Submitted by “In various older studies, Microsoft and some analysts claimed Linux has a higher total cost of ownership than Windows. They attributed the difference mainly to higher system management costs, and concluded that the higher TCO outweighed the much lower license and acquisition costs for Linux. However, in a new study of over 200 Linux enterprises, Enterprise Management Associates found that this perception is no longer accurate. Sophisticated management tools now allow Linux management to be fast, effective, and inexpensive. With far lower acquisition costs, Linux is now a cost-effective alternative to Windows. EMA analyzed the cost factors cited in previous studies and found the following results.” Please note that this study was sponsered by OSDL, so take out your salt while reading this.
Stuff like this is not worth the disk space it is stored on.
Microsoft-sponsored study: “Windows is teh win, Linux is teh lose”
OSDL-sponsored study: “Linux is teh shizzit, Windoze is teh lame”
Come oooon. It’s not like this is going to change soon.
At least the symmetry (which this study hightens) makes it more obvious how little these studies reflect reality.
I can only agree.
There will always be cases where Windows has the best ROI, and there will be cases where Linux has the best ROI.
So generic studies makes no sense to begin with.
Stop feeding the industry analysts!
right on. i’m getting kinda sick of being told that my OS is either shit or the shit. And while I obsess over keeping linux, The only computing solution out there that really lower the cost of performing power user tasks(like CGI or video editting for instance), is apple.
Kinda sad that Microsoft hasnt gotten around the whole “delete” button when it comes to their own code, cause maybe they’d be able to surpass apple for quality. If only billions of dollars could be properly used. and pigs could fly….
“Stuff like this is not worth the disk space it is stored on. ”
I agree but your talking about the study , and I target your comment.
“Microsoft-sponsored study: “Windows is teh win, Linux is teh lose” ”
Its what Microsoft whant people to believe but the reality is in the data offered.
For one Microsoft sponsored study are never marked sponsored by Microsoft. Second they use the price of
UNIX engineer , where as GNU/Linux Engineer are much much much more cheaper to hire. Third they never mention the price of License per computer or per CPU which on 50 machine can be something like 50 000 dependaing on th task , Fourth they never mention the License vérification cost ( if you have someone installing an illegal license or are unable to show the license that goes with the machine you will get fined ,
they usualy use the priciest solution availaible ( IBM mainframe ) and compare it to an intel Middle of the road solution ) , they always forget training cost of the staff , and never discuss the cost of deployment or addition of new machine , They never mention the fact that the staff and employee can make themself a copy and bring it home fro personnal use and study and training. GNU/Linux is truely cross platform where as Microsoft solution are tied to one or two type of cpu solution. Microsoft stop upgrading amd offering security solution after five years and no one else can offer to continue to do them.
I agree , its never going to change for someone like you , your not interested in the truth in this discussion your interested in bashing GNU/Linux.
Listen, you should get off the drugs some time.
I was hinting at the fact that “studies” are always going to be biased towards the party that paid for them — whether this is Microsoft, IBM, OSDL, or Redhat. I wasn’t defending Microsoft and bashing GNU/Linux. You obviously were bashing Microsoft, and for no reason at all.
Someone needs to hop on the clue train.
I dont take drugs of any kind , thanks for caring for my health 😉
“I was hinting at the fact that “studies” are always going to be biased towards the party that paid for them”
Its not always the case some people value there respectability and reputation based on trust and will not compromise there ethics and value.
“I wasn’t defending Microsoft and bashing GNU/Linux.”
Actually you where , you keep on saying that there is any truth to the report that Microsoft as brought forward even after you read them and know where the errors where made.
“You obviously were bashing Microsoft”
Not exactly , I whas reminding everyone that Microsoft report so far have been false and done in such a way that no real data or information can be gathered from them.
“and for no reason at all.”
Please its a Multi billion USD market that they are trying to maintain and keep. There lying and anyone who disagree is either stupid or as been paid by them , I dont say that in some occasion Microsoft cant compete or is not the best solution availaible , but when someone try to peg me as a fool by comparing oranges with banana , I whant to see honest comparaison on same hardware and with same budget from similar vendors type.
“Someone needs to hop on the clue train.”
That would be you 😉
I can’t do anything but agree with your title. It is a perfect description of your comment.
