Linux is not after Microsoft. Linux is chasing Unix out of the enterprise server scene. Why not Microsoft? Well, its because Microsoft apparently has not completely succeeded in the data center league.
Linux chasing out Unix In the enterprise server O/S battlefield
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
linux is growing very rapidly in server installs for bigger biz and enterprise biz.
ms windows likewise is still growing in server installs in big biz and enterprise.
all studies show unix’ market share shrinking in a fairly substantial way.
considering that unix got its start in the 70s, and linux and ms both got started in the server markets in a significant way in the 90s, its no surprise that in many ways they are both still attacking unix’ dominance.
http://www.enterprisenetworksandservers.com/monthly/art.php/738
“Linux server sales grew revenues at 56.9 percent and unit shipments at 46.4 percent, the seventh consecutive quarter of double-growth revenue growth.
Windows servers posted double-digit growth, as revenues grew 16.4 percent and unit shipments grew 26.5 percent in year-over-year comparison.
Unix server revenue declined slightly, dipping 3 percent worldwide year-over-year; but Unix server revenues showed growth in Japan and Asia/Pacific, where Unix-based IT infrastructure and telecommunications are expanding.”
No problem, turn your 24/7 IT shop over to an unproven OS. I have a friend that did. She needs a job now. After she got fired, the whole university’s IT infrastructure runs Solaris on SPARC and X86.
Linux on the desktop makes a lot of sense. Saving a couple of hundred per seat, that adds up fast. More secure, less problems, easier to support, no viruses. But with the cost of IT infrasture in most big shops, saving a couple of dollars on the OS makes no sense. BTW, have you priced a Linux server OS lately? Kind of high isn’t it.
— “No problem, turn your 24/7 IT shop over to an unproven OS. I have a friend that did. She needs a job now. After she got fired, the whole university’s IT infrastructure runs Solaris on SPARC and X86.”
In my experience, Linux can be as reliable as the admin running it is skilled, which is pretty much the case with any server. Im sure your friend blames Linux but, unless she was specifically fired for using Linux for some reason, I suspect it was her not having sufficient skill and experience to run a completely reliable Linux server.
“No problem, turn your 24/7 IT shop over to an unproven OS. I have a friend that did. She needs a job now. After she got fired, the whole university’s IT infrastructure runs Solaris on SPARC and X86. ”
get a clue. a OS that runs google sure the hell is damn proven.
http://news.netcraft.com/
http://uptime.netcraft.com/up/graph/?host=google.com
is that unproven for you?
Where can I download the version that Google uses?
Simple examples for the simple minded here {Anonymous (IP: 61.95.184.—)}
http://www.redhat.com/apps/commerce/rhel/as/#x86
Intel x86 Standard $1499 (American)
SuSE is similarly priced. So to quote you, “get a clue.“
http://www.kernel.org
It occurs to me, that one of the main reasons Linux is beating the metaphorical crap out of Windows in _replacing_ UNIX Enterprise install’s is portability. It’s almost certainly a lot easier to move an application from a commercial UNIX to Linux, than it is to move that same application to Windows.
Add in the fact that if Linux is missing an API call or two, it’s a lot easier to add in those API calls to Linux, than it is to request Microsoft to do the same to Windows.
“It occurs to me, that one of the main reasons Linux is beating the metaphorical crap out of Windows”
you must not read the statistics very well. yeah linux is doing super dandy in the web serving space…email too to large degree.
but they are far from beating the crap out of windows.
windows remains the overwhelmingly most installed server os in the world.
“It’s almost certainly a lot easier to move an application from a commercial UNIX to Linux, than it is to move that same application to Windows”
well you forgot something very important:
Windows Services for UNIX (SFU) 3.5 provides the tools and environment that IT professionals and developers need to integrate Windows with UNIX and Linux environments.
http://www.microsoft.com/windows/sfu/default.asp
download it for free at:
http://www.microsoft.com/windows/sfu/downloads/default.asp
Anonymous wrote:
> is that unproven for you?
Google’s demands on the reliability of the software and hardware which powers it are minimal at best. Their use of Linux demonstrates nothing except that it was the cheapest way to build a cluster which can be made highly tolerant of node failure. The below USENET post puts it quite well:
http://groups.google.com/groups?selm=rshu_20040705122020%40stra…
Just as well you used the words “server os”
> windows remains the overwhelmingly most installed server os in the world.
since the most installed O/S in the world would be something like Symbian or one of the embedded platforms (not even Wince) since cellphone numbers are rocketing past both desktop and server platforms.
