Post a Comment
Well I believe there's something of a misconception among people that HP/UX, Solaris (to a degree), IRIX and AIX are all for the most part dead. Obviously Solaris isn't, and HP/UX and AIX are still put to use in a few big iron environments and still have some market share. GNU/Linux has displaced UNIX in many high-volume uses but it certainly isn't dead.
Indeed UNIX is not dead we even see file servers that were previously Linux moved over to solaris 10. Sure support for Solaris is more expensive than most Linux Free forms etc but you can get support for any platform for free or close to it if you know where to look.
Unix add's value where linux distros normaly cant afford to Big Iron why? They don't make the hardware... Linux still has people complaining about NV still not open sourcing their IP Rights. Traditional closed source don't care what is contained in those drivers as long as they have good performance which is also the aim of the manufacturer.
I must admit that AIX is heading for a path that SGI know well.... It needs a community and thats being stolen by the other OS's
I do not mean to be critical, but this needs to be said. The management behind the original SCO became Tarantella. SCO was a damned good Unix. Absolutely tops on x86 at the time. The CEO made a nasty public swipe at Linux, back in the day. But to his credit, he quickly learned from it and apologized. They decided that their Unix days were numbered and went into a sort of virtualization effort. Caldera, a Linux company for whom I used to have some respect, but associated with the very shady "Canopy Group" purchased the OS division of SCO, hired a new upper management staff, and the rest, as they say, is history. We should not tarnish the name of SCO. We should tarnish names like Darl McBride, Chris Sontag, and Canopy Group. Canopy Group was pulling the strings. And Darl was their front-man. And a very willing one in my view.
SGI is doing some really impressive work, with Linux, on SSI supercomputing systems. Sun is adapting Solaris to this new day and age.
Yeah, I'm a Linux fan. But, honestly, I'm a Unix fan first. Linux and the *BSDs have been the unifying forces that Unix has needed for so very long. It pains me when we still fight amongst ourselves, now that we have a chance to do something wonderful... again.
Edited 2008-10-23 17:00 UTC
+1
I don't even think Darl was pure evil. When he first became CEO the stuff he was saying is what you would want to hear from the CEO of a major Linux company in this day and age.
One of the things I like about your posts is that you are a linux fan that actually knows some history. It is interesting to see how a company could start with so much street cred, and end as one of the Great Satans. Or how a bunch of evil bastards like IBM could completely turn around their image with unix geeks the way they have.
Not a company. A name. "The SCO Group", commonly and erroneously referred to as "SCO" today, is a very different company than it was before. Not the same people at all. My recollection is that Darl did, indeed, start out as completely evil. It was Ransom Love whose comments were, perhaps, somewhat debatable. He was much hated by the community. But I tended to agree with his comments regarding the respective benefits of, say, BSD vs GPL licensing. Darl was no Ransom. And yet the new SCO was *all about* ransom. And very mistakenly so. Yet the old SCO... and by that, I mean the *really* old SCO... the one that predated Caldera, Ransom, and ransom, was pretty cool. The original Caldera was also reasonably cool. The current incarnation is the completely evil one. But is also, in the immortal words of the beloved Douglas Adams... Mostly Harmless. ;-)
Edited 2008-10-23 19:58 UTC
A very long time ago, yes. Before my time, actually. Which made for a lot of"What the hell is this!?" glances from me 20 years ago when I was just getting started with SCO Xenix, as most of the shell scripts began with an MS copyright.
What, exactly is it that you disagree with regarding Ray Noorda?
None of the rest of your post is worth responding too. My only question there is how you can stand to be so very hateful. Hateful thoughts drain the life out of me. I find it far more pleasant to avoid them. You have done a great deal to damage the reputation of the Linux community here on OSNews through the expression of your hatred of BSD. Possibly more than anyone else I can think of. I have tried to reach you in the past, through public and private communications, through reason. I've exhausted all the reasonable strategies. So perhaps it is time to resort to an *unreasonable* one:
You can stick your head up your ass, Moulinneuf. Or, actually, I think you already have.
Edited 2008-10-24 02:05 UTC
" We should tarnish names like Darl McBride, Chris Sontag, and Canopy Group. Canopy Group was pulling the strings. And Darl was their front-man. And a very willing one in my view."
False on both account ...
I cannot help but feel that you have something significant that you are trying to say here. But perhaps you should be a little more explicit. Are you condemning me for having had some respect for Ransom Love? The "Old Caldera"? My original post was about a very different company. I'm not sure I really care to argue over Caldera. They were never my *favorite* Linux company. Just one that I had some respect for a long time ago.
I *think* we're still kind of on topic here. And apologize to those who may reasonably disagree.