Actualy, RTFM would be good for you before posting bull. This was actualy first TCO study that posted contras to MS ones and in the same time tried to avoid any conclusion “X is better than Y”
Microsoft-sponsored study: “Windows is teh win, Linux is teh lose”
OSDL-sponsored study: “Linux is teh shizzit, Windoze is teh lame”
Come oooon. It’s not like this is going to change soon.
From the end of this study:
However, this study is not exhaustive, and does not mean that Linux is the best or correct choice for every implementation, application, or enterprise. The choice of platform must account for many more variables than just resource costs, management effort, or even TCO. Windows in particular has many available and proven applications. On the server side, Windows has made good inroads on UNIX, and has an opportunity to become a true enterprise platform. UNIX, while it has lost ground to Windows, remains a strong contender on the server, with proven stability, scalability, and application durability. Mainframe operating systems also continue to defy their critics, and IBM’s z/OS in particular maintains a stronghold in the large enterprise server category with massively scalable architecture and applications. Enterprises should examine their specific needs and use this data as a starting point before deciding on an enterprise platform.
Conclusion of complete survey: In translation for the lazy people like you.
There is NO lower TCO when you consider OS. TCO comes in question when you take in variables like: services you need, administrative tools you can use, administrators that will manage this system, hardware that will run specific jobs, OS used to do the job. As you probably notice OS is only 1/5 of the equation.
Where everything can vary? (on any OS, and study posted all this vars too)
Services? Well, if you need good e-mail or web server you’re probably get much better TCO with decent linux setup, but if you need for example calendaring (not web, I ment decent client for Outlook or other e-mail clients)? Probably still the lowest TCO with Exchange.
If Admin is not adequate. Good admin can script and automate his system to the last level to lower the job needed. Bad admin is just example how one can waste time with useless clicking. Good admins exists in both worlds, although from the nature of Linux there is probably better chance that you find people with more knowledge (MSCE, RHCE and such don’t mean anything here. after finishing MSCE, RHCE one still doesn’t posses 0.1% of knowledge to be effective admin. It is milleage that is needed, not some exams)
Best administrative tools. One can use sophisticated tools to be effective, one can use his brains. Both occasions are good. Having no administrative tools requires better admin. Sometimes results are more positive with having better admin than with having good tools. But better admin costs more.
Hardware? Ok, Linux wins here definitely. But hardware is usualy smaller part of equation.
OS? If OS is poorly choosen, job will suffer anyway, there is no Windows or Linux magic rule.
“Probably still the lowest TCO with Exchange.”
Can I say wrong ?
Chandler
http://www.salesforce.com/
http://www.sugarcrm.com
Can I say wrong ?
Yes, you can
As safe as sex gets when one is using condom 
Ok, Chandler is still far away, but other two look damn sweet. Although I couldn’t find salesforce pricing in this short time.
Damn, I wish you’d mention this two weeks before. I had to deploy Exchange against my wishes.
p.s. A question if you actualy used them. Does those two enforce their own e-mail server or they use server already present on the box?
Edited 2006-02-13 22:57
A quick read of the article mentions multiple times about “sophisticated management tools” but doesn’t specify which tools are used other than Tivoli, Unicenter TNG and similar “deep pocket” management software. My experience with Tivoli is the only thing that limits you is your imagination and whether your machines all have Tivoli Endpoint software installed.
If you used Tivoli or similar software in a Windows environment I would expect a similar level of uptime and ease in administration. But Tivoli, CA Unicenter, Levanta Intrepid, OpenView and ESALPS are not Microsoft products and Microsoft likes to sell their “solution” as much as IBM does. I wonder how the results would look if a Windows and Linux environment were configured in the same fashion using similar hardware and software, including management tools. I haven’t used SMS in years, but I feel that it doesn’t compare to Tivoli or Unicenter.
Edited 2006-02-13 18:37
There are: lies, damn lies, and statistics.
The OSDL and open-source community use all three to try to take down MS. Happily, they continue to fail.
One company says their solution is cheaper, the other company says its cheaper. Its the tale as old as time.
As a Linux advocate, I would like to say that I agree. While it’s nice to hear a counterpoint to MS’s ubuquitous “Get the Facts” propaganda, a good solid, truly independent study would be more persuasive to third parties. At least OSDL is up-front about sponsoring the study.