So justifying the alleged “superiority” of Windows based on the fact it is the most installed seems a moderately flaky argument to me.
Pleas reread my comment. and don’t take things out of context. I was talking about Linux and Windows replacing existing Enterprise installations. This means where a company has things like Oracle and DB2 running on HP-UX, and S/390s, and Windows and Linux (in theory) have now grown up enough (along with the CPU’s) to run those said databases on relatively generic hardware, instead of massive parallel servers.
This isn’t about email or webserving, it’s about databases and ERP software (which interface with those databases), and the software that a companies engineers will have built up around those databases.
Regardless of the relative merits of WSFU, I find it exceedingly unlikely that it’s ever going to be as good at providing a UNIX like environment, as something that, for all intents IS a UNIX.
The comment about “windows remains the overwhelmingly most installed server os in the world.” means very little in this context. Show me the number of Windows servers handling tens of thousands of simultaneous users or connections per machine. If they exist, I would put money on the odds that are many more UNIX machines at this level than Windows machines.
“Google’s demands on the reliability of the software and hardware which powers it are minimal at best.”
thats nonsense
read
http://news.netcraft.com/
Linux vs Microsoft is no “fair” fight.
These comparisons between OpenSource and Microsoft never work. The best technical os does not win. It’s all a lopsided unfair business.
I will tell you why. Microsoft is not just an OS. It has software that it can leverage(handout) to potential customers.
I think EBAY uses microsoft server. Does anyone know the real deal in how these companies select their os. I don’t think you do. Microsoft probably gives huge discounts because they want the big companies on their side so they can show the ‘small’ business that microsoft is used by the big boys.
Microsoft guys are shrewd as hell. They are incredible business people. Although office and vb are cool.
“The comment about “windows remains the overwhelmingly most installed server os in the world.” means very little in this context. Show me the number of Windows servers handling tens of thousands of simultaneous users or connections per machine. If they exist, I would put money on the odds that are many more UNIX machines at this level than Windows machines.”
that is the last bastion for unix.
and just as ms has only been in servers for about 10 years in any meaningful way, they only started to move on the big iron scene about 5 years ago.
but this shows ms is now there with windows server 2003 datacenter edition
http://www.microsoft.com/windowsserver2003/evaluation/features/comp…
support for:
up to 64 way itanium II 64 bit cpus
supports up to 512gb of ram
all the big iron pros like ibm, hp, nec, unisys, dell offer enterprise class solutions ( http://www.microsoft.com/windowsserver2003/partners/oems/64bitoems…. ) that can handle the worlds most demanding erp or db needs….
http://www.theinquirer.net/?article=9755
“the record breaking TPC-C results from the HP 64-way Superdome, leapfrogged NEC when HP, Microsoft and Intel achieved the world’s highest-ever single-system TPC-C benchmark on a non-clustered 64-processor HP server based on Itanium 2 Madison, Microsoft Windows Server 2003 Data Center Edition and SQL Server 2000 Enterprise Edition (64-bit). The result of 658,277 transactions per minute (tpmC) exceeded the previous record, set by a 128-processor RISC system.”
not bad for the new kids on the block.
Anonymous wrote:
> thats nonsense
Really? If a node fails, what happens? Google is basically a giant farm of web severs behind load balancers making queries to an equally large farm of database servers (maybe they’re even the same machines). If a machine fails you the system just switches over to one of it’s identical brethren until it can be rebooted or replaced. If some HTTP connections are dropped in the process they will only be a tiny drop in a vast ocean and the affected clients will simply retry. There’s no penalty for queries returning non-current data so database writes from the crawlers can propagated at leisure.
Given this, the database can simply be redundantly distributed across the cluster. As any given node carries neglidgable state it doesn’t matter if it fails. The important thing is to buy the hardware with the best price/performance ratio and enough of it so that only a small portion of the cluster will have a software or hardware failure at any one time. Google is excellant example of how to build a reliable whole from many unreliable parts. However, as the referenced post pointed out, these techniques are only applicable to limited set of problems.
I’m not claiming that Linux is unreliable, merely that Google is an extremely poor example of Linux’s ‘success’ in an enterprise enviroment. Linux clearly can work well in small configurations such as desktops or web sever farms. When you start looking at larger and mission-critical systems where having a support contract is more a necessity than nicety, Linux has no real cost advantage over commericial UNIX offerings.
:I’m not claiming that Linux is unreliable, merely that Google is an extremely poor example of Linux’s ‘success’ in an enterprise enviroment. Linux clearly can work well in small configurations such as desktops or web sever farms. When you start looking at larger and mission-critical systems where having a support contract is more a necessity than nicety, Linux has no real cost advantage over commericial UNIX offerings”
you simply said its unproven. its not. it has very real cost benefits. nobody is switching to unix to linux out of love. there are good business reasons why people like ibm,novell, sgi and hp are doing it
:”” When you start looking at larger and mission-critical systems where having a support contract is more a necessity than nicety”
mission critical stuff dont depend on outside support. they have their own team working and customising on it. they are a whole lot of enterprises using linux mostly in the form of redhat or suse linux. they have their reasons.
Since Linux *is* a Unix(-like) OS this would mean it is chasing itself out. It would be more correct to say that it is chasing commercial Unix systems out.
“It would be more correct to say that it is chasing commercial Unix systems out.”
no. thats not accure. redhat and suse linux are commercial linux systems. it would be accurate to call them proprietary unix systems.
support for:
up to 64 way itanium II 64 bit cpus
supports up to 512gb of ram
I’m sorry that just fails to impress me at all. http://www.sgi.com/servers/altix/
http://www.arc.nasa.gov/aboutames-pressrelease.cfm?id=10000087
This is Linux 2.4 on a 512 CPU system and I think 8TB RAM. They should be aiming for at least 1024 CPUs and 32TB with Linux 2.6.
So if you want to play the big willy game, first make sure you don’t have the smallest one. OK?
Anonymous wrote:
> you simply said its unproven.
No, I never said that. You’re confusing me with ‘nobody’ who started this thread. I only said that Google didn’t require reliable hardware or operating system software. A statement I feel I’ve adequately justified in another post.
> its not. it has very real cost benefits. nobody is
> switching to unix to linux out of love. there are good
> business reasons why people like ibm, novell, sgi and hp
> are doing it
Again, I must criticize you choice of examples. All of those companies are operating system vendors, not customers interested in Linux becuase of the high cost of commercial UNIX licenses or support. Novell and SGI were clearly either unable or unwilling to bear the cost of developing their own operating systems. IBM and HP, on the other hand, are much larger companies and likely view Linux as an oppertunity to expand their business rather than to displace their existing UNIX varients. They both also sell a number of non-UNIX derived systems in niche markets.
This is Linux 2.4 on a 512 CPU system and I think 8TB RAM. They should be aiming for at least 1024 CPUs and 32TB with Linux 2.6.
Oh, and before anyone jumps down my throat here… It *is* a modified 2.4 kernel, but much of it is stuff that has come from 2.6 (eg, O(1) scheduler) or is otherwise GPLed. I think the main SGI code is their XSCSI layer – which apparently the 2.6 kernel won’t need.
or pay attention to what this article and thread are about
you post a link to nasa and nasa is not a business/enterprise. last time i checked nasa has a gigantic budget funded by the american taxpayers and isnt known for giving a return on investment the way corporations do.
scientific computing is a different animal all together.
“”IBM and HP, on the other hand, are much larger companies and likely view Linux as an oppertunity to expand their business rather than to displace their existing UNIX varients”
point is linux sells in the enterprise area too pretty well. its not unproven.
I know what the thread is about. You trying to make out how awesome windows 2003 is because it can handle OOOH 64 CPUs and AAAHH 512GB RAM.
Sorry, better luck next time.
Anonymous wrote:
> mission critical stuff dont depend on outside support.
I must say that’s an extremely strict definition of mission-critical. Perhaps ‘more critical’ would have been more appropriate. In any case, even if you’re running a mission-critical server and going to the lengths of having a dedicated in-house support team and stockpiling spares for rapid response, that’s no reason not to have a support contract.
My point stands that businesses tend to have their servers under vendor support contracts. Once you start paying for support, Linux has no longer has a clear cost advantage over commercial UNIX varients, particularly when on IBM-PC compatible hardware (e.g. Solaris/IA32).
My point stands that businesses tend to have their servers under vendor support contracts. Once you start paying for support, Linux has no longer has a clear cost advantage over commercial UNIX varients, particularly when on IBM-PC compatible hardware (e.g. Solaris/IA32).
Oh is that your point? Well Linux must have definite advantages other than cost then, if so many are migrating. Or perhaps you think that the industry has managed to pull the wool over the eyes of all those people…
Anoymous wrote:
> Oh is that your point? Well Linux must have definite
> advantages other than cost then, if so many are migrating.
This is a very complex issue. I suppose that a large proportion of Linux marketshare is in unsupported (relativly) low-end boxes in web server farms and the like. It would be interesting to see where all these new Linux users are comming from. Vendors whose platforms are dying? SCO, Novell, SGI, and HP(Tru64). Vendors who are pushing Linux in addition to their own platforms? IBM and HP(HP-UX). Or Sun?
> Or perhaps you think that the industry has managed to pull
> the wool over the eyes of all those people…
To a certain extent, yes. Linux has a lot of fans and they’ve been pushing it hard. It would be foolish to suggest that purchasing descicions are made purely on the merits of the products and are without regard to the popular support Linux currently enjoys.
Sun’s idiocy in delaying the productization of Solaris 9 on Intel hurt them a lot. SPARC severs are beautiful machines but Sun was too slow to realise they needed to have a competitive solution on Intel hardware too. If Sun open sources Solaris, things will only get more interesting.
Robin Kay wrote:
> Anoymous wrote:
>> Oh is that your point? Well Linux must have definite
>> advantages other than cost then, if so many are
> migrating.
>
> This is a very complex issue. I suppose that a large
> proportion of Linux marketshare is in unsupported
> (relativly) low-end boxes in web server farms and the
Well, this article for example says it is pushing Unix out of the “enterprise”. Also, most commercial Linux distributors make their money with support contracts.
The issue is only a “very complex” one if you try to piece everything together in a way that supports your arguments. Otherwise it is very simple.
> like. It would be interesting to see where all these new
> Linux users are comming from. Vendors whose platforms are
> dying? SCO, Novell, SGI, and HP(Tru64). Vendors who are
> pushing Linux in addition to their own platforms? IBM and
> HP(HP-UX). Or Sun?
Why would that matter? If they are dying, it is because they are no longer economically viable for the company. That is how the capitalist ecosystem works. So, take SGI for example: for them, Linux is obviously up to the job of running their hardware, so it no longer makes sense to continue to put money into IRIX, right?
Again, the issue is actually very simple.
>> Or perhaps you think that the industry has managed to
>> pull the wool over the eyes of all those people…
>
> To a certain extent, yes. Linux has a lot of fans and
> they’ve been pushing it hard. It would be foolish to
> suggest that purchasing descicions are made purely on the
> merits of the products and are without regard to the
> popular support Linux currently enjoys.
>
So before the likes of IBM, SGI, HP, etc pulled the wool over everyone’s eyes, they must have had the wool pulled over their own eyes to take up Linux in the first place. I don’t think so.
There is a blindingly simple reason why all these people are starting to use and support and invest in Linux. The issue need not be very complex… you are just failing to see the forest for the trees.
> Sun’s idiocy in delaying the productization of Solaris 9
> on Intel hurt them a lot. SPARC severs are beautiful
> machines but Sun was too slow to realise they needed to
> have a competitive solution on Intel hardware too. If Sun
> open sources Solaris, things will only get more
> interesting.
Where did this come from? Sun has been hurting long before Solaris 9 on Intel was delayed. SPARC servers may be beautiful, but they are expensive, and can no longer keep up with Intel or IBM chips at the high end.
As far as interconnects and IO, Sun’s sunfire (or whatever it is) interconnect is a crossbar: a big, dumb, expensive, unscalable dinosaur, which is why they’ve been stuck at 18 nodes for years. Unfortunately, the only way they can compete with an Itanium is with 2 SPARCS, and now that the likes of SGI and HP have come along with some decent interconnects, Sun gets left in the dust. Oh, watch the POWER5 too.
Anyway, why you brought Sun into the argument is beyond me… what is your point there?
mission critical stuff dont depend on outside support. they have their own team working and customising on it. they are a whole lot of enterprises using linux mostly in the form of redhat or suse linux. they have their reasons.
I doubt you’ll find many “mission critical” enterprise systems that *don’t* have support contracts with the software and/or hardware vendor.
As someone else pointed out, Google are a unique example, it is not correct to generalise from their very specialised requirements and environment.
I find it very interesting to hear that Linux is doing such good way to take out the Unix competition (as it claims). I personally would believe most of the usage which raises the bars are because A LOT of small webservers use Linux as it is so simple to set up.
You could just as well continue this article by writing Linux steal from proprietary *Nix, however BSDs gains a lot of users who previously chosen Linux. So is Linux really a long term choice or is it something temporary while people learn how to master Unix? Sysinstall hardly graceful compared to many of the distros for Linux, but after a while you might consider technical merits not eye candy merits ….
When you start looking at larger and mission-critical systems where having a support contract is more a necessity than nicety, Linux has no real cost advantage over commericial UNIX offerings.
That’s an individual judgement which isn’t necessarily supported by others. I’m still wondering why you think this is universally true. Have you compared the support contracts of a commercial UNIX vendor offering versus the one from a Linux distributor like SuSE or RedHat? Or perhaps on a broader scale: SGI Altix versus some newer SUN SPARC?
> Oh is that your point? Well Linux must have definite
> advantages other than cost then, if so many are migrating.
This is a very complex issue. I suppose that a large proportion of Linux marketshare is in unsupported (relativly) low-end boxes in web server farms and the like.
Large number of low-end computers are unsupported? Where do you get your statistics like this one from? Why wouldn’t they have a support contract from eg. SuSE or RedHat? You’re basically saying Linux/x86 is inferiour on the issue of stability. Anything to back that up?
To the Australian person replying to you: In contrast to HP and IBM, SGI is a special case as compared to these 2 UNIX vendors; as a smaller company, it barely survived the dot-com aftermath so they’ve chosen to outsource lots of software to the FLOSS community while developing/porting some necessary applications to support their old cause on a “new” platform: Lintel. Maintaing and developing MIPS, IRIX is just too costy for them (they’ll drop support in 2006 and the MIPS product line has already been dropped. No more R18000.). As huge companies, IBM and HP had less problems with this aftermath so they’re able to form a more pragmatic style. SUN is somewhere in the middle of IBM, HP vs SGI — but to me they only appear to realize, approve that recently.
To the Australian person replying to you: In contrast to HP and IBM, SGI is a special case as compared to these 2 UNIX vendors; as a smaller company, it barely survived the dot-com aftermath so they’ve chosen to outsource lots of software to the FLOSS community while developing/porting some necessary applications to support their old cause on a “new” platform: Lintel. Maintaing and developing MIPS, IRIX is just too costy for them (they’ll drop support in 2006 and the MIPS product line has already been dropped. No more R18000.). As huge companies, IBM and HP had less problems with this aftermath so they’re able to form a more pragmatic style. SUN is somewhere in the middle of IBM, HP vs SGI — but to me they only appear to realize, approve that recently.
Well, exactly. That was my point. It is more cost effective for SGI to scale down IRIX development and use Linux. There is no way they could do this and still survive as a hardware company if, for example, Linux isn’t “ready for the enterprise” or isn’t scalable or proven.
Umm no, Linux is technically superior to BSD in almost every way. There are just about as many people migrating to Linux from BSD as there are from BSD to Linux.
Not a very intelligent response… Did you ever consider stability as something technical?
Not a very intelligent response… Did you ever consider stability as something technical?
I did, and I do. There is no evidence of BSD being more stable than Linux anymore, other than questionable anecdotes from BSD zealots. I’m sorry – prove me wrong… I’d love to see an attempted objective test.
Note, I didn’t say Linux was more stable either, because I don’t have data. It isn’t unstable though, mind you.
http://www-106.ibm.com/developerworks/linux/library/l-rel/
Take a look at that for example. Machine runs for 60 days at full CPU load, doing networking, nfs, io, etc.
Linux is going through these sorts of tests all the time at IBM, and I’m sure probably SGI and linux distributors too. Also, OSDL has a lab that is basically on 24×7 doing lots of regression tests against every kernel snapshot that is released. I think their automated test hardware includes 4 or 5 8-ways with I think 40 or so disks, and a handful of each 4, 2 and 1 ways…
If someone says “Linux must be technically superior to *BSD because now there are more linux users than bsd users,” then he is saying “Windows is superior because there are more windows users out there.” Bah!!!
Honestly, I don’t know which OS is superior at all…but I do love both OS…Linux is like a brand-new sports car…lots of tunings and experimental stuffs; while BSD is like ol’ good classic cars…might look old and out-of-fashion…but neverthless GOOD.
Anyway, I run FreeBSD in my home pc while managing a Debian GNU/Linux server. Both OSes ROCKS. And Windows? It’s a game console. really. yeah.
Well, we’re talking about the respective kernels here…
Linux has wider driver support, it supports much more architecture support than FreeBSD (on par or more than NetBSD), it is more performant in all the recent benchmarks I have seen, it undoubtedly scales to more CPUs and larger memories, it supports journalling filesystems (softupdates really haven’t shown to be even as good), has lower overhead for basic things like syscall and context switch latency and end to end network packet processing, supports direct IO, asynchronous IO to name a few things.
How on earth does a discussion about linux taking server share off Unix at a faster rate than Microsoft turn into ‘Linux is Better than *BSD’ debate, a topic which has been so utterly exhausted on these forums and elsewhere. Please keep to the topic huh?
Agreed, that wasn’t the point with the debate from the start. The issue was that Linux isn’t really grabbing over that many BSD users but probably picking from AIX, Solaris, Unixware etc… and I highly doubt that BSD users are stealing from…. Windows???
That obviously means that the BSD users also come from some place, and if we’re speculating a bit, it’s likely they come from Linux. Hence why I said from the start that Linux can also be considered a middle step before moving back to another Unix again namely BSDs…
So the issue was NOT about performance nor licenses or anything, just about the trends in the computer society, just like the article was talking about a trend as well…
> Large number of low-end computers are unsupported? Where
> do you get your statistics like this one from? Why
> wouldn’t they have a support contract from eg. SuSE or
> RedHat?
That’s not what I said or meant. For starters, it was a supposition. I was supposing that Linux would have it’s biggest wins in enviroments which are cost sensitive such as web server farms. In such an enviroment where high-availability is not required, you have lots of hardware doing the same thing and can easily fall over on to another machine if one fails, you might well not purchase vendor support or have a minimal level of such support. Linux having no licensing cost might therefore be more cost effective than a commerical UNIX system.
> You’re basically saying Linux/x86 is inferiour on
> the issue of stability. Anything to back that up?
I have not said nor will I say that Linux is unstable.
Anonymous wrote:
> Why would that matter?
There’s a difference between a customers that has been forced to migrate to one seeing such a great advantage in another platform that they’re willing to undertake a costly migration.
> So before the likes of IBM, SGI, HP, etc pulled the wool
> over everyone’s eyes, they must have had the wool pulled
> over their own eyes to take up Linux in the first place. I
> don’t think so.
Again, you’re putting words in mouth. IBM, SGI, HP, et cetera certainly havn’t had the wool pulled over their eyes. SGI may have had no choice, but IBM and HP have merely taken advantage of an oppertunity to profit. Even Sun supports Linux on their Intel servers and the JDS presently uses the Linux kernel.
> There is a blindingly simple reason why all these people
> are starting to use and support and invest in Linux. The
> issue need not be very complex… you are just failing to
> see the forest for the trees.
Not at all. I can see as clearly as all those companies that there’s money to be made from Linux. Customers want to cut their infrastructure costs and have heard Linux is a means to that end, perhaps without considering all the options.
> SPARC servers may be beautiful, but they are expensive,
> and can no longer keep up with Intel or IBM chips at the
> high end.
True enough. Fortunatly, there are more important things to most server applications than raw processor performance. SPARC servers are still selling well enough and Sun is still reporting customer wins. The APL and throughput processors will arrive before long.
> which is why they’ve been stuck at 18 nodes for years.
Presumably you’re talking about the number of ports on each crossbar switch?
> Anyway, why you brought Sun into the argument is beyond
> me… what is your point there?
Apart from UnixWare, I believe Sun is the only vender with a UNIX derivative on commodity Intel hardware. That’s the point.
I recently purchased SuSE Enterprise Linux 8 to run on an IBM p615 for an application in my shop. I’d read so many good things about it, and *LOVE* SuSE on my pc both at work and at home. However, after going 3 weeks with SuSE support and finally the devs at SuSE, my problem was never *REALLY* resolved. They sent me a workaround, but there were some manual steps in the event of a drive failure, which defeated the whole purpose. And before anyone who doesn’t know speaks, you can do fully automatic drive mirroring on a pSeries box with AIX. When you get right down to a basic function like basic drive mirroring, and it doesn’t work, then I’d say it isn’t ready. I now am of the mind (and others that have responded on the SuSE mailing list) that we do not believe it’s not ready. And, I hate to keep slamming SuSE, but it’s a problem with Linux on the pSeries, not just SuSE.
Maybe I am a complete idiot, but what the hell is the much sought after “enterprise environment” I hear discussed so often. I have worked and am working in 2 fairly large industires, Insurance and DoD. I have yet to see something in either of the 2 places that Linux could not handle. Just my $.02
That’s an individual judgement which isn’t necessarily supported by others. I’m still wondering why you think this is universally true. Have you compared the support contracts of a commercial UNIX vendor offering versus the one from a Linux distributor like SuSE or RedHat? Or perhaps on a broader scale: SGI Altix versus some newer SUN SPARC?
Sorry having a support contract is SOP in large corporate shops. Why would i want to have two different vendors supporting my server install one for hardware one for OS? when i can get a package deal from HP, SUN or IBM? if i get the package deal where the cost difference between HP-UX and Linux on the same hardware is just a couple hundred dollars it would be foolish to choose Linux over HP-UX.
all these heavy improvements of GNU/Linux are not becoz of Linux or OSDN or Redhat or anything else, its clearly becoz of the Mr.Richard Stallman’s GPL . without his ‘free’ open source concept, Linux kernel would have been in the ditch for these many years. the whole world should be thankfull to him. Long live his ideas !!!
Idiot. Linux could well have done just as well with any number of open source licenses. The GPL isn’t the be all and end all.
I have worked in both the DoD and commercial environment and can tell you, Linux isn’t totally excepted yet, and won’t be. The NSA has done some experimentation with SELinux, but it still isn’t secure, and most organizations within the government either run Solaris, and more recently BSD to handle servers considered to be “mission critical” and where security is the main concern.
Some might argue that the government is behind, I beg to differ. I saw things in the DoD that my commercial company hasn’t implemented yet, as well as other companies I have seen. So far the DoD is the most advanced and secure scenario as far as “enterprise” level architecture is concerned, and undeniably the biggest as well.
Linux is great for web servers, MTA’s and other stuff, fact is it just isn’t accepted as much as open-source zealots would like it to be. Windows 2000 is a fantastic OS and until Linux can come up with something that can beat active directory there isn’t really a comparison.
Instead of competing with Windows, Linux needs to be able to integrate with Windows environments because let’s face it, as much as most of you fat pizza eating hairy nerds want it to happen, Windows will never leave the “enterprise” , and will always be around.
“I have worked in both the DoD…”
Liar
Key points:
0. Is this ignorance or is it marketing arguements boys?
1. Linux *IS* after Microsoft. Microsofts products are aligned squarely at the SMB market. Linux will challenge at every level of the industry due to the open code base and the ease of extension.
2. Clustered Linux works well for those sorts of problems.
3. Open source development will improve security and reliability to a higher order than any of the old UNIX versions.
4. Linux runs on the widest number of hardware platforms than any of the other OS’s mentioned. Intel should *FEAR* this and so should anyone else with cack x86 based chips. Power chip platform shows promise.
5. Linux and the overall opensource programming space is expanding to fit the entire possible code market. It will become defacto like the web & ip protocols. The miriad of Linux distros allow it to morph into and incredible range of market spaces
6. The Linux kernel and other Opensource components which make up the distro can be constructed from scratch by a reasonably capable tech. Its not impossible and this means ultimately the OS is free to a student. Its also able to be scrutinized by any one.
7. Bill will bet the company and change Microsoft down the Linux path just as he did with IE. If Balmer stops him then its curtains for MS.
8. Sun is confused. Its stance on Java will change to opensource or Java will lose its customer base with more security worries.
9. Open source is the ultimate in security. True security to know whats in your servers code base. True ability to patch it if you can. Maximum eyeballs to scrutinise and attempt to hack eventually evolving the platform to an extremely hardened level. Security is understanding you system so you can defend it not living behind an iron curtain.