It's is explicit and perfectly understandable , you said that SCO (santa cruz operation ) was not responsible for taking out Caldera as SCO and that the "Shady" Canopy group was responsible.
Neither are true ... But then you got that part ...
"My original post was about a very different company."
NO , the Santa Crux organization and SCO are the same company ... Not with exactly the same people , but still the same company.
"I'm not sure I really care to argue over Caldera"
There is nothing to argue here it's recorded history.
"They were never my *favorite* Linux company."
So for that your trying to change SCO recorded history ? Your loyalty and former liking of the santa cruz organization might be seen as admirable for others , but your history rewriting as no place here.
"I *think* we're still kind of on topic here. "
You don't think very hard here ...
"And apologize to those who may reasonably disagree. "
Disagree with history ? No they insanely disagree with reality that GNU/Linux won by itself and that Unix is not GNU/Linux as not helped it now or in the past , neither did BSD and that there not united.
Some people including yourself have trouble with people who are gonna disagree with them. Even more so when those people introduce a reality they don't like.
It's M o u l i n n e u f : My real life name ... but then coward like you have been here long enough to know that. Must be your copy paste that is broken again ...
"link to Moulineuf's postings "
http://osnews.com/user/uid:2598/submissions
VS
http://osnews.com/user/uid:2598/submissions
BTW what's so shameful with Lars Hansson ?
"Number of Comments: 2496 VS Number of Comment Votes Applied: 1200 "
Wow You spend your days commenting here ? You Seem to comment just as much as you vote on people.
"Moulineuf is so ridicules he is a self parody"
Real man have no trouble ridiculizing themself , not that I am ridicule here , because you never *laugh* at my ridicule post ... You don't get that there is more then one point of view to every discussion.
"Awhile back I used to try and enumerate all the OSNews terms of use that he was violating in his posts "
As I said then and now , None , the fact you don't get it is rather interesting , some people are "special cases" and don't have to follow the rules other have to at all the time. I am sure you did the math and are still shaking your water head that I am still able to post here ...
"personal attacks, inflammatory language, etc) but he just doesn't get it."
How can I be personal with a coward named Google Nina ?
"but he just doesn't get it."
Am I not getting it or are you not getting it ?
It is rare that you do not personally disparage the person you are talking to.
II. No personal attacks on story authors, other readers, or news editors of this web site.
Not only do you disparage them, but you tend to attack them to.
III. Even if you are in violent disagreement or have strong feelings, find a way to keep your comments calm, and try to explain your reasoning, instead of just ranting.
I don't think I have ever read a comment by you which isn't you flipping the hell out.
You troll every single bsd story posted on this site
According to the rules, the Admins should have banned you ages ago. Many people will break one of those rules here or there, I don't think anyone else breaks four every single time they post.
http://www.osnews.com/user/uid:266/comments
reality disagree with your lies ...
http://www.osnews.com/user/uid:266/comments
reality disagree with your lies , again ...
Something we can both agree on ...
It's seem the problem is you here , I post and I am always flipping out according to you ... analysis show I am never flipping out.
http://www.osnews.com/comments/19947
http://www.osnews.com/comments/20031
http://www.osnews.com/comments/20104
This very Article and thread
Title : "Linux's Next Victim"
http://www.serverwatch.com/eur/article.php/3779821
not a single BSD in the article yet : there is above 10 BSD mention without counting mine ...
None of witch I Broke ... Maybe there is a page 2 and 3 and 4 to the rule that you don't know about.
The reason : because Google Ninja ( or insert other name here ) the coward don't like Moulinneuf and is comment apparently is not a valid one.
I agree you don't think , if your interested in the banned , suspended and other statistic contact the site admins ...
Edited 2008-10-24 17:03 UTC
SCO was a good UNIX in a time where serial terminals ruled. A i386 SCO server with a bunch digiboard serial connectors could serve a company of around 256 people.
But like Microsoft they were overrun by the internet and missed the boat. Unlike Microsoft they hadn't the money and power to catch on. From mid-nineties on, SCO was a farce.
When Windows will be gone, probably computing as we know it will have changed fundamentally.
If Microsoft were to go bankrupt today, Windows would be alive even in 2020 (at least). Reverse engineering, other companies buying Microsoft's intellectual property, etc.
Windows' legacy is huge, even though some people try to deny the obvious 
While I don't see Unix disappearing, its role has clearly diminished. The place I got my first sysadmin gig at in 1997 started out running purely Unix, a year later a couple of Linux boxes had been added. A few years after that again it was all Linux except for one Unix server. So while they still needed a Unix admin, they needed him because of that one server as opposed to the entire shop.