At least OSDL is up-front about sponsoring the study. —sbergman27
Agreed.
And THAT is why I usually feel more comfortable with an OSDL study than one that purports to be impartial only to turn out to be some sock puppet for Microsoft. Ideally we’d be able to find someone with out a bias or some gain or another who could attempt to be neutral or impartial, since that is just not an option I prefer to listen to OSDL because I already know what their bias is and am able to adjust for that. This isn’t an option with Sock Puppet consultant firm #23456.
–bornagainpenguin
Someone please correct me if I am wrong, this is not a flame.
Windows Management for 20 servers:
For each server:
1.) Load up Citrix and login.
2.) Point, click, wait, click some more.
3.) Set up and configuration done.
This takes a long time.
Linux Management for 20 servers.
Provided that you have ssh with key based authentication set up + ssh-agent running on your workstation.
for server in $(cat ~/serverlist.txt); do
ssh $server -c “command to run && command2 to run”
done
Linux allows me to have more servers with less administrators. Granted, you have to have more competent administrators, but less are required.
Edited 2006-02-13 19:52
Windows Management for 20 servers:
For each server:
1.) Load up Citrix and login.
2.) Point, click, wait, click some more.
3.) Set up and configuration done.
This takes a long time.
Well, actually.
Citrix is a company name ( http://www.citrix.com ) which has nothing to do really with remote management – instead it provides added value to (and is the original ‘driving force’ behind) the terminal server technology nowadays found in Windows.
I suspect you mean “Load op mstsc and login to the remote desktop”. Your proposed way is in fact one of the ways you can administrate an Windows 2000 and higher machine and is perfectly suitable for small scale changes.
For larger scale changes Windows has a very nice feature set in the way of standard RPC based management utilities (sc.exe is one in this category which allows one to control the service system on Windows systems), WMI/VBScript scripts (eventquery.vbs (in your %Systemroot%System32 directory on XP/2003) is a fine example of this) all of which are commandline based.
To really answer your question though – basically – the same statement on Windows could be done like this:
for /f %i in (serverlist.txt) do (
psexec %i command-to-run
)
(this assumes usage of psexec from SysInternals, although rcmd.exe from the Windows’ resource kit might suffice as well).
Personally i prefer using SysInternals’ excellent “pstools” tools which allows for some more control in how something is done and which is a free (as in beer).
Anyways – when someone is serious about administrating a larger set of Windows server you really should have the Resource Kit utilities and be aware of the SysInternals’ utilities – they are invaluable
Editted to correct a mistake involving rexec and rcmd
Edited 2006-02-13 21:26
Actually.. NIX Mgmt is more like this:
http://www.cfengine.org/
One tool that makes it easy for a single person to manage hundreds of systems all over the world….and its FREE, totally customizable, unlike stuff like Tivoli etc.
I’ve looked into that before, but the management is sceptical about deploying it. I.T.I.L. change management process is a major headache sometimes.
Great software though, +1 for you.
[ Note: a post about this study also appeared on lwn.net earlier today. Following is a cut-n-paste from my comment there. ]
Excellent! And wanted to point out also that most (all?) of the advantages sited in the study apply to *all* the open-source UNIX OS’s, not just Linux. For example:
– Open Solaris distros (http://www.genunix.org)
– FreeBSD (http://freebsd.org)
– NetBSD (http://netbsd.org)
To name three.
Eric Boutilier
OpenSolaris
http://opensolaris.org/os/blogs
Another study into TCO. Who is this for really for ,anyway? Organisations are so different and TCO is dependant on so many factors , is measured differently by different groups and is so difficult to accurately measure that the figures don’t have much ‘realism’ value to them. You have to ask yourself the question what for ?
The main factor that drives TCO is the <italics> person in charge of the implementation </italics>. He/She is the one who raises the TCO up or down due to the correct /incorrect usage of the technology being employed.
It’s stupid to make decisions based on TCO studies or extol the wonderful virtues of a particular product ,it’s also stupid to try and use these metrics to make objective comparisons of different systems. Every situation is different, and a brain inside the IT officer is the only and I repeat the only objective factor useful for such comparisons ….
