never compared Solaris & Linux performance :-(
>
> With my usage, Linux always has been slower or at most only as fast than
> Solaris.
>
> Linux has poor memory management and scheduling. How can such an OS
> behave better than Solaris?
Interesting question, but there are some not entirely uncommon
circumstances where it does. Particularly, process initialization -
startup, scripts, builds, configure scripts. You can get up to a factor
2 slowdown in Solaris compared to Linux.
I would agree with much of the sentiment above - push them, and Solaris
does behave much better. But under light or modest load, and there are
cases where Solaris performance really isn't quite up to par.
One problem is that it's that sort of behaviour that governs many
users' impressions of a system. It's something we need to work at.
(As part of OpenSolaris, I've tried looking at the scheduler, because I
know there are interactions there. But I don't have the familiarity
with the code or the skills yet to understand it.)
--
-Peter Tribble
MRC Rosalind Franklin Centre for Genomics Research
http://www.rfcgr.mrc.ac.uk/~ptribble/
Joerg Schilling
Peter C. Tribble <ptri...@hgmp.mrc.ac.uk> wrote:
>>>Solaris 10 was supposed to address the poor performance issues with previous
>>>versions. After installing it, testing it I can say that "it doesn't"
>>>address the performance issue at all... It may scale pretty well, but it
>>>pokes around under light load as if it was carry the world...
>>
>> Looks like you never compared Solaris & Linux performance :-(
>>
>> With my usage, Linux always has been slower or at most only as fast than
>> Solaris.
>>
>> Linux has poor memory management and scheduling. How can such an OS
>> behave better than Solaris?
>
>Interesting question, but there are some not entirely uncommon
>circumstances where it does. Particularly, process initialization -
>startup, scripts, builds, configure scripts. You can get up to a factor
>2 slowdown in Solaris compared to Linux.
For Solaris 2.6 and older I would definitely agree that startup of programs
was slower.
Solaris and Solaris apps if linked in a SVr4 compliant way need more dynamic
libs than Linux. This should have been reduced wwith never versions.
I don't believe that the Solaris schedular is related to this.
Dan Espen
> In article <cpq4pd$jjn$1...@helium.hgmp.mrc.ac.uk>,
> Peter C. Tribble <ptri...@hgmp.mrc.ac.uk> wrote:
>
>>>>Solaris 10 was supposed to address the poor performance issues with previous
>>>>versions. After installing it, testing it I can say that "it doesn't"
>>>>address the performance issue at all... It may scale pretty well, but it
>>>>pokes around under light load as if it was carry the world...
>>>
>>> Looks like you never compared Solaris & Linux performance :-(
>>>
>>> With my usage, Linux always has been slower or at most only as fast than
>>> Solaris.
>>>
>>> Linux has poor memory management and scheduling. How can such an OS
>>> behave better than Solaris?
>>
>>Interesting question, but there are some not entirely uncommon
>>circumstances where it does. Particularly, process initialization -
>>startup, scripts, builds, configure scripts. You can get up to a factor
>>2 slowdown in Solaris compared to Linux.
>
> For Solaris 2.6 and older I would definitely agree that startup of programs
> was slower.
>
> Solaris and Solaris apps if linked in a SVr4 compliant way need more dynamic
> libs than Linux. This should have been reduced wwith never versions.
> I don't believe that the Solaris schedular is related to this.
I'd like more info.
Here's what I see, Mandrake 10.0 vs. Solaris 2.8:
linux> ldd /bin/ls
linux-gate.so.1 => (0xffffe000)
librt.so.1 => /lib/tls/librt.so.1 (0x40025000)
libtermcap.so.2 => /lib/libtermcap.so.2 (0x4003a000)
libc.so.6 => /lib/tls/libc.so.6 (0x4003e000)
libpthread.so.0 => /lib/tls/libpthread.so.0 (0x40186000)
/lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000)
linux> ldd /usr/bin/make
linux-gate.so.1 => (0xffffe000)
libc.so.6 => /lib/tls/libc.so.6 (0x40025000)
/lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000)
linux> ldd /bin/cat
linux-gate.so.1 => (0xffffe000)
libc.so.6 => /lib/tls/libc.so.6 (0x40025000)
/lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000)
solaris> ldd /bin/ls
libc.so.1 => /usr/lib/libc.so.1
libdl.so.1 => /usr/lib/libdl.so.1
/usr/platform/SUNW,Ultra-5_10/lib/libc_psr.so.1
solaris> ldd /usr/ccs/bin/make
libintl.so.1 => /usr/lib/libintl.so.1
libnsl.so.1 => /usr/lib/libnsl.so.1
libsocket.so.1 => /usr/lib/libsocket.so.1
libw.so.1 => /usr/lib/libw.so.1
libm.so.1 => /usr/lib/libm.so.1
libc.so.1 => /usr/lib/libc.so.1
libdl.so.1 => /usr/lib/libdl.so.1
libmp.so.2 => /usr/lib/libmp.so.2
/usr/platform/SUNW,Ultra-5_10/lib/libc_psr.so.1
solaris> ldd /bin/cat
libc.so.1 => /usr/lib/libc.so.1
libdl.so.1 => /usr/lib/libdl.so.1
/usr/platform/SUNW,Ultra-5_10/lib/libc_psr.so.1
Looks like a mixed bag, but I can see some flaws in
this test, Solaris make seems to be network aware,
gnu ls provides color.
