The leading OpenBSD programmer, Theo DeRaadt, has accused Sun Microsystems of hindering development of the open-source software for its newer computers, causing Sun to scramble to cooperate with the project in response.
Sun Accused of Shading Open Source
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Why does Mr University of Alberta want to buy a 2 to 8 processor machine and run an operating system that can’t use more than one processor?
I am fully behind Theo and his desire to get docs from Sun, but I don’t see the point in soliciting quotes from a guy that wants to run a uni-processor only OS on an 8 way box.
Sun is far from open source. Even their crypto contributions have had all kinds of hidden licence restrictions that are just ridiculous. Either open source your stuff, or don’t bother. Half assing it doesn’t help you or the open source community.
“Why does Mr University of Alberta want to buy a 2 to 8 processor machine and run an operating system that can’t use more than one processor?
Oh, it gets worse than that. The article claims:
University of Alberta’s Bob Beck said he is forced to buy out-of-date UltraSparc II-based E450 servers instead of newer UltraSparc III-based V880 machines for the university’s SunSITE software exchange.
Well, that’d make you think that SunSITE is running OpenBSD. But it’s not… it’s running Solaris. I don’t see what’s precluding them from *purchasing* UltraSparc III systems if they’re not even running OpenBSD to begin with.
“I want to run OpenBSD on (UltraSparc III systems), because I can do stuff with it for routing, security and traffic control that I cannot do even with a $100,000 Cisco box,” Becker said in an e-mail interview.
This certainly begs the question of why they’re purchasing Sun hardware if they don’t intend to run Solaris. Solaris will always be leaps and bounds ahead of any open source operating system on Sparc. Even if Sun were to release the Ultrasparc III docs, OpenBSD will only function in a rather crippled manner and will not realize the full potential of the hardware, at least for many years to come.
If he wants to run OpenBSD, perhaps he should invest in some x86 servers.
This is just another case of Theo de Raadt raising hell because he’s not getting his way.
This is nothing new, the Plan9 port to Sparc was abandoned after Sun denied
important info to BellLabs(Plan9 supports MIPS, ARM, x86, Motorola
MC680XX, Alpha, PowerPC and a few more arches) Ironically Sun was quite helpful
with the Sparc 64 Linux port (IIRC they even donated some HW to Debian to help
with the effort).
Even more ironically if you think that without Unix, that BellLabs invented, Sun
would have never existed.
Sometimes I wonder how Sun holds together with so many internal conflicts and
philosophical contradictions.
Just another example: they invest considerable resources in helping OSS
projects like Gnome and Mozilla(and don’t make any fuss about it!), and at the
same time they give Apache Jakarta(the only thing that keeps the java community
alive) all kind of trouble to implement JCP specs and ignore their
(superior) projects when adding functionality to the JDK(eg Log4J, RegEx, etc),
yay for NIH.
Not to mention that they have too many lawyers(I calculate that they have 4
lawyers for each programmer)…
Anyway, sad to see so much talent wasted… I just hope that some day they
give up their NIH complex and learn from their past mistakes.
/me dreams of the day that Java and Solaris are *really* opensourced.
\K
“Why does Mr University of Alberta want to buy a 2 to 8 processor machine and
run an operating system that can’t use more than one processor?”
Maybe because it’s the most secure operating system that exists?
I agree that probably Solaris has the most advanced kernel of all *NIXs, but
it’s security record leaves much to be desired(specially if you account for the
whole OS and not only the kernel).
And for the cheap crapy IA32 option: there are things you can do with
OpenBSD+Sparc that you can’t do with OpenBSD+IA32(let alone hardware
reliability)
Eg.: non-executable heap.
\k
Right well if the U of A buys a 2 proc UltraSparcIII there isn’t really _that_ much wasted. One CPU, so what? NetBSD has basic SMP support now (from what I know) and since OpenBSD is very closely related to NetBSD it can’t possibly be _too_ long before this code trickles into the OpenBSD source tree. This of course is an assumption.
Also — for those of you that don’t know — The OpenBSD project has strong ties with the U of A. The U of A hosts the OpenBSD website & ftp stuff. I’m not entirely sure of all the details of their relationship but I can speak with first hand experience that the U of A does make heavy use of OpenBSD. For example when visiting my sister, who lives in residence at the U of A, I was surprised to discover that they log into an OpenBSD machine (over the lan using Putty) to authenticate for their internet.
Just thought you might find that interesting. I don’t know how OpenBSD use in other Universities/Colleges compare?
Even Java is no real open source yet:
To my best knowledge, there is still no native FreeBSD version of Java available as binary distribution.
One still has to download the original source tarball from Sun and to download a patch set for FreeBSD, both after clicking through some legalese.
Regards,
Marc
Well Sun has certainly right about requesting every other company to be open, but they should better act on this openness themselves.
Right well if the U of A buys a 2 proc UltraSparcIII there isn’t really _that_ much wasted. One CPU, so what?
Well, let’s see, he wants to run OpenBSD on a SunFire V880. With the minimum configuration that’s a $37,000 computer. (2 * 900MHz UltraSPARC III, 4GB RAM)
He’ll also need support for the FCAL controller (see http://www.openbsd.org/sparc.html#hardware for supported hardware)
Even if Sun were to give OpenBSD the information they need, it will be a long time before OpenBSD will even be able to boot on a V880, let alone be ready for use as a server.
I agree that probably Solaris has the most advanced kernel of all *NIXs, but it’s security record leaves much to be desired(specially if you account for the whole OS and not only the kernel).
Bah, security issues arise from the services the admin chooses to provide.
Obviously if an admin is choosing OpenBSD over Solaris the phrase “secure by default” means jack shit as the admin is deviating a great deal from the default system configuration. Hearing people say “I’m going to use OpenBSD instead of Solaris on my V880 ‘cuz Solaris gots a worse security record” makes me cringe.
There’s no reason why the admin couldn’t install Solaris without any of the default services and build every service he wishes to provide from source.
What it comes down to is: If you’re buying Sun hardware, you sure as hell better be running Solaris if you want to get your money’s worth.
Right well it doesn’t make any difference either way.
The fact of the matter is Sun isn’t giving the OpenBSD project documentation which they _have_ given to other linux projects.
Do you have any idea how much money Universities throw at stuff? A $37,000 computer is NOTHING when considering the grand scheme of things. Chances are putting OpenBSD on it right now wouldn’t be a great idea — but that’s not the point — the point is there is a demand for it. How will OpenBSD ever boot on that bloody machine if the OpenBSD developers can’t get a hold of the docs? I believe Theo even mentioned previously that he had intentions of getting one of these machines for the project.
Sun is just being stupid.
We _know_ that article has mistakes, and we know that it was angled in such a way as to encourage the reader to agree with the OpenBSD side — so what? Your 1 or 2 points don’t hold any water because they’re just _your_ opinion. How do you know what the U of A is going to do?
… then, they should open the source of the apps they bought from lighthouse design ! more infos on :
http://www.petitiononline.com/laafs/petition.html
It cracks me up every time I hear people hollering at Sun about not dropping its pants everytime open source is mentioned. I’m an open source advocate but I’m also pragmatic. I look at company the size of Sun and see they have contributed a whole bunch of things to the community at large from NFS to Open Office (sun spent over 500M to purchase Star Office and then contributed to code to the community). Of course with the turmoil in the economy and just simple fact of life in a large corporation they are not going to be 100% every time on open source issues but are they to be pissed on for misstepping on 1 out 10 opportunities? Folks need to get a grip and encourage Sun, IBM, and anyone else to open things up and show them the benefits. I can’t see why Sun – or anyone else for that matter – is going to be upset if only 10 people can’t run OpenBSD on their systems.
Yah. I agree that sun should provide the docs for their hardware.
I don’t agree with a shoddy article. The article would have been substantially better without Mr U of A’s quotes.
And yes, I am fully aware of the U of A’s openbsd environment. I know numerous people that attend the university, and I know they make heavy use of openbsd. I just don’t think it is feasible to run it on an 8way server.
Anyway, lets hope that Theo gets his docs, and that they are also able to get smp going within the next couple of years. I would honestly prefer to see smp on that 8 way machine that that crappy intel hardware
“Do you have any idea how much money Universities throw at stuff?”
