Once Linux overtakes Unix, as Aberdeen analyst Bill Claybrook predicts, Linux and Windows “will be duking it out for everything.” There will be “unbelievable competition between Windows and Linux” in areas like back-end database servers, he says.
Behind the Scenes: Microsoft’s Server Dominance
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
This article argues that Windows compatibility will be a major factor and corporations will choose a Windows over Linux for server software as a result. Samba 3.0 takes Linux another step forward and can do almost all tasks required in a Windows world from sending a document to a local printer to logging onto an Active Directory domain. I’ve used Samba for the last 2 years and 3.0 is a BIG step forward.
As GUI admin tools improve (for the newbies among us), these issues cited in the article will erode even further.
I won’t start a flame war over this. I’m a fan od FreeBSD over Linux, but let’s face the facts. Both are highly capable server OSes and each has a few small shortcomings and strengths compared to each other. In other words, at the server level, I think they’re both equals and the choice between the two is personal choice and in some rare instances, app requirements.
I think they’ll both be around, and used for a long time to come.
Again, they are comparing dollars against dollars, not installations.
I’m a Linux advocate and I don’t see your comment as flame invoking at all.
The way I’m looking at the matter is ….
When commonly speaking of the label “UNIX”, in most cases freeBSD is not counted into the figures becuase it is not counted among the likes of “proprietary UNIX”.
so looking at it from your point of view, I agree.
Interesting article, however, the first quesiton I wish to ask is this; What server application is missing that is available on Windows but not on Linux?
1) The article talks about a so-called “integrated approach”, could some one explain to me the real benefits of integration between Office and Sharepoint? are there any real world examples out there where real data can be pointed to as proof that such a model improves efficiency?
Most people I know, who have deployed Sharepoint have simply used one or two of the most basic functions, namely, the discussion forum and instant messaging. Both of these are provided by mature software already available. For the forum, all one needs is a database, say Posgres, PhP and Phorum.
2) People put too much emphasis on “initial cost”. If I were to move from one area to another, I could simply argue that I might as well stay in the same house because I would have to pay for stuff to be packaged up, trucked to the new house and then unpacked. Basically, it puts the idea that the property *COULD* increase in value completely out of the equation.
When moving from Windows to *ANY* platform, the cost is going to be higher. Why? because Microsoft is incompatible with the rest of the known universe. In the *NIX world, you can float form one *NIX to another with very small disruption, however, moving from Windows to *NIX is alot harder as Windows has not been submitted to a standards body, like UNIX, and thus documented so that other companies can implement the API.
What these companies should look at is the long term cost. Not one or two years but five and six years down the track. If one were to take a 6 year snap shot of the business world, just look at how much money was wasted on unneeded IT equipment purchased by over zealous managers who some how think that by buy more IT stuff will make their business operate more efficiently.
3) If the issue was purely based on ease of use for administration then MacOS X would win out over Linux, but that isn’t the case. Linux isn’t hard to administrate as a server. I am certaintly no guru, however, with a bit of reading and practice, I can how setup servers; FreeBSD/Solaris and Linux, without any problems. The fact remains that we have so-called “analysts” assuming that because something is slight different, there for it must be inheriently more complicated to use.
4) Good administrators cost money. The TCO worked out on Windows is absolutely bogus to say the least. It assumes that one can get a good Windows admin for a set price when in reality, that isn’t the case. If one were to get a good Windows admin, it is going to set you back atleast the same amount it would if you had UNIX.
Once you step outside the Bill Gates/Steve Balmer/Microsoft reality distortion field and start doing REAL hiring of people you actually realise that no, you can’t pick up a good admin for $30,000 a year.
5) Linux will come onto the desktop. SuSE is already a major partner both with IBM and SUN. The default desktop for SuSE is KDE and ultimately the end user, with in a typical organisation does not need anything more complicated than OpenOffice.org, Kroupware or Konqueror for their daily work.
I have NEVER see a end user create a database, write a macro or use a high end graphics tool. What there needs to be, however, is the ability to batch convert large numbers of templates, macros and documents over to OpenOffice.org format
Macros being the bain of existance, however, if basic templates can be converted from VB to OpenOffice.org BASIC with minimal fuss, the future for OpenOffice.org as a viable alternative over MS Office will be alot brighter.