I did not read the article but just reading the “Licensing Cost” section gives an idea about the writer’s attitude towards Microsoft …
ISA Server as “Web Server”!?
Internet Information Server (IIS) is the Microsoft Windows Web Server, and it is included into the OS cost. ISA Server is a Web Cache/Proxy/Firewall/VPN Solution but not a web server! and they even compare it against Apache!!
Take the Operating System cost as example: For Windows the “Enterprise” licensing cost is shown while for Linux they choose the “Standard” (AS) version, instead of the “Enterprise” (ES)
Perl and PHP are compared against the price of Visual Studio. Visual Studio is a high quiality development platform and Perl and PHP are language script last time I check their official web sites… nothing makes sense.
This study is so biased that the Microsoft “Get the Facts” papers look even great.
Yorch
Edited 2006-02-13 22:11
Take the Operating System cost as example: For Windows the “Enterprise” licensing cost is shown while for Linux they choose the “Standard” (AS) version, instead of the “Enterprise” (ES)
Wrong! Enterprise for MS and Enterprise for RH haven’t got the same meaning:)
Red Hat Enterprise Linux AS
The top of the line server solution for large departmental and datacenter workloads including database, ERP, and CRM servers.
Red Hat Enterprise Linux ES
Ideal for a wide range of entry-level and departmental workloads such as network, file, print, mail, and web servers.
From page:
http://www.redhat.com/en_us/USA/rhel/details/servers/
Also check the prices of those and notice that ES is the cheaper one:
https://www.redhat.com/apps/commerce/rhel/server/
Perl and PHP are compared against the price of Visual Studio. Visual Studio is a high quiality development platform and Perl and PHP are language script last time I check their official web sites… nothing makes sense.
Would you be happier if they named Mono, GCC, FPC, Python, Ruby, Java, Eclipse, NetBeans, Anjuta, Monodevelop, Nvu, Bluefish, Screem…?
Price is the same: $0
Would you be happier if they named Mono, GCC, FPC, Python, Ruby, Java, Eclipse, NetBeans, Anjuta, Monodevelop, Nvu, Bluefish, Screem…?
All of these (or at least, almost all) are available on Windows too. So what is your point?
MySQL runs on Windows, too. Apache, too, NetBeans, Eclipse and Java too… All of them cost $0 on Windows, too.
Just one thing: Mono and Monodevelop simply can’t be compared to .NET and VS. Mono is not even complete yet (WinForms is missing).
All of these (or at least, almost all) are available on Windows too. So what is your point?
My point was to ask rethoricaly: “What is your point?”
Just one thing: Mono and Monodevelop simply can’t be compared to .NET and VS. Mono is not even complete yet
One thing invalid.
So, .NET is not complete yet since it hasn’t got Mono namespace? And because it isn’t crossplatform?
I use .NET quite a lot. But only crossplatform parts. WinForms sucks here, GTK# rocks. MS.NET runs on MS only and this is why Mono is way more complete than MS.NET /*for me*/. You just have to consider ones needs to talk about completeness.
WinForms is missing
Missing? For who? I don’t miss it, I don’t use it, hell, I don’t even want it. I have to install GTK# under Windows too, since it doesn’t get installed by default. And since I don’t use WinForms, VS is just as usable as MonoDevelop for me.
Missing? For who? I don’t miss it, I don’t use it, hell, I don’t even want it. I have to install GTK# under Windows too, since it doesn’t get installed by default.
Why are they implementing Windows Forms on Mono then?
[i]And since I don’t use WinForms, VS is just as usable as MonoDevelop for me.
LOL VS as usable as MonoDevelop?
If I am not wrong, MonoDevelop is port of SharpDevelop. SharpDevelop is good, but is not as good as VS. Maybe one day.. maybe.
Why are they implementing Windows Forms on Mono then?
Beats me, but in a more serious tone: to be able to run Windows.Forms apps on Mono. Since Mono and .Net share binary compatibility, some (for example) Delphi (Delphi has complete environment tweaked for Windows.Forms) apps already work on Mono.
LOL VS as usable as MonoDevelop?
Yes, if you read my comment (where I specified that this is my personal, not public opinion) you can notice that I don’t need or want Windows.Forms designer, which is in fact the only good thing in VS. Oh, yes and I don’t use debugger if you think of putting it out next.