That doesn’t matter because you’re going to tell me regardless!
“A $37,000 computer is NOTHING when considering the grand scheme of things.”
Well gee, guess what my job is (look at the source address of my previous posts for a hint)… why, it’s the same as Mr. Beck’s. I’m a system administrator for a state university. So unless they do things radically different in Canada, $37,000 is a shitload of money.
This Bob Beck guy is obviously in their NOC, and most likely he had a part in administrating the SunSITE server (running Solaris) which got hacked when the OpenSSH distribution got trojaned.
So let’s see, the only person Theo the Rat can find to advocate the use of OpenBSD on Sparc is a guy who admins the site where OpenBSD is distributed, and furthermore one who is partly responsible for OpenSSH getting trojaned. That’s not a very convincing argument Theo.
Your 1 or 2 points don’t hold any water because they’re just _your_ opinion.
Umm, let’s look at the facts then:
Fact: Mr. Beck stated in the article he wants to run OpenBSD on a SunFire V880
Fact: OpenBSD will not boot on a SunFire V880
Fact: Even if the OpenBSD developers had the UltraSPARC III documentation and added support to OpenBSD, OpenBSD would not boot on a V880.
Fact: The UltraSPARC III is well documented in *code*, but Theo is whining that that’s not good enough.
Fact: If you work for a NOC and request a $37,000 machine, you sure as hell better be able to justify your purchase.
“How do you know what the U of A is going to do?”
I’ll tell you what they’re not going to do: run OpenBSD on a SunFire V880 in the forseeable future.
Provided Mr. Beck is serious and isn’t just being a de Raadt mouthpiece, I suppose he’ll have a small server farm of V440s which will make about as effective a server collectively as a single V880 running Solaris.
Did anyone of you whining monkeys read “I want to run OpenBSD on (UltraSparc III systems), because I can do stuff with it for routing, security and traffic control that I cannot do even with a $100,000 Cisco box,”?
that means:
1) He can’t use Solaris for the same tasks.
2) He can’t use as a server farm of instead of one server.
Did anyone of you whining monkeys read “I want to run OpenBSD on (UltraSparc III systems), because I can do stuff with it for routing, security and traffic control that I cannot do even with a $100,000 Cisco box,”?
No, here’s how I read it:
“Hi, I’m the admin of the FTP site that OpenBSD is distributed on. Theo told me to say that if only OpenBSD supported UltraSPARC III, I could get rid of crappy Solaris and do all sorts of neato security stuff on my servers”
And apparently you’re missing the point: He could still buy SunFire V880s because they’re using Solaris on SunSITE anyway, not OpenBSD.
Of course the V880 can run Solaris — and of course if he buys one he’s going to leave Solaris on it. That’s not the point. What part of this simple equation do you not understand? Isn’t it already accepted that the comment was said merely to leverage the argument in Theo’s favour? I’m NOT debating that. Is it misleading journalism? Yes. But does it matter? Who cares, the end result is that it pressures Sun to give the OpenBSD guys the docs — Inturn this results in faster OpenBSD support for the V880 — Which is a good thing for everyone right?
You keep tripping on the minor details and fail to see the big picture.
I’ll fact you:
FACT: Theo wants SunFire V880 OpenBSD support
FACT: Theo previously commented that he was intereted in purchacing one of these machines for the OpenBSD project.
FACT: OpenBSD does not run on the V880
FACT: Theo wants OpenBSD to run on the V880
FACT: Theo cannot get ahold of enough docs to make OpenBSD run on V880 (fair enough, no? giving the docs to the OpenBSD project isn’t going to cause any damage to sun is it?)
FACT: Theo gets pissed off because the same docs he is after have been given to some Linux projects
FACT: A somewhat misleading article is written to make sun look bad.
Chances are the U of A guy _does_ want to run OpenBSD on the v880 _sometime_ in the future, but that’s never going to happen if Sun doesn’t just give Theo the damn docs to make him happy.
This is a stupid argument. It’s called government funding, maybe that doesn’t happen down in the US?