Sun's 'Project Copy Linux' goes commercial

The first supported first version of Sun Microsystems' OpenSolaris, AKA Project Indiana, makes its debut today with additional backing from Amazon's Elastic Computing Cloud. Sun is introducing three levels of paid support for the OpenSolaris 2008.5 code drop for developers and end users. Support starts at $49 per incident for …

[Matt Bryant]

"Packaging company"! When will Sun learn?

This is typical Sun, bad-mouthing their competition rather than relying on a technical argument. Red Hat has developed a measure of respect in the Linux community from hard and co-operative work, and from the business community by supplying a good product and services, but there are still Linux people that accuse Red Hat of being "the Microsoft of the Linux world"! How does Sun think the Linux community will react to their aggressive and sneering approach? I expect some take up from Slowaris shops desperate to retain their expensive Slowaris admins and ditch their SPARC kit (well, up until they hit the wall of non-compatibility between x86 and SPARC Solaris anyway), but zero take up from existing Red Hat, SuSE or Ubuntu users.

As for Sun bragging about their support, having experienced what was supposedly their best with fully-supported commercial Solaris on SPARC in a mission-critical environment, let's just say I would much rather have Red Hat's which has never let me down. And I assume since support is coming from the people that created it, that ZFS support will be delivered by NetApp then? :P

[Anonymous Coward]

That's odd

"Red Hat, which Sun called just a "packaging company"."

That's a really odd thing to say about Red Hat at this point. Some of the major contributions are listed at

http://fedoraproject.org/wiki/RedHatContributions

[Brian]

Inflammatory title?

You're just BEGGING for attention here, aren't you?

[Anonymous Coward]

Wake up Sun

The opposition isn't some other Linux company, it's Gates. Not that any other Linux outfit is notably better in acknowledging this...

[Anonymous Coward]

Did someone >really< say this?

So which of the 30,000 people at Sun called Red Hat "just a packaging company"? Did someone >really< say that or is El Reg just out to start a fight? If someone did say it, why didn't their name make it into the article, like all the other attributed quotes -- none of which said anything like that?

[Don]

Hmmm....

Looks like Sun really needs to quit trying to make Solaris anything other than a proprietary OS.

[moronatwork]

redcrap?

<i>This is typical Sun, bad-mouthing their competition rather than relying on a technical argument. Red Hat has developed a measure of respect in the Linux community from hard and co-operative work, and from the business community by supplying a good product and services, but there are still Linux people that accuse Red Hat of being "the Microsoft of the Linux world"!</i>

??? What kind of respect do you offer a company that designs a broken packaging system, and refuses to fix it? And if your rpm database gets corrupted, well, reinstall - HUH? And I've had open tickets that stayed opened for more than a year (2.5 years now, if they still haven't closed it) *WITH* support acknowledging the issue, but refusing to work on it.

Bah. Redcrap sucks.

[James Henstridge]

packaging

So when Sun provides Python, Apache, PHP, Ruby, etc with OpenSolaris, that is totally different to what mere packaging companies like Red Hat do?

[Antoinette Lacroix]

Well said @Mommy

Actually, there might be a driver for USB toasters. Open Solaris has a bustling community, they only don't post here. I was able to find drivers for all of my 'more or less exotic' gear. Some had to be build from source, but they are there.

[EnigmaForce]

I'm quitting Solaris 10...

...for Opensolaris 2008.05 - downloading now 8-)

+2 @mommycalled for good measure, well said indeed :-)

[Anonymous Coward]

Well said @Mommy @Antoinette

So so true, makes me wonder how many people who are complaining about solaris have actually used it properly, instead of just crying and going back to ubuntu when they can't find 'top' ;)

I'd rather not have a driver for 'crappynetworkcardX' and have to go and get myself a supported card than for it to look like it is all working and then randomly get around 10% of expected throughput and packet losses.

Not that I am totally dispelling linux as a 'useful' tool, It has it's use cases, I just think this whole fanboy 'will not hear a bad word about it even though I have not tried the alternatives' thing is getting irritating.