There's a difference between being able to support a platform and really knowing how the platform works. I mean I've worked with at least a half dozen different *nix variants in my career, but I wouldn't claim to know the ins and outs of all of them down to the last detail.
They didn't need an Admin to only support that one Unix server, but they did need someone who knew Irix (in this case), and could also help with supporting Linux. Not all Linux admins have experience with Irix. And while it is true that most things are similar enough that a Linux admin could work out many things about Irix (I know I've done it), it's always better to have someone who actually knows what they're doing.
If I had an important Irix server and was hiring an additional sysadmin, I'd choose an experienced Irix admin, who could also help out with running the Linux servers over an experienced Linux admin who could teach himself Irix on the job.
A short statement about job offers, just from my individual point of view: There seem to be more UNIX than Linux offers. This is because (I think) employers have a strange opinion about what's the difference between Linux and UNIX. UNIX is complicated, very technical, CLI driven; it needs a proper admin to be run. And Linux is user friendly, the 14 year old son of the boss could install it. :-) Okay, fact is: In order to run a system correctly, you still have to know many things, even in the age of "user friendly Linux". (Just to avoid misinterpretation: I always thought that Linux is user friendly.)
Beside the implication where usage share shrinks and where it grows, if you have fundamental knowledge about UNIX, you can benefit from it administrating any Linux. Those who administrate different UNIXes already know that there are differences between them, such as they are between UNIX, the BSDs and the many different Linusi.
Having said this, even if the "commercial" UNIX variants would completely disappear, the knowledge and experience they taught to today's admins will still live, and make their life easier.
Wow, you mean you went from one hardware architecture to a completely different one and it was faster because your software was better suited to run on that new architecture? Shocking!
I also find the cost argument amusing since you can run Solaris 10 for free -- you can't legally do that with RedHat Enterprise Linux.
"In order to use the Solaris 10 Operating System for perpetual commercial use, each system running the Solaris 10 OS must have an entitlement to do so." - Sun Solaris License Legal requirements.
. . . .
And you get your entitlement sent to the email address you used when you registered to download Solaris 10 off the Sun website...for free (as in beer).
My shop was a big Solaris HPUX shop for a long time. Linux and Free/Open-BSD did small jobs though.
Now we are going to Linux in a big way.
There is the impression that Linux is cheaper than Solaris but that is perception more than reality. Red Hat licenses are expensive and Red Hat support is more expensive than Sun OS support. As much as Linux has improved from an administrator's point of view I still like Solaris better. Solaris seems more robust, has a better performing IP stack, it has ZFS, and like I said is actually cheaper in a commercial environment when the license and support costs are compared.
One of the last big projects that popped up I suggest that we do a proof of concept on both Linux and Solaris to see which performs better. We could set up 2 virtual machines easily and know which works and performs better with very little investment of time or resources. I can't remember the last time one of my ideas were shot down so quickly. These days it seems that if "Linux" isn't the right answer you aren't asking the right question.
I don't want to seem impolite, but some years ago, this statement, having only replaced Linux by "Windows", has been considered completely true. If you said words like Linux or BSD, people would have looked at you as if you came from another planet. You can achieve this result today by saying Solaris, or z/OS. :-)
But I think this has to be seen within the development of time. In Germany, there's still the belief that nothing exists outside "Windows", and Linux is still a kind of geek toy. But slowly, those who run server stuff have recognized the advantage that Linux can run fine where others still run commercial UNIX, and "Windows", of course.
In my opinion, there's room for everything. For many requirements, Linux is an excellent solution, but there are applications where I would prefer Solaris or one of the BSDs without any further question. This is even true for servers.
"In order to use the Solaris 10 Operating System for perpetual commercial use, each system running the Solaris 10 OS must have an entitlement to do so." - Sun Solaris License Legal requirements.
Ah, but the entitlement IS free. So you're wrong :-)
Unless RedHat has changed their license since I was a subscriber, that isn't true. They told us that unless we stripped out their copyrighted material (which at the time consisted of several specific packages) we couldn't continue to use it.
By contrast, you'll every major release for free from Sun and security fixes in the meantime.
So again, you might want to look closer :-)
For those that want something that is completely open source (for all the parts Sun owns) but is fully redistributable, look here:
http://www.opensolaris.com/
...commercial support is coming soon too.
1) Solaris is no longer propietary, and unlike other historic unix operative systems, it's actively developed.
2) UNIX does have some features than Linux doesn't have — single-system RAS capabilities and the ability to deallocate failing processors online, for example."
Linux has been able to deallocate failing proccesors online for a long time (I can find commits about cpu hotplug from 2004), see /sys/devices/system/cpu/cpuX/online
Well we've seen 11.31 and see the technology previews of 11.4x. And then we also see linux.