Casper H.S. Dik
>(As part of OpenSolaris, I've tried looking at the scheduler, because I
>know there are interactions there. But I don't have the familiarity
>with the code or the skills yet to understand it.)
The Solaris scheduler is fairly complicated; however, it's been an
O(1) scheduler (a much touted Linux 2.something feature) since about
Solaris 2.4.
Peter C. Tribble
Casper H.S. Dik <Caspe...@Sun.COM> writes:
> ptri...@hgmp.mrc.ac.uk (Peter C. Tribble) writes:
>
>>(As part of OpenSolaris, I've tried looking at the scheduler, because I
>>know there are interactions there. But I don't have the familiarity
>>with the code or the skills yet to understand it.)
>
> The Solaris scheduler is fairly complicated; however, it's been an
> O(1) scheduler (a much touted Linux 2.something feature) since about
> Solaris 2.4.
That it may be. The problem we saw was (as far as I recall - sorry Gary
if I mess up) due to a scheduling race in the following scenario:
Shell script runs through many short lived processes. When new process
is launched, parent sleeps. When child exits, two scheduling events are
generated - parent wakes up and tries to find a cpu to run on, and the
cpu the child was running on is now free.
Needless to say it didn't always get it right. Especially if this is a
multiprocessor system with something else happening. (The fix was to use
processor sets or even a simple pbind to stop processes wandering
around between cpus - in other words, to override some of the
scheduling decisions.)
We were seeing a slowdown of something like a factor 10 in some extreme
cases. I mean, you would think that if you had a multiprocessor machine
then you wouldn't be able to detect much of an impact on system
responsiveness provided there was at least one CPU free. The reality
was otherwise, unfortunately.
Must run some tests on build 72, to see how it's changed
recently. Things are getting much better in the last couple of Solaris
express releases.
Al Hopper
You're not comparing like with like. The biggest issue with this
comparison is that you're comparing an x86 (based) architecture with a
SPARC arch. That does not make sense.
On an older Solaris 9 box:
$ cat /etc/release
Solaris 9 8/03 s9x_u4wos_08b x86
Copyright 2003 Sun Microsystems, Inc. All Rights Reserved.
Use is subject to license terms.
Assembled 17 June 2003
$ ldd /bin/ls
libc.so.1 => /usr/lib/libc.so.1
libdl.so.1 => /usr/lib/libdl.so.1
On a newer Solaris 9 box:
$ cat /etc/release
Solaris 9 9/04 s9x_u7wos_09 x86
Copyright 2004 Sun Microsystems, Inc. All Rights Reserved.
Use is subject to license terms.
Assembled 28 June 2004
$ ldd /bin/ls
libc.so.1 => /usr/lib/libc.so.1
libdl.so.1 => /usr/lib/libdl.so.1
On a Sol 10 box running Beta 70 in 32-bit mode:
$ cat /etc/release
Solaris 10 s10_70 X86
Copyright 2004 Sun Microsystems, Inc. All Rights Reserved.
Use is subject to license terms.
Assembled 22 October 2004
$ ldd /bin/ls
libc.so.1 => /lib/libc.so.1
libm.so.2 => /lib/libm.so.2
Conclusion: Give a more recent version of Solaris a spin on an x86
server, ideally the latest available Solaris Express release (B72)
and I think you'll be impressed.
Al Hopper Logical Approach Inc, Plano, TX. a...@logical-approach.com
Voice: 972.379.2133 Fax: 972.379.2134
"It's a good thing that HP never acquired the rights to penicillin. If they had,
mankind would have perished from widespread disease while HP tried to figure out
how to integrate it with anthrax." TD (fullname unknown).
Al Hopper
You're not comparing like with like. The biggest issue with this
comparison is that you're comparing an x86 (based) architecture with a
SPARC arch. That does not make sense.
On an older Solaris 9 box:
$ cat /etc/release
Solaris 9 8/03 s9x_u4wos_08b x86
Copyright 2003 Sun Microsystems, Inc. All Rights Reserved.
Use is subject to license terms.
Assembled 17 June 2003
$ ldd /bin/ls
libc.so.1 => /usr/lib/libc.so.1
libdl.so.1 => /usr/lib/libdl.so.1
On a newer Solaris 9 box:
$ cat /etc/release
Solaris 9 9/04 s9x_u7wos_09 x86
Copyright 2004 Sun Microsystems, Inc. All Rights Reserved.
Use is subject to license terms.