@Matt, laughing at your random mishmash of half-assed arguments, open ended points and general statements backed up by nothing whatsoever always makes me chuckle.. keep up the good work :)

Regards,

A.C

[James Henstridge]

@mommycalled

I remember one of the Sun engineers reporting some results from DTrace that showed that one of the GNOME applets was regularly polling a file at regular intervals. When people tried to reproduce the results on Linux with less sophisticated tools like strace, they couldn't see the polling.

The reason for this was that Linux provided a file change notification API to user space, which allowed the app to do its job without polling, while Solaris did not.

Of course, today Solaris has exposed a user space file change notification API, and people are working on DTrace-like tools for Linux. To say that innovation only occurs on the Solaris side or only happens on the Linux side is deluded.

[Anonymous Coward]

Well, RedHat *is* a packaging company

Flame me all you want, but I will have to agree with the statement that RedHat is by and large a packaging company. This statement has definitely truth to it when you actually try to get support from RedHat - during many support calls I've logged with RedHat I could get a definite feeling that they don't have anyone on their staff who can properly answer the questions pertaining to the area I'm interested in, instead deferring to developer forums and such external to RedHat. In other words they seem to be doing a fine job packaging software invented and built somewhere else, but they don't own or control the overwhelming majority of stuff that is in RHEL and it shows in their ability to support it. If you have a contract with RedHat, just for laughs sake try to log an RFE against some obscure area of the Linux kernel and see them respond back with some obscure references to developer forums trying wriggle themselves out of a situation. This sort of thing doesn't happen with either Sun or HP that actually own the software they sell (Solaris and HP-UX respectively) and therefore you get more definite answers if you want something fixed or implemented. So yeah, RedHat *is* a packaging company in many respects, they will never be able to stand up to AIX, HP-UX or Solaris with their RHEL simply on that basis alone, unless of course they dramatically grown their engineer headcount to properly develop and support their product.

[E]

ZFS

ZFS looks rather nice. Linux does have LVM but it is not idiot proof enough, at least for this idiot. Tru64 had a disk pool / data set scheme too, and I never managed to delete my data using it ;-) Maybe ZFS will be as idiot proof.

The Solaris release runs nicely in VBox, so maybe I'll learn all about Sun's clustering / parallel programming offerings.

[Jeff Tickle]

Title?

Such mudslinging; the title, the comments, and Sun... isn't Free software about choice? Maybe not JUST about choice, but isn't that a part of it? I don't think I'm going to be replacing Linux with OpenSolaris in a production environment anytime soon, but I'm also not going to deride the work that went into it. Say what you want about Sun as a corporation, but your human brethren made this software. Be nice.

[Sabahattin Gucukoglu]

Misc

mommycalled: you must of course be aware that SunOS/[Open]Solaris is a BSD derivative with some big-iron features in it (some specific to Sun hardware). So do stop spreading the "Sun are original inventors" tripe. Thank you. If I wanted a BSD, I'd go look for one and spare myself the mess that OpenSolaris on X86 is right now. And it'd be free, fully-supported and rock solid too (my preference is NetBSD at the moment). And I'd still end up with an OS whose kernel team was a mailing list away, and who in turn are closely connected with the rest of the core team for an integrated OS (which Sun thinks is an exclusive for a commercial vendor even though companies like RedHat have their own developers to work on QA, distribution-specific patches, etc, etc, etc).

Cheers,

Sabahattin

[Anonymous Coward]

Re: Did someone &gt;really&lt; say this?

We have updated the article to make clear who said this.

Drew

El Reg

[mommycalled]

More foolish from people who don't know

Sabahattin Gucukoglu says

"mommycalled: you must of course be aware that SunOS/[Open]Solaris is a BSD derivative with some big-iron features in it (some specific to Sun hardware). So do stop spreading the "Sun are original inventors" tripe. Thank you. If I wanted a BSD, I'd go look for one and spare myself the mess that OpenSolaris on X86 is right now."

Just another fool. Solaris is SysVr4 based. Sun changed from a BSD based OS to SysVr4 around 1988, that's 20 years ago.