6) Accoutability is another interesting issue that comes out of these sorts of discussions. When was the last time you, as a administrator, saw Microsoft take the issues you have raised with your sales rep as a serious issue? how many here have reported a bug and found that it is never followed up because you are referred to as the “unwashed masses” and considered “not important enough” to get “special treatment” or what an old school like me would refer to as “customer service”.
7) Although security is an issue, one has to ask, when was the last time you applied a security patch only to find that either you computer starts crashing radomly, slows down or worse, opens the computer up to an even worse attack.
As a person clued into IT, I can appreciate that operating systems are complex beasts and patches will need to be issued, however, making a patch and ensuring that half a dozen things don’t go wrong is a completely different thing all together.
Microsoft issues patches for vulnerabilities, however, what happens if the security patch breaks an application on the server which is a required piece of the organisations day to day function? or worse yet, the patch is applied and instability occurs? One only needs to look at the last Windows XP patch which caused simplier results.
….I find most pro-Windows advocacy articles insulting. I’ve worked in all manner of administrative work on all manner of computers, and I have *never* seen a mor arrogant bunch.
I used to like Windows, until I stopped working on my *duties* long enough to see what kind of god-forsaken culture I’d wound up in.
Funny thing is, I was infected with it for awhile; I didn’t get “cleaned up” until I got into Linux, needed help, and was told I ought to *learn to help myslef.*
I used to do that all the time, before I found myself slipping down the slippery slope of “point and click, drag and drop, let us handle it for you.”
You want to know what I think? I think that if Microsoft wins over Linux, it’ll be AT&T all over again, but in the computer business instead.
Except that the first time there’s a major economic collapese over a lot of lost data, the Corporate Boardrooms will ask themselves: “How did we get taken in by these people?” Then, they’ll want to do some of it themselves again.
Only by then, all the skilled people will be dead or retired (or sued out of their livelyhood.)
So there. Go be compatable and integrated.
The last two weeks of problems with our W2k servers have brought us to the end of the line for Microsoft (one employee opened an email before Norton had a virus id for then new Sobig virus) – We are a non-profit and have toyed with Linux some as we can get older computers free but they come from corporation that had site licenses and thus have no OS installed. While our W2k network is in constant need of my time – the Linux machines just keep on going without problems. We will not be purchasing any further product or licenses from MS and as the MS products become outdated or the computers running W2k fail they will be replaced by Linux or open source software products or proprietary software that works with Linux. Within about a year we will be switched except for one box left with W2k just in case someone has Win only software. We do have one program from a state agency that is a Win16 program but it only runs once a quarter.
No computer can work unattended.
Be sure you patch your linux box too. Yeah, the there are no “Blaster” like worms for Linux yet, but you don’t want to take the risk.
montra: Patch, Patch, Patch.
No computer can work unattended.
Be sure you patch your linux box too. Yeah, the there are no “Blaster” like worms for Linux yet, but you don’t want to take the risk.
montra: Patch, Patch, Patch.
But there is line between what is baby sitting and what is invesnive care. If one, as an administrator spends half their time intensively caring for a server, one really has to wonder whether his time and skills are being utilised in the most efficient mannor.
Sure, as an admin, I don’t expect an environment where by things “just work”, but I do expect the ability not to worry about something for longer than a few hours.
I’ve had this same discussion on a different board, and more or less we came to the same conclusion as you. In general the MCSE people are in the state of mind where they think that they are gods gift to the computer field, which is the furthest thing from the truth. UNIX users tend to be more inquisitive [sp?], and a bit more humble. There are of course exception in both camps, just visit any Linux/*BSD/UNIX bbs and you’ll find the vocal minority acting disnnoant toward inexperienced users. Then there are the Microsofties which visit ZD-NET, Slashdot, and OS-News and badmouth whichever( non Microsoft )OS in the headlines today. To be more on topic I think that GNU/Linux and MS Windows are aiming for different server markets. GNU/LInux is great if you want server with the smallest amount of overhead, customization out the ying-yang, and ability to peek under the hood. MS on the other hand is alot more useful to people who have very little or no experience administering a network/server type of set–up, and for those people who have money to blow on the latest hardware. Please don’t troll or start a flame-war this are my point of views and I don’t want to pass em off as fact, merely as an observation.
“Again, they are comparing dollars against dollars, not installations.”
As Aki pointed out the information is useless. (Unless you wish to infer that Windows is either; more used or more expensive.)
The problem is that people don’t realize that every product has it’s place. I hate M$ products, but I do realize that there are plenty of commercial applications written specificly for it that won’t be ported to Linux anytime soon. It would take such a large shift in the market that software ISV’s would have to rework their software to run on Linux. And this is where things get difficult.
