Windows to Linux defections to outpace Unix shifts in 2011 Linux server deployments are expected to take slightly more business from Microsoft than Unix in the next year, according to latest data. A Linux Foundation poll of major public and private sector organizations using Linux has found that 76.4 per cent plan to add more Linux servers during the next year, with just 41.2 per cent …
"I don't own stock in any companies involved with the Frankfurt Stock Exchange."
I can understand your relief. With reliable and responsive Linux based systems the chances of you making money from arbitrage trades on the Franfurt stock exchange will be nil. However, there's still time to take advantage of the crappy performance and instability of the London stock exchange, as they're running Windows for a while yet.
You honestly are not comparing the FUD created by Windows to that of Linux users would you? Would you risk your reputation on crap like that? After all the manipulation of studies by M$ against Linux & OSS, you wouldn't honestly be taking that pathetic a dig at Linux?
You tosser!
So you like to have a go a windows for creating FUD, but if someone acting on behalf of your beloved Linux does then that's fine.
How about being balanced? How about saying FUD is pathetic no matter who it comes from?
And I'm sure the poster was not having a dig a Linux, but at the study - which let's face it is, at best, a little flawed....
So what if Microsoft or Linux create FUD it is both the same in the end - FUD.
He did not take a dig at Linux (most of us are big enough to use what is needed and where) but took a dig at the Linux Foundation's useless survey.
Let me see if I ask 1900 UNIX users what servers they will deploy most what will they say? Repeat this with Linux, Winidows, OSX whatever operating system you like.
Oh you used a swear word at the end and used a $ in Microsoft. Didn't realise you were 5.
To be meaningful, the survey should have hit MS shops (and other *nixes shops) as well. Telling us that Linux users are going to continue and perhaps slightly increase their use of Linux tells us nothing at all, other than the fact they're going out to buy more kit.
We're a hybrid - 1/3 of our systems now are Linux, with the bulk Windows. We HAVE in fact increased our Linux server fleet by nearly 30 boxes in the past year while reducing our Windows servers (due to a major LOB application replatformed from Windows to Linux), but this survey wouldn't have teased out that information either.
...tell the new owners of the company I am just about to leave, who took one look at our Linux-based infrastructure and immediately decided it had to move to Windows servers for no other reason than Linux freaked them out (and they wanted to shut down my IT department). They are spending £16K alone on Exchange licences and probably around £10K on hardware. And then there's the Intranet server, helpdesk, customer survey system, Nagios network monitoring, Asterisk VoIP platform, data collation system, distributed across-the-wire backup servers.
Sigh.
In my previous contract, a new young IT manager decided everything had to move from SUSE to Windows (it is what he knows). The email migration alone swallowed an entire year's IT budget and two years later, they are still fiddling with webservers, mysql databases, etc. The new IT manager has now been moved sideways and I guess everyone is trying to figure out how not to look stupid when the directors ask them what they are doing and where all the money went.
What a waste of public money!
Look at the level of discussion that the Linux community has descended into: when a new distro is being considered the first thing anyone talks about is whether they have adopted KDE or Gnome as the UI as if that is in some way important. Kind of how people do with a new version of Windows.
Then you have the large, Windows-style, monolithic packages of hundred of megabytes that seem to get bigger and slower each revision for no discernible reason. Kind of like Windows. Even the Linux kernel is not immune - different versions of the 2.6 kernel differ by more lines of code than are actually in any of the other free Unix kernels in total. What does that bloat do but make it more like Windows?
Linux was a reasonable OS ten years ago and the Linux community could legitimately crow about how much better than Windows it was. They've spent the last ten years dumbing it down and making it just like Windows instead of so much better than Windows.
If it looks and stinks like Windows then it is only logical that is where the majority of conversions are going to come from. Modern Linux has devolved to a point where it bears only a passing resemblance to a proper Unix system.
You don't seem to have a very clear picture of the Linux landscape there.
"Different versions of the 2.6 kernel differ by more lines of code than are actually in any of the other free Unix kernels in total"
More than half of the kernel is drivers. Linux has better hardware support than most free Unix kernels. You can choose what you compile in if you really care; Ubuntu runs beautifully on my Sheeva plug and even clips along at a reasonable speed on my EEE 900, complete with Gnome. OpenOffice.org starts from cold on my EEE in less than 5 seconds. Hardly bloated.
The 'dumbed down' Linux distributions (let's take Ubuntu as a canonical (sorry) example) are there for people who need or want them. If you want something rawer and leaner, you can still have it.
...exist to lure those who're turned off by Windows for various reasons.
There are less dumbed down distros out there. If you're feeling adventurous, go for Debian or Arch. Try Slackware if you really want to see a distro that presents Linux and it's apps the way it's meant to be seen. And if you're really feeling masochistic, Linux from Scratch is right up your alley.
Yes, another myth that is repeated so often it is accepted without question. For bread and butter stuff none of the free Unix systems are really lacking. Linux may have the edge on random consumer tat but nothing more than that. If you want hardware support look at NetBSD with support for dozen of different architectures in a single kernel (Linux is a differnt kernel for each new system) as well as support for devices that simply doen't exist in the world of x86.
It's still a tiny fraction of the size of the Linux kernel. So what was that point about it all being hardware support?
Like it or not as Linux has matured it has grown disproportionately. Yes, it is now huge and still growing for comparatively little in terms of substantial benefit. This has only considered the kernel since that is what is most relevant to server apps, but the user interface stuff is if anything else worse. I agree with the "Spectacularly Refined Chap": there is a bloat problem there but pointing anything like that out will be instantly shot down by the religious Linux advocates instead of measures being taken to actually address what are growing into serious issues.
When it comes to servers, the only reason I even bother to leave a GUI on the damned thing is to keep my DBA happy (so he can run his Oracle bits). Otherwise, all you see at my server consoles is a friggin' bash prompt.
KDE or Gnome... for frig sakes. Are you an MCSE or something?
You get this time and time again from Reg commentards....
This is about servers, not desktops. With the level of confusion between the 2 expressed here ("Linux will never take over from Windows - it can't [play my fave game] etc") I truly wonder about who The Reg's audience are.
I always thought this site was read by mainly people in the tech industries but I'm coming to the conclusion that most commentards are arm-chair Sysadmins without a clue about how computers are used in real businesses.
And for the record, my company (one of the largest banks in the world) is going exactly this way too. Of course I don't expect Windows to be gone in 18 months... but there is a very definite trend.
I wouldn't say that a GUI is ridiculous in a server environment: it's a way for people to talk to the computer, and it has its own set of strengths and weaknesses. I only have a standalone Linux box, so of course I don't know what I'm talking about. I just make typing errors.
A Linux box with a GUI can be an extremely smart terminal console. I suppose there might be an ultra-cheap video chipset on the server's motherboard, but if things have gone that pear-shaped, safer to stick to text.
And if you're competing with Windows, I don't think you can ignore the GUI, server or desktop.
One reason not to have them is that the server is better located elsewhere than the company, person or organisation that uses it, i.e. in a data centre designed and serviced to provide redundant connections to the internet backbone, on site power generation and specialised staff in server maintenance and maintaining hardware. Another is the memory soaked up in this expensive environment by GUIs. A third reason is that sharing the cost of hosted servers works well with virtualisation, but this also limits memory available to a single virtual server instance.
For those not liking command line administration there are a couple of options. Many server applications can be administrated using web applications these days, and file managers on Linux desktops work very well over SSH, so you can drag and drop files over the network and edit them locally on something that has a GUI if you wish.
"For those not liking command line administration there are a couple of options. Many server applications can be administrated using web applications these days, and file managers on Linux desktops work very well over SSH, so you can drag and drop files over the network and edit them locally on something that has a GUI if you wish."
Understandable, but imo if someone doesn't like command line administration, maybe they're in the wrong job?
You clearly don't use any flavour of Linux. I would even go so far as to say the "dumbed" down Ubuntu can be changed to your hearts content and using your own thought train, made "pro".
Please, do us all a favour and go crawl back under your rock, you know absolutely nothing about Linux clearly. I'm writing this from what was originally Xubuntu, took pretty much all the standard stuff off and have built it back up. Why you ask? The canonical repositories are quite awesome and up to date.
It's the repositories that make Linux a community, simples. I hazard a guess to say (from the way and what you write) you are using a Mac. Well guess what, it's based on Linux...
"Well guess what, it's based on Linux..."
The sound you are about to hear is the fail buzzer. OS X is based on a grab bag of UNIX-ey technologies, Mach, OpenStep, various flavours of BSD, none of which are - or have ever been - Linux.
Probably feell like hiding under that rock you mentioned now, eh ?
Of course I've never used Linux. That is precisely how I was able to draw comparisons between Linux now and ten years ago. That is why I still have a Slackware install CD from 1994 in a drawer here. That is why I have submitted perhaps a dozen kernel patches over the last 16 years.
As for the GUI stuff - yes, it is not important. If you bothered to read my original post you would have seen I asked if how that "is in some way important". That's a distinct point to whether you have a GUI or not - I see another respondent couldn't even make that distinction. The bloat inherent in modern Linux distros is endemic: if you can't see that it is you that is either ignorant or too blinded by fandom to see what is as clear as day. Why was it than when we replaced a client's ten year old machine running 2.2 with a modern system running 2.6 packet forwarding times got longer? This is a measured fact not opinion or religion. Why is it that Linux used to blow Solaris away in terms of performance but nowadays they are broadly comparable on the same hardware?
The refusal to acknowledge this is actually part of the problem: look at my original post - 28 thumbs down! You wouldn't get 28 thumbs down for claiming 2+2=7. For the Linux apostles the supremacy of their chosen OS is a matter of simple faith rather than analysis. It has to be perfect. Linux is free! It's open source! It's sticking it to The Man! Therefore it must be good!
That attitude is fine for a rebellious and pubescent teenager whose opinions ultimately count for naught. In the real world where money rides on decisions made there is a duty to be more critical. I'm not the only one who has noticed the bloat issue: consider glibc for instance. It is an official policy of the development effort that if any libc, anywhere, ever supported something then glibc must support it too. Does that make sense for a key library that all Linux apps depend on? Linus Torvalds is on record rallying against that, pointing out how ridiculous it is. Does he know nothing about Linux too?
I think perhaps you missed the point - organisations that have already adopted Linux for some applications are more likely to use it for increasingly more applications, because their experience of it has been good. If their experience had been bad they'd be looking at other operating systems.
Well, this comes from the Linux foundation so take stats with a pinch of salt. I am sure Larry would tell us that Oracle has 90% of the database business. I do agree that there is a move to replace Unix with Linux. Windows to Linux sounds more challenging. A brand new project may choose whatever.
Not clever, Frankfurt.
Linux is supposed to be used to displace inappropriate use of those desktop/consumer OSes from Redmond, not to displace real (albeit terminally untrendy) OSes whose wicked step-parents just want to see them fade away ASAP.
E.g. use Linux like they did at the London Stock Exchange, after the LSE dropped Tandem NonStop for a flagship Windows (FFS!) setup with Accenture implementing. Windows obviously was the wrong tool for the job, and after five or so years of trying and failing to make it work right, the LSE person in charge finally admitted defeat and duly moved on (that's not something you see very often!). And Linux went in instead of Windows.
If you want a simple always-on multi-site disaster-tolerant setup, VMS is still the one to beat. It'll be better still when after the IA64 retirement is announced and the Proliant version comes out (the HP web page below can't count and seems to think VMS and HP-UX are already on Proliant: "HP ProLiant servers running five operating environments – HP-UX, HP OpenVMS, Red Hat Enterprise Linux and Microsoft® Windows Server 2003" :)).
http://www.itwire.com/opinion-and-analysis/the-linux-distillery/28359-london-stock-exchange-gets-the-facts-and-dumps-windows-for-linux
http://h71028.www7.hp.com/enterprise/w1/en/solutions/storage-disaster-proof-solutions.html
Linux won't cost near as much (esp. for hardware) than buying a Sun or AIX box would run you.
OTOH, FreeBSD is my personal favorite for those bits that you positively, absolutely must secure (like external-facing DNS servers, as a start). Scared the hell out of the boss when I put 'em in, but 18 months on and (unlike the previous Windows (!?) ones), these have been humming along w/o a single successful crack attempt.
A suit here thought much the same thing about the cost. However, we had to buy RedHat licenses, along with other licenses to quickly build servers, along with other products to add HA and other features already in Solaris. In the end, it was more expensive, and the solution isn't as reliable, and we're looking to migrate back to Solaris.
As far as hardware costs, don't forget Solaris will run on x86, so your hardware costs could be the same. Yes, SPARC tends to be a bit pricier, but you tend to get more features, and the box can usually push more data. Yes, x86 is starting to add those features, but the cost is putting it inline with midrange UNIX servers.
In the end, it all depends what your requirements are and what you need. Windoze *gasp* may even fit the bill. Unfortunately, suits tend to fall for glossies and logic goes out the window...
All OSes have their particular strengths (and weaknesses). For things like VMS I would suggest that a particular problem can be the availability of software packages for that platform. Certainly the commercial software systems that I work with are very back-level on their VMS support. The vendors just don't put the effort in for the "minority" systems.
Truth is with Linux (and Solaris etc) you are generally going to be able to get the software you need.
As far as I know the project hasn't gone live yet. It was supposed to but not yet... So why the Windows project with SQL Server failed? Can you trade on LSE millions of trade a day? If you look at the deal this is purely with money. They bought a Sri-lanka company for 15milion outright, so makes economic sense if it’s so cheap to have all the technology as well so maybe can sell to other exchanges. Good luck to them though when it goes live. Tamil Tigers Server anyone?
Cost may have been a factor, but the current Windows/.Net implementation at the LSE failed spectacularly on several occasions. On at least one of those the exchange had to close for a full day. That is a *lot* of lost business - and all this despite Microsoft parachuting in the best .Net/Windows experts to make the project fly. This was supposed to be there flagship project to show .Net could cope as a large scale enterprise platform. It failed, big time.