Assembled 28 June 2004
$ ldd /bin/ls
libc.so.1 => /usr/lib/libc.so.1
libdl.so.1 => /usr/lib/libdl.so.1
On a Sol 10 box running Beta 70 in 32-bit mode:
Alan Coopersmith
|linux> ldd /usr/bin/make
| linux-gate.so.1 => (0xffffe000)
| libc.so.6 => /lib/tls/libc.so.6 (0x40025000)
| /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000)
| libintl.so.1 => /usr/lib/libintl.so.1
| libnsl.so.1 => /usr/lib/libnsl.so.1
| libsocket.so.1 => /usr/lib/libsocket.so.1
| libw.so.1 => /usr/lib/libw.so.1
| libm.so.1 => /usr/lib/libm.so.1
| libc.so.1 => /usr/lib/libc.so.1
| libdl.so.1 => /usr/lib/libdl.so.1
| libmp.so.2 => /usr/lib/libmp.so.2
| /usr/platform/SUNW,Ultra-5_10/lib/libc_psr.so.1
Eh, not sure how useful a comparison that is - if you link with -z
lazyload, many of those will never get loaded at startup (and some,
never at all). The linker guys have done some amazing things to
speed up loads when the right options are used.
--
________________________________________________________________________
Alan Coopersmith * al...@alum.calberkeley.org * Alan.Coo...@Sun.COM
http://www.csua.berkeley.edu/~alanc/ * http://blogs.sun.com/alanc/
Working for, but definitely not speaking for, Sun Microsystems, Inc.
Dan Espen
> On Wed, 15 Dec 2004 15:24:20 -0500, Dan Espen wrote:
>
>> j...@cs.tu-berlin.de (Joerg Schilling) writes:
>>
>>> In article <cpq4pd$jjn$1...@helium.hgmp.mrc.ac.uk>,
>>> Peter C. Tribble <ptri...@hgmp.mrc.ac.uk> wrote:
...
>>> Solaris and Solaris apps if linked in a SVr4 compliant way need more dynamic
>>> libs than Linux. This should have been reduced wwith never versions.
>>
>> I'd like more info.
>>
>> Here's what I see, Mandrake 10.0 vs. Solaris 2.8:
>>
>>
>> linux> ldd /bin/ls
>> linux-gate.so.1 => (0xffffe000)
>> librt.so.1 => /lib/tls/librt.so.1 (0x40025000)
>> libtermcap.so.2 => /lib/libtermcap.so.2 (0x4003a000)
>> libc.so.6 => /lib/tls/libc.so.6 (0x4003e000)
>> libpthread.so.0 => /lib/tls/libpthread.so.0 (0x40186000)
>> /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000)
>> solaris> ldd /bin/ls
>> libc.so.1 => /usr/lib/libc.so.1
>> libdl.so.1 => /usr/lib/libdl.so.1
>> /usr/platform/SUNW,Ultra-5_10/lib/libc_psr.so.1
>>
>> Looks like a mixed bag, but I can see some flaws in
>> this test, gnu ls provides color.
>
> You're not comparing like with like. The biggest issue with this
> comparison is that you're comparing an x86 (based) architecture with a
> SPARC arch. That does not make sense.
It looked to me like the biggest issue
was the lack of equivalent function.
> On an older Solaris 9 box:
>
> $ cat /etc/release
> Solaris 9 8/03 s9x_u4wos_08b x86
> Copyright 2003 Sun Microsystems, Inc. All Rights Reserved.
> Use is subject to license terms.
> Assembled 17 June 2003
> $ ldd /bin/ls
> libc.so.1 => /usr/lib/libc.so.1
> libdl.so.1 => /usr/lib/libdl.so.1
All your examples show 2 shared libs needed on a basic
Solaris x86 executables where the minimum for linux seemed to be 3.
Above it says Solaris needs more dynamic libs than linux,
but for the simple cases, the opposite seems to be true.
Hamilcar Barca
06:29:52 +0000), Rich Teer wrote:
> Who says Open Solaris will use CDDL?
What's the point, if any, of the CDDL? Is Sun just testing the waters
to see how much negative feedback is generated? Why do they go to the
expense of creating a new "open source" [sic] license if they don't intend
to use it?
Who says Open [sic] Solaris won't use the CDDL?
> Repeat after me until it sinks in: Open Source != GPL.
The proper relation is, and remains:
Open Source" < GPL
> I suspect that some Sun code (legally) is already in Linux.
There may be. However, there isn't a single Sun Microsystems copyright
notice in the 2.6.9 kernel, so Sun's contribution, if there is one, is
miniscule. (OTOH, so what?)
--
"I think [Microsoft] have a pretty good story, but I tell you, game's on.
We've got to prove ourselves, and some people are choosing Linux."
-- Steve Ballmer. http://news.com.com/2008-1082-998297.html. 25 Apr 2003
Hamilcar Barca
+0000), Joerg Schilling wrote:
> One reason is that the Linux development is dominated by Linus.
Your penis envy is becoming obvious.
Hamilcar Barca
+0000), Joerg Schilling wrote:
> Well, with CDDL people like Linus have more problems with trying to dominate
> a project.....
Do you then claim that all arguments in favor of the C