You can definitely replace all of the productivity apps from M$ with StarOffice, Mozilla, etc. However, there are always apps that do a specific task that only runs on Windows. I’m sure that most of us have seen such applications that already exsist in companies that are stuck on Mainframes or on OS/2 systems. So it takes time to for things to be replaced by newer versions or better products. And change is something most companies avoid as much as possible, because it’s expensive.. especially in this economy.
I don’t think that Linux will replace the commercial Unices, like Solaris, AIX, and HP-UX. They are designed to run on very reliable platforms and are used for major application suites like Oracle ERP or SAP. I think what will happen is that the commercial Unices will take the place of the Mainframe systems of the world. And I see Linux taking over the simple tasks like file/print, proxy, E-mail, etc. And it’ll have it’s place on the desktop.
The big issue faced by Linux is that in order to take on the desktop, there will have to be a high level of integration, a better GUI, and ease of management. While most Unix ppl, such as myself, think that Linux can accomplish most of these tasks, we still have to realize the difference between windoze users and Unix users.
Windoze users are generally use to the OS taking care of things and it’s a “click okay” type of world. When something goes wrong, they don’t have a clue as to why. I see this happening on the Windoze side of the company I work for. You have a large staff of admins and help desk ppl supporting all of the Windoze users and servers. They spend most of there time putting out fires. The sad thing is that they have to spend more time putting out fires than being proactive and helping their users become more proficient.
Unix users are generally use to figuring out things on their own and typically will only have questions related to applications. Less of a UNIX sysadmin’s time is spent putting out fires. Most of it is spent being proactive and helping the users become more productive.
When I was an sysadmin at Sun, everyone used a Sun box as there desktop, nowadays it’s a Sun Ray. Most of the help desk tickets were questions about applications, account settings, office moves, desktop upgrade, and requests for restores. Because of this, we had a few admins to handle on-site help desk issues, while most solved these issues from their office remotely. While the senior level admins worked on day-to-day maintenance on servers and projects. The number of users, desktops, and servers that could be supported by a dozen admins was in the thousands across geographicly seperated areas. Compair that to where I’m at now, it takes a dozen windoze admins to babysit 300 users and about 30 servers. It’s pretty sad.
The one thing that I always thought was sad is that not many of the commercially available applications for Linux run on any of the other platforms Linux runs on, much less a version of BSD. And that is bad for the Unix community as a whole. If you wanted to run an application on your FreeBSD PPC box or your Tru64 DEC Alpha box, you might run into the problem that it’s only available from the vendor for x86 Linux. As such I think this breaks the choices a company can make as Linux becomes more popular. People will be force to run applicatons on a PC running Linux, instead of their perferred hardware and OS platform. While it doesn’t matter for the freeware GPL applications, since they can be recompiled.. it does matter with the commercial apps where you don’t have the source code.
On the patching issue. You should always patch your systems regardless of what OS they are. Now some OS patches are bad, it’s bound to happen. That’s why you should always test them out first before deploying across your datacenters.
But if you want to hear about some insane patching.. look at Oracle 11i DBA’s. They spend most of their time applying patches that are gigs in size!
I am not a fan of .Net. Take in to consideration that companies are constantly forced to upgrade there product for the latest M$ OS. When these companies upgrade to .Net there is a possibility to use the software under Linux, BSD, Unix, etc… giving people an alternative on products that are historically M$ based. I know .Net or mono is not ready under linux, but considering Novell’s purchase of Ximian and the demand for these applications increase, it may be soon.
I won’t start a flame war over this. I’m a fan od FreeBSD over Linux, but let’s face the facts.
Yes, BSD rae more safer, but Windowsa sevrer are eaasier to admin and to use (usability) … after 2 more years, Windows server will be much more suitable and easiyer to admin.
[i]The one thing that I always thought was sad is that not many of the commercially available applications for Linux run on any of the other platforms Linux runs on, much less a version of BSD. And that is bad for the Unix community as a whole. If you wanted to run an application on your FreeBSD PPC box or your Tru64 DEC A[7I]
That’s evolution over determinability !
Not to mention server side scripting. Linux has plenty (Python, Java, PHP … etc.) but so does Windows 2003 server , and many more (ASP …).
I have to see Windows gainining Market share and Linux gainning share at the expenses of Comercial Unix (Nothing new here!),.