How it failed when the system is being used right now and every day for the past few years? Yes it did fail once with the 2008 lehmans collapse but nobody claimed windows or .net or sql server was at fault and there were special circumances. And anyway, there are so many difference between implementations, most likely development is at fault not the platforms. I would suggest that you wait for the new project to go live and endure some hard stock market collapse days before triumphilising prematurely for something that’s behind schedule already.
It failed many times, one of these times for 8 hrs.
It was a huge disaster. It does not matter if it works 90% of the time. This is not your Vista laptop that you can restart 2-3 times a day and still get away with that.
You seem not to comprehend basic things like high availability.
Did any other exchanges crash on the day of Lehmans collapse? Its at times of financial crisis that these tools absolutely must work. How would you feel if you were a trader and you were unable to trade on the day Lehmans collapsed. TradElect and the LSE failed at a critical moment; other exchanges running different software (on Linux/Unix, no less!) carried on. Where would you want to invest your money? At the apparently flaky LSE, or in New York or Frankfurt? LSE had to do something to retain investor confidence.
These were software crashes, too, not due to volume of traffic - or at least not that alone:
Crash one: 9th September 2008
http://blogs.computerworld.com/london_stock_exchange_suffers_net_crash
Crash two: 26th November 2009 (Did anything significant happen on Nov 26th to justify this one?)
http://www.guardian.co.uk/business/2009/nov/26/london-stock-exchange-crashes
I've a feeling there are other, smaller outages that have not been widely reported.
In contrast, MilleniumIT that the LSE bought is already in use (TradElect had no other customers AFAIK) and has a proven track record on various exchanges around the world, including NYSE Euronext.
London being the second biggest most probably exchange in the world crashed on one of the most volatile days ever, what did you expect to crash the Harare Stock Exchange? So here is a link of other exchanges that crashed...
http://royal.pingdom.com/2008/10/30/when-trading-stops-major-stock-exchange-outages-this-decade/
Maybe some of them run Unix/ Linux.
Software is complicated, programmers make mistakes, the cleaner unplugs the cable, Cisco messes up the network, there is a variety of things that can possibly go wrong. As I said in my original post the deal made financial sense, they bought the company for £15m when they probably had to pay accenture £15m a year plus can sell it to other exchanges. Lets see how the new system goes, I've heard the new linux LSE sub-system went down for 2 hours the other day. Goggle it. So what? i am not saying is nice but software is complex, it doesn't mean the platform is at fault, many things can go wrong. Without knowing the exact problem is unfair to criticise linux or .NET or whatever. At least they realised that outsourcing doesn't always make financial sense.
Will it usher in a whole new round of chair-throwing? Oh, I hope there's pictures!
Hey if we can run with the chair-throwing motif .... It'd be like UFC meets Silicon Valley, all over again - the Royal Rumble Royale! ><
...because, most of all, Big Brother cares for you and his certified marketing engineers ><
As someone who worked on the project, some info I can tell you regarding OpenVMS is that apparently HP only have around 10 support staff in the entire world. Ten, FFS... And the rumour was that HP want to kill it off.
Their choice to use Linux was not a bad one at all, that said, the mish-mash of talented developers and TATA newbies with no knowledge of Linux caused problems. If the project would have been far better and developed faster had it been that they had recruited experienced EU contractors for the entire project.
"Even the Linux kernel is not immune - different versions of the 2.6 kernel differ by more lines of code than are actually in any of the other free Unix kernels in total. What does that bloat do but make it more like Windows?"
*shrug* nothing. A huge majority of that "bloat" is additional drivers. It's essentially inert unless you actually have the hardware that needs that certain driver; unless it's for a hard disk controller it won't even suck up any RAM since it'll be a module that is just not loaded.
"Look at the level of discussion that the Linux community has descended into: when a new distro is being considered the first thing anyone talks about is whether they have adopted KDE or Gnome as the UI as if that is in some way important."
You're just looking at the wrong forums. Some reviews have comparative benchmarks, compare what kernel the distros have, what features are enabled in those kernels, and what new features the newer kernel has compared to the older one.. compare memory management, caching, if it has any interesting new filesystems (usually benchmarked and stress tested as well), any new support ("new and improved" virtualization support, or whatever), that the distro has X server version foo+1 and waht new features it has compared to X server version foo, and so on. They'll have like a 15 page review and spend maybe 1/4 page discussing the GUI.
@Linker3000, that's damned unfortunate. Besides them moving in the wrong direction, it's also a bad decision to rip-n-replace an existing, working infrastructure.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