Most of the times, when there is no other way, we still use HP-UX. Howeber, being BSD based, having for instance nog recursive grep etc....
We always install gnu tools, bash, gcc, beause the base system is crippled. We also see that 11.31 is somehow buggy. Unless we have to, we switch over to linux. SLES 10 to be more precise. Much cheaper. Still on HP hardware though (C7000 enclosures w/ BL4/8 series).
For the PA-RISC stuff we keep HP-UX.
Another consideration is the media of HP-UX, will cost you over $800 for the media, with some s/w assurance.
The hardware is OK, HP-UX is.... well IMHO dying. Not of this century (like most BSDs aren't, like solaris)
Pity UNIX is dead as I liked the idea of running ZFS without a need for FUSE.....oh hang, plenty of datacentres are running ZFS.
Ah well, at least the article was right about Linux killing any chance of UNIX running on the desktop.... except of course if you own any Apple products.
In short: What a bullsh*t flame-bait of an opening line.
Sigh............. Are they? I'm afraid you'll find that people do not start formatting existing disks and partitions in somewhere like a data centre to put in a new filesystem where everyone around Sun is jumping up and down saying "Look, this is so cool you have to rip out all those Linux, and hopefully x86 systems, and go back to the very Solaris, and hopefully SPARC, systems that you left behind after the dot com boom because it cost less and you got more performance". It's not enough.
The article is simply old news and Sun has suffered badly at the hands of Linux and x86 over the years, but the fallacy for Linux companies is that they can keep seeing Unix as the 'low hanging fruit'. There will still be Unix systems around, but at best, it is a stagnant market.
I have a stupid question: How is Linux killing UNIX?
The last time I checked the UNIX trade mark is owned by the Open Group (www.opengroup.org). Any system may use the UNIX trade mark if they pass a Certification test.
The newest UNIX on the market today is Mac OS 10.5 which is a certified UNIX. It has the same certification as AIX 5L V5.3, HP-UX 11i V3, & Solaris 10 (the UNIX 03 Certification).
Apple had to change a few of the GNU tools to pass the certification; however, I believe they sent the patches back to the GNU team.
If anyone wanted to spend the money and time, it would be easy for GNU/Linux to pass the certification test. Please note that the Linux Standard Base Specification 2.0 and the Single UNIX Specification 1.9 are in the process of being unified.
May I ask a question? Overall, how is this sub-thread likely to benefit the world? I, personally, do not see how it is likely to. These Linux vs *BSD threads are, in my opinion, non-constructive. They hurt all communities involved. We need to learn when to ignore any perceived trolling from whichever side we consider to be the "other" one and remember that we are, essentially on the *same* side. And walk away from these silly and useless exchanges.
Edited 2008-10-23 21:37 UTC
RE[5]: Why Linux vs UNIX?
... BSDs and commercial UNIXes are still strong in the servers world....
... and in the desktop world, Linux has 1% market share approx and Mac OS/X (a true UNIX), has a 7% market share...
... so, I really do not know if Linux is killing the other UNIXes around... and being a little sarcastic, the only true Linux victim until now has been Minix...
Do not misunderstand me, I like Linux a lot (I adore Gentoo), but saying Linux is killing the other UNIXes, when in some markets the other UNIXes [MacOS X and iPhoneOS] have a better marketshare does not make sense at all.
Calling OS X a true UNIX is a bit of a stretch. They bought the certification, most GNU/Linux distros could too, if it made sense. The underlying OS is an even worse mix mash than most Linux implementations, not that average users care about that, Apple scores big on the stuff that matters to Joe Blow.
To be unix certified it's not just a case of purchasing a certificate from the nearest unix cert reseller, the OS has to meet certain standards and i find BSD/Mac OSX a little neater than Linux when it comes to layout.
The unix cert is not really for the desktop users, from what i can see it's forethought in apple positioning itself to sell more and more of their server software and hardware to businesses.
Agree with the other posters, this article is pretty poor, from what i can see unix will be around for a very long time to come, yes linux has got into certain datacenters, however there will always be a place for Unix, like theres a place for Windows Server.
What i mean by adding in Windows into that line is that it's annoying to see the age old argument that for one OS to win the others have to fail. This has been the argument on the desktop for a long time. However as many have said including Jobs, Gates etc.. computers are prevalent enough that there is enough room in the market for these systems to continue to thrive without the death of another.
OSX is very much a UNIX! Just as much as Sun OS, AIX, HP/UX...
Having worked on kernels for many years! there are _no_ OS's out there that I know of that are more of a cluster phuck! than the Linux kernel. Including FBSD, Darwin, vxWorks.
Linux is what you get when you have too many cooks in the kitchen! Spaghetti everywhere!