What's different between NetBSD kernel development team being an email away and the Solaris Kernel development team being an email away. Bzzt NOTHING. Besides I don't have to deal with Theo de Raadt

By the way Dtrace and ZFS were Sun developed FOSS. I guess releasing the Solaris source into the wild is nothing. By the way ZFS is part of FreeBSD as of April 2007. Since Sun has contributed nothing you won't be using ZFS in your BSD system will you?

[juddy]

Sun controls their kernel - Redhat does not.

Redhat is a packaging company in that they are re-releasing freely available components, some of which are extended or provided by Redhat.

Trouble is, of the packages listed on the Fedora contributions list, not all are applicable to the rest of the Linux companies. Novell/Suse suffer from the same lack of *complete* compatibility.

This point is moot when you think about the interoperability of Solaris packages or apps as applied to Linux, however.

Take a look at some of the videos of round table meetings of Sun engineers. It's immediately clear (to me) that the disconnects that exist in the Linux world wrt problem solving, best practice, and road maps are of a completely different nature, if they can even be compared directly.

Looking at Sun's hardware business, the width and breadth of that hardware, the services that run on them (it's still a Sun, IBM, and HP world within the Fortune 500), and this effieciency in communication and execution within Sun's OS division, it appears that Redhat is truly little more than a "packaging company" (with crummy package management)

Inflammatory, to be sure, but this is the big time.

[Sabahattin Gucukoglu]

Re: More foolish from people who don't know

Mommycalled: "Solaris is SysVr4 based. Sun changed from a BSD based OS to SysVr4

around 1988, that's 20 years ago."

I wouldn't call Linux a BSD derivative (more like SysV), but it still owes a lot to BSD. Thanks for the heritage tip (I stand corrected), but my point is that Sun don't have to write everything to be famous. And that's what you think commercial companies can do exclusively. No, Sun are standing on the shoulders of giants. That doesn't mean their contributions aren't valuable or that they don't have great engineers, but many FOSS developers are very good (including those in BSD land who aimed for quality).

FTR: I use Linux on my server for various reasons (primarily, accessibility in textmode). NetBSD's hardware support still has a bit to be desired, but it works on my playboxes, both textmode and GUI, network or no. But OpenSolaris really *was* a mess the last time I looked, and I'm going to take special care not to look at it again until it runs on enough everyday desktop hardware. Like, you know, RTL8169 cards embedded on mainboards (yes, I know, but that doesn't change anything - it still didn't work). And all that aside, I have issue with Sun's oily friendliness towards the OSS world. I'll wait and see just how "Open" they mean by Open Source, and maybe then I'll reconsider. If Sun run OpenSolaris on opensolaris.org, for instance.

Mommycalled: "What's different between NetBSD kernel development team being an email away and the

Solaris Kernel development team being an email away. Bzzt NOTHING. Besides I don't

have to deal with Theo de Raadt"

Theo is OpenBSD, not NetBSD. And yes, while he does froth at the mouth a bit, his contributions can hardly be considered valueless.

There is *no* difference between the reachability of the kernel teams, silly! That was the point - one's free OSS and one's commercial OSS; both provide a complete core system and neither is a packager of outside stuff, at least not exclusively and in the sense that RedHat are. Don't you know how much NetBSD is putting into standards-conformance in their shiny new base system and tools? Don't you know that NetBSD tools are now in Apple userland? Silly, silly, silly!

"I guess releasing the Solaris

source into the wild is nothing."

The licenses have to be agreeable and Sun has to *mean* what it says. Giving bits away gradually, keeping back premium features while exploiting the community, etc, etc, are not how you show your openness. It's like MySQL - the source is certainly there, but no commitment. It's not FOSS - not really. I know that's sort of nebulous, but Sun want to undo the hurt they've done others before they can be trusted. And as long as Linux has the real market leader, you'd be silly to deny that Sun craves a piece of the spotlight. OTOH, they're certainly doing their best, especially recently. But it's not always easy to say it's been about mutual benefit.

http://www.theregister.co.uk/2007/06/13/schwartz_v_torvalds/

Cheers,

Sabahattin