With Mono and MonoDevelop I get my environment out of the box. I have to bother extra to get it in VS.
Again, this is my personal bias
If I am not wrong, MonoDevelop is port of SharpDevelop.
Yes, you are wrong. Project started like that, but in early stage they completely split and gone their own way.
MonoDevelop != SharpDevelop
“Wrong! Enterprise for MS and Enterprise for RH haven’t got the same meaning:)”
You are right, I missunderstood the names (double fault since I support some of the RH boxes at work!
“Would you be happier if they named Mono, GCC, FPC, Python, Ruby, Java, Eclipse, NetBeans, Anjuta, Monodevelop, Nvu, Bluefish, Screem…?”
No. Let’s be serious here. None of the tools you mention is _by far_ as good as VisualStudio for software development.
Even if you are not a Microsoft lover you have to admit that.
Yorch
Nah, I probaly forgot to entag it in sarcasm. This point was flawed anyway. If you’re considering server side TCO. What is the point of specifying IDE?
Even if you are not a Microsoft lover you have to admit that.
Not that I want to bash VS. But I don’t find it usable. I preffer smaller special environments over bigger. Although this is my opinion (and I alays state when I’m posting my personal opinion) and I can understand why many people like it.
For example, I left out of Delphi for fpc and VS for GCC. Reason was the same. Simple editor of my choice, controlled environment with crossplatform ability. 99.9% of my time IDEs like VS and Delphi are too bloathed to be usable. But that is probably defined with my needs, which would say that GUI is about 0.1% of my work other is special clustered client-server HPC. I even had to write my own debugging and profiling environment to be suitable for me.
Fair enough?
If I run a simple home webserver, and I would use Windows Server 2003, I must buy a license and buy a PC that is able to run it. If I would use linux. I download debian and run it on an Pentium II 350 with 512 MB ram. Now choose, which one costs more.
What I am trying to say is: Total cost of ownership is relative to the ones using it.
Take a look at this:
Web Server
Microsoft ISA Server 2004 Enterprise Edition ($5999 per processor) = $23,996
vs.
Apache/JBoss = $0
MS ISA Server is NOT a web server. God, what are we talking about here?
Good study. No.
Wow.. somebody please tell me I am wrong about this too:
Microsoft Windows Server 2003 R2, Enterprise Edition (Includes 25 CALs) = $3,999
vs.
Red Hat Enterprise Linux AS 4.0 = $1,499
Shouldn’t the price for Red Hat EL be $1,499 per year?
CentOS 4 ( http://www.centos.org ) is the same thing, minus the mandatory support contract, for 0$ per year.
I use it and love it.
[couldn’t resist]
Wow.. somebody please tell me I am wrong about this too
Since you insist on MY answer.
Yes, you are wrong and no, it shouldn’t be. What is Enterprise for MS is not Enterprise for Redhat. AS is more expensive than ES
https://www.redhat.com/apps/commerce/rhel/server/
[/couldn’t resist]
Shouldn’t the price for Red Hat EL be $1,499 per year?
That’s a good point, and I would think they should also be including the annual fee for Microsoft’s Assurance support or whatever it is they’re calling it now.
I have no idea how MS support pricing works, but in my experience annual software maintenance contracts work out to anywhere from 10 – 40% of the value of the product, so in addition to the initial capital outlay for purchasing a license, you have ongoing annuity expenses that need to be factored in as well.
One thing about Microsoft is that in their arrogance, they often ignore companies or technologies that could pose a risk to their business.
The other thing is that once Microsoft figures out that a company or technology is a threat to it’s business, they will spend millions upon millions on clever marketing campaigns explaining why they don’t consider that company or technology to be a threat.
The mere fact that Microsoft is targeting linux with it’s “Get the Facts” campaign is reason enough for open minded IT customers to take a closer look and try and get the real facts. One can simply draw the logical conclusion that if Microsoft is telling you linux has nothing to offer, then linux must have something to offer that Microsoft doesn’t want you to know about.
That’s validation enough, why waste money on counter campaigns when those scarce resources could be better spent? MS is a marketing machine, taking it to that level is a mistake because you’re fighting on their turf against an unlimited warchest.
Just my 2c.