KRR
RE[3]: I do not want to disappoint but...
uptimes? What uptimes? If no one is using the OS, does it matter how long it will stay up?
I find it both funny and sad to see BSD zealots defending their turf against linux. I actually like the BSD's. They have some remarkable features. But to sit there and bash Linux when so much of the source code between the two OS's is shared is pretty silly. KDE is KDE no matter what you run it on.
I find it both funny and sad to see BSD zealots defending their turf against linux. I actually like the BSD's. They have some remarkable features. But to sit there and bash Linux when so much of the source code between the two OS's is shared is pretty silly. KDE is KDE no matter what you run it on.
See Netcraft uptimes.
KDE 3.5.x is great on both! 4.x sucks rocks! It's completely unusable IMO and has a terrible UI. Too much eye-candy and flash that gets in the way of doing work! besides KDE is not part of an OS IMO.
One of the reasons the documentation sucks is that everything in Linux is a moving target! Developer knowledge only lasts a few months before it gets stale! BSD's documentation (and OS) is much better as their development cycle is better defined and the additions to the OS are part of a well defined long-term strategy!
KRR
Those who work as developers sometimes get upset when searching for proper documentation in Linux. Sometimes, there's a manpage, usually not. The source code isn't as tidy and explaining as expected, in regards of useful identifiers or comments. That's different from what is usually known from the "commercial UNIXes" and the BSDs. There, you usually get excellent documentation, either it is already there (locally) or you can buy it. For example, the BSDs have one of the best documentation philosopies I've ever seen: great handbooks, FAQs, and manpages for everything, not only for the system commands, but for configuration file layouts, kernel interfaces, system calls, maintenance procedures and device drivers. Everything is consistet. But that's usually not the case with the many Linusi and especially with the "two big desktop environments" where documentation is left to others, scattered around in forums, Wikis and even Blogs.
At least to me as a developer, documentation is one of the most important things. I hope that Linux will improve in this regards, but I fear that the attitude "It already works, why write documentation?" will become more and more comfortable...
The Linux kernel now exceeds 10 millions lines of code, including comments, empty lines, etc.
If you only count source code, then Linux kernel is 6.4 million lines of code.
I vaguely remember that entire Windows NT was 10 million lines of code? Is that correct?
Is linux kernel growing like mad? Is it difficult to have one monolithic kernel with 6.4 million lines of code bugg free? Hell yes. Is it not as stable as it could be? Yes.
Read what Linux kernel developer says about buggy Linux kernel:
http://lwn.net/Articles/285088/
If you only count source code, then Linux kernel is 6.4 million lines of code.
Don't get me wrong, please. I didn't want to focus on a kind of "BSD vs. Linux kernel documentation" approach. It's just about my individual observations, and it's far more than just kernel, or system. Have a look at "modern" applications like all the KDE stuff. Try to "man kde" (or a KDE related program), or try "man grip" (for a Gnome related one); in opposite, try "man opera" or choose some arbitrary kernel interface, library function or device driver. On a BSD system, you're usually lucky. Then, have a look at the system's source code (and not only the kernel one's): The source is very tidy and quite good to read. This is something I really applaud: Taking the time to document what's already working, instead of "just let it work and let others write some documentation on the Web".
I don't know about "Windows", I have none. :-)
I wouldn't try to say anything else. In regards of functionality, Linux has walked a long (and good) way to get POSIX and other compliances.
http://lwn.net/Articles/285088/
Thanks, very interesting!
"... and in the desktop world, Linux has 1% market share approx and Mac OS/X (a true UNIX), has a 7% market share... "
In the US, and only in the US. Worldwide, Linux and OS X each have a market share of around 1-2%. For example, where I live, Apple penetration is very low, and online statistics for the largest sites here show OS X usage less than 1%, and even lower for Linux. I wouldn't bet on high Mac OS X usage in China, India, or other developing countries
Apple is pretty US focused. And the rest of the world DOES MATTER, despite what americans might think (that's why Acer for example is one of the largest PC manufacturers) 
The article is referring to a case study of Sabre, if you had cared to read it, or even just the summary.
For lots of applications x86 platforms surged ahead of SPARC and other platforms in terms of performance and people realised this around eight years ago. I saw a 1.4Ghz desktop Athlon thoroughly cream an UltraSPARC III running Zope at an organisation at the time, and that was a sign that they'd move. Despite saying they never would, two years later they moved to a x86 server and Red Hat because they had no choice.
PowerPC was generally insulated from this because it was a high-end niche and used in different workloads, and still is, but Sun sold a lot of kit to dot com companies and in areas where x86 and Linux then took over. They refused to put Solaris on x86, or improve it, because it harmed their hardware sales. However, in the long-run it is debatable whether the costs of maintaining AIX and HP-UX can be justified even in areas where such high-end platforms are sold.
Was the zope stuff you saw related to the Zope on Solaris stuff I did?
http://www.zope.org/Members/glpb/solaris
This was why I was so excited when Sun started x86 servers as it meant you could get x86 performance for things like this, yet a Sun logo on the front to keep them happy
-Matt
http://www.zope.org/Members/glpb/solaris
Probably related to it, but not that specific instance. There was mumblings for quite a few years slightly before and after that, but neither the Zope nor the Python developers had the inclination to try and work out what was going on. How could they? They didn't have access to SPARCs and weren't particularly interested in troubleshooting what was a closed OS, and still is in many areas. They wrote code and compiled primarily on Linux systems, and Windows ran fine. I seem to remember a Sun consultant's answer somewhere was to compile with Forte. Not helpful.
I do hope OpenSolaris will help with this and Sun will put more effort into getting lots of open source code out there to work. It's been a long time coming.
Yes, at last ;-). The trouble was that it has taken a long time to trust Sun's commitment to Solaris on x86. I still haven't forgiven them for trashing Cobalt and their Cubes, and it is disappointing to think of the success they could have had in that area and the money to be made. Many small businesses had those systems that could never have afforded Sun kit.
This is plain FUD.
"3 fold speed gain" means that they have thrown out a couple of 10 year old SUN servers and instead brought in 5 times as many cheap state of the art x86 servers. I suspect.
I remember reading an article where they said the same thing "Our Linux solution is much faster than our Solaris solution". Upon closer scrutiny, the company had got rid of 800 old SUN servers for 3000 shining new x86 servers.
No wonder you have 3 fold speed gain. Pure FUD. If you run Solaris on the same new x86 servers you would also see major speed gain.
FUD.
1. Linux is as much Unix as any OS from IBM or HP. So saying it's Unix vs Linux is kind of a stretch. It should probably be Linux is becoming #1 Unix. Yes I know that's not 100% accurate, but for most people that pretty much sums it up.
2. If systems using Cobol were still around in 2000 (and even today) you can bet that all these other Unix OSs will be in use for many years to come.
3. I don't think companies like HP and IBM really mind that Linux is replacing their brands of Unix. They still get to sell their hardware with it, and they get the support revenue that goes with those sales. But now they get to share the R&D and development cost with everybody else. Keeping up with hardware drivers and whatnot for your own brand of Unix is not an easy or cheap task, and sharing this work with a larger development community has to be a big win for everyone.
Edited 2008-10-23 17:55 UTC
Decent analysis of the situation. Those who are protectionist of their particular Unix will probably care. Those that aren't probably won't be as long as money is coming in. After all, people use Linux because it is Unix-like and was the way that it was designed.
Linux and x86 will probably replace a few more Unix systems where they can, but there will always be some big systems sitting around running Unix somewhere and they won't be going away. The problem for some companies like Sun is that they are, and certainly were, caught between the x86 Linux systems that people can more than adequately run, and the big iron systems that IBM in particular run that exist in their own market. It's sometimes not clear where they fit.
Unix will never die though, and that refrain has been going on for years. Sadly, it has been coming from some stupid Linux companies who think that they can keep grabbing the low-hanging fruit of existing Unix installations rather than looking at where they should be looking - competing with and replacing Windows Server. Novell in particular should be doing this, because Netware is gradually being eaten and pretending you're a Linux company to try and replace the market you're losing is not a great idea I don't think.
Red Hat is far from free of costs, Sun support prices should be very competitive in most cases
I guess that the money saved are due to other factors than changing the OS alone. E.g. when you change your OS you are likely to fix other things that you have found to be problematic in your old system. It could be anything from changing hardware to changing management procedures.
I remember Linux folks as always complaining about Microsoft spreading FUD about it. It looks like things are different now. The Linux foundation seems to be pushing Linux campaign very aggressively everywhere it can reach. First that Zemlin guy and now this. Another article full of FUD and misinformation comming from the Linux side. I'm beginning to think that these are paid articles.
Linux in present state lacks many enterprise features, like FC stack, IO multipathing a stable kernel API (thats a laugh!) ability to handle high workloads and large number of cpus,... Also Solaris is cheaper than Red Hat now and runs very well on x86 hardware. Linux's success on low end servers doesn't mean that Unix will die. But it's obivous that some people would like it very much.
I also wonder what is wrong with osNEWS these days. I don't see any NEWS in that article. Is there a need to link to every pamphlet there is just because it favours Linux?
I'm sorry to sound negative, I'm just getting fed up with Linux propaganda. Btw I'm a desktop Linux user myself, althougt I'd avoid it on server.
I'll not comment on anything else that has been beaten to death already, but I do agree with that to a certain degree. The replacement of Unix systems with x86 and Linux is definitely not new at all, it has badly affected companies like Sun over the years but I do think many Linux-using organisations just need to drop it and move on.
Replacing Unix systems like Solaris has been the refrain from Red Hat and other companies like Novell for quite a while because they think that it's been easy money, but I believe that has now just about run its course and they need to start dropping it. Being a merely a Unix replacement is just not good enough.
If anything, you can take this kind of article as a sign that Linux organisations desperately don't want to try and compete with Windows Server, for reasons best known to them, by flogging a dead horse.
Miracle?! Linux's FC support is a nightmare: closed source drivers, binary incompatibility between minor kernel versions, vendor dependent tools, etc.
Linux needs a stable kernel ABI to be "Enterprise ready" and compete with UNIX.
Sure I'm sure! Linux is fully dependant on FC vendor's drivers and tools, which lack performancea and are not as tightly integrated with the system. They can't compete with the support Unices have, e.g. Leadville FC stack and MPxIO in Solaris 10 which are present by default. Solaris also installs with FC drivers. (Qlogic and Emulex at least)
AFAIK, Linux does not provide an API to publish the information in /proc programmatically. If I want to handle such information in my program, I must parse the output and process it. AIX provides a C API to do such stuff.
The Linux kernel API for drivers is not backward compatible.... If I have a driver for, let's say, Linux 2.6.10, it probably will not work in my Linux 2.6.27 ... Solaris drivers written for old OS versions are still working in the bleeding edge builds of Solaris and OpenSolaris.
DTrace, ZFS or HAMMER (the DragonFlyBSD filesystem) are being developed in UNIXes. Quartz and Aqua are developed on top of Darwin, the BSD heart of MacOS X.
So, several great technologies or features present in the UNIXes are still missing in Linux... I do not understand the "tone" of the article showing the superiority of Linux over its "ancient" and "old" ancestors.
Edited 2008-10-23 19:48 UTC
I don't really understand how someone could argument that Solaris is going away.
Just look at the increasing numbers of Software available for Solaris
and the download numbers and community increase ... very constant increasing
http://opensolaris.org/os/community/advocacy/metrics/
And if you are not aware I think the best Sun could do is also more investing in Solaris x86.
I mean don't get me wrong Linux has it's advantages ... but features like
- Zones
- ZFS
- Service Managment Facility
- Dtrace
- binary compatibility from Solaris x86 (old version) to actual version
- binary compatibility from Solaris sparc old 2 new version
- source code compatibility from Solaris sparc and x86
havn't you ever had the trouble with the wrong glibc version ...
just my 5 cent
and looking forward to upcomming Opensolaris 2008.11 with a bunge of new features which are going definetly in the right direction...
http://www.phoronix.com/scan.php?page=article&item=opensolaris_2008...
The "price ZERO" argument is valid for me, a simple person with one computer and no much money... but is not valid for a big company which prefers to pay a lot of money but having support, updates and the warranty their software and ther data will be available as much as possible.
A big company will pay all the needed money to have their info secure and available, that's why the old UNIXes or the enterprise supported Linux [Suse, RedHat] or MSWindows will never lose their place in the enterprise market.
Companies and projects have their ups and downs, it is just business as usual. Things can change a lot in ten years. It is all too early to claim that UNIX would be fighting for its survival, not to mention because of Linux, especially as the challenges UNIX flavors have had are often not related to Linux nor to UNIX quality but to other things.
Big changes in the computer hardware business have caused many challenges: mainframes and many old UNIX platforms have gone out of fashion causing troubles to UNIX. There are also huge amounts of new MS Windows servers that have been eating UNIX market share maybe more than Linux has.
Linux is, of course, growing fast, but if you view Linux only as a new flavor of UNIX, this isn't anything new. Also in the old UNIX days there were many ups and downs and much competition between the various Unix flavors and companies.
UNIX - in its many flavors - has still a very strong hold especially of the high end server rooms, and that is also where the big money often is. For example, many university server rooms still happily run Solaris and that is not likely going to change any time soon.
Linux has gained lots of popularity, new users and developers by strongly emphasizing the open source development model and desktop usage besides of servers. But if you consider Mac OS X a UNIX too, and remember popular new desktop Unix flavors like PC-BSD, there is a lot going in the desktop UNIX front too. OpenSolaris is an example of a traditional UNIX trying to reap the benefits from the open source movement. Time will tell how they will succeed but the first early steps they are taking now is not the time to say the final word about it yet.
The often more centralized and unified development model of UNIX can have its benefits. "Bazaars" (like what the Linux developer community is often said to be like) can be innovative and lively places but tend to be a bit chaotic too and do not necessarily provide the best solutions to everything (especially if it happens to be a "cathedral" that you want to build instead of a "bazaar"). Both "bazaars" and "cathedrals" have their own benefits, pros and cons - there's room for both. Mac OS X, by the way, is an example of a tightly controlled "cathedral" type of OS project and development model.
I also agree that Linux is very much like UNIX itself. UNIX and Linux can coexist, and they could also cooperate much more for the benefit of both - if only there was enough mutual will for it to happen in the both camps.
In the old UNIX days, UNIX lost many good chances it had in order to become the dominant operating system both on desktops and on servers because of too much competition and too little cooperation between various UNIX flavors. I hope that people learn something from history and that a similar attitude in the Linux & UNIX world now won't ruin many good new chances of both Linux and UNIX again.
Edited 2008-10-23 22:16 UTC
Longest uptime
http://uptime.netcraft.com/up/today/top.avg.html
tgpsubmit.persiankitty.com longest time 1620 Days
Next Server 2003
The first Linux comes in outside the top 10 and there is a total of only 2 linux boxes that are in the top 50.
Bring on the boo's :S
Hmmm. While Solaris has it's place I have not heard anyone mention that it lags behind Linux where device drivers are concerned. I think this is one of the major stumbling blocks to it's wide scale use. So the fact that it has features that Linux doesn't is pointless if it doesn't work on your hardware.
OSX would flourish more if it wasn't built for "proprietary" hardware. However, this model has served Apple well so far because it cuts helpdesk requests.
I disagre! It would become much more buggy (linux and Windows like) if it had to support the increased variations in hardware. It has improved stability because it only has to support a well defined set of hardware. Linux does an okay job of supporting the HW but it's a much much larger target and is moving in unpredictable directions. There is almost nothing that can be done about this!
Sabre probably got most of their cost savings by moving to cheaper hardware. How much of the cost of their old system was for the OS and how much for the hardware? In any event, with OpenSolaris now out, the "Linux cheaper than Solaris" argument goes out the window. The only difference will be who's support contract is more - Red Hat or Sun.
BeOS? I think Apple & Mac OS X played a role in the demise of BeOS, and, of course, also the dominance of MS Windows as a desktop OS, and many wrong decisions made by Be Inc. too - but Linux had practically nothing to do with it. Read here for more information: https://secure.wikimedia.org/wikipedia/en/wiki/BeOS#History
During the time when both BeOS and Linux were still young, those projects had quite different userbase and near term goals. BeOS was a commercial, closed source multimedia OS, Linux started as something totally different from it. Only after BeOS (or at least Be Inc.) had already went belly up has Linux started to really emphasize multimedia support too and to gain notable marketshare on multimedia desktops too - but actually multimedia support (for example, video editing) may still be quite far from Linux' greatest fortes.
Edited 2008-10-25 16:34 UTC
Despite all of the critics and not-agreeing that UNIX is dead, it's going down, I think in terms of sales - If I'm to buy systems, I'd likely choose / assemble hardware of my own, x86 based, and then install SLES or RHEL on top of it, with commercial support, if I need it. It's definitely less expensive than AIX, HPUX, Solaris, etc. For example consider this - adding 4 gigs of RAM on P4/5 IBM server costs around 800-1000$ ? How about I purchase simple RAM module for 50-100 for my x86 based machine ?
Also, no one could convince me that old proprietary UNIX systems are better for supercomputing and large-scale systems - see Google and Cray.
Old Sun stuff is not optimal for super computing stuff, that is correct. Linux is simpler to tailor for super computing than Solaris kernel. But Linux scales quite bad in everyday work. If you do number crunching, a simpler Linux kernel suits better. Unless you want to tailor the well scaling Solaris for that, of course. But Solaris is more complex and this is not easily done. Of course SUN could tailor the Solaris kernel for number crunching if they want to. It is easier to tailor a kernel to a specific simple task. And it eases the task if the kernel is simple enough to modify too, that is Linux. It has a naive approach as you can see by the fact that Linux becomes unstable when you stress it hard.
It seems that lots of companies goes to Linux and when they grow, they notice that Linux doesnt cut it, with high utilization. Then they switch to Solaris and report no problems.
And Solaris support is cheaper than Linux support. And you can of course install Solaris on the same hardware as Linux. The good thing is that when you know Linux or Solaris, it is quite easy to switch OS. They are quite similar - compared to, say, Plan9 vs Windows or Windows vs Haiku. I prefer Solaris + ZFS on my home fileserver than Linux + ext3. Who wants silent corruption?