>Not to mention server side scripting. Linux has plenty >(Python, Java, PHP … etc.) but so does Windows 2003 >server , and many more (ASP …).
I think this was because the article was about more datacenter or other applications that face the internal network rather than the external internet.
My take (just an opinion) is that while Linux can replace Windows in these positions (perhaps less easily than it can commercial Unix) in functionality the windows lock on the desktop is going to be a problem for that replacement to happen. The sysadmins will know that Linux may be a more cost effective choice, but will the CTO and accountants? Who are most likely not to have much knowlage about the alternatives and never to have used Linux, but they will know all about Windows as that is what they personally work on.
Linux will win!
Who wants to put money on it?
“I have NEVER see a end user create a database, write a macro or use a high end graphics tool. What there needs to be, however, is the ability to batch convert large numbers of templates, macros and documents over to OpenOffice.org format ”
Really? There are several people where I work that if they lost their macros it would decrease their productivity by almost 90%, and I’ve also have users who are looking to create a database to hold all of their expense data.
OpenOffice needs to create an OpenOffice document format that others can look at and implement.
Anyway, my biggest thing with *ixs in general is that they don’t inegrate well. We have several Samba servers running, but in order for people to get to them they have to reset their password to gain access. It’s a hassle, and the password sync program our SysAdmin created generally doesn’t work.
Does any one know if there is an *ix equivalent to Active Directory? What do I have to do to get an Unix domain running? I mean is that even possible? This is just out of curiousity. I may want to do it in the future.
It’s absolutely normal that Windows is losing shares in front on Linux on big datacenters. Everyone who tried to “patch” an SP on a Windows server know he will have a HUGE amount of compatibilities problems (reto actions, comportements behaviours,…).
In fact, keeping a Windows center “on the date” is more expensive than a Linux one. And in the last case, I saw less capacities problems and more efficien programs.
That’s why i don’t mention anymore my “Windows certification” on my resumee : lots of companies in Europe see you as an costing factor.
“OpenOffice needs to create an OpenOffice document format that others can look at and implement.”
The OpenOffice document formats are open XML formats, full specifications are available from OpenOffice.org. The XML document is gzipped so if you want to look at your own sxw’s just gunzip them and you have humanly readable XML.
Corel has expressed some interest in the format for Wordperfect and has given some preliminary support to Sun in its attempt to get the OOo/SO formats accepted as formal web standards.
“I have to see Windows gainining Market share and Linux gainning share at the expenses of Comercial Unix (Nothing new here!),.”
In dollar terms the rate of increase in Server sales for Linux,on a year to year basis, is twice that for Windows. Much of this is replacing proprietary Unix – but it is also effectively stopping wintel replacing RISC/UNIX by substituting lintel instead.
Linux is also replacing a lot of Windows servers as more windows shops crack and realize the advantages of Linux. In the big Unix shops Unix based management can use lintel edge servers as a way of getting rid of the Windows boxes that have been creeping in.
Another important factor is that Oracle now recommends Linux as the preferred operating system for running Oracle databases. This is having a big impact on companies that originally considered installing their new Oracle databases on Windows. They are switching to Oracle on Linux instead on the basis of advice from the Oracle Corporation (mine did).
The dollar server sales statistics ignores all the OS less servers sold which will almost certainly have Linux or a BSD installed on them not Windows (note you can buy OS less servers from Dell), plus what about all the existing boxes where an old copy of NT is wiped and replaced by Linux to upgrade the system.
Overall it is clear that Linux is replacing Windows in a big way in the server space not only proprietary Unix. We should not be talking about “Microsoft’s Server Dominance” but instead about Linux’s Server Dominance. Linux is now the dominant server operating system.
I think things will change quite a lot in the upcoming years.
First of all Windows becomes more competitive all the time and I would definitely think it’ll grow on behalf of Linux and some *nixes. Just like a few companies might try to adopt Linux on the desktop.
However in the longer run there is a demand for huge changes in the world of OSes.
First of all, SUN will definitely gain market share again as they, without doubt, has some really interesting solutions keeping TCO really low.
Secondly Microsoft will grow stronger on server side and loose on desktop to nisched OSes. I can see Linux taking a small share, OBOS/Zeta getting a bigger share (especially small offices) and QNX for instance picking up some part of the community.
Furthermore, I think we’re about to see a decrease in Linux activity on the server side and as security becomes an issue, systems such as OpenBSD will see a growing amount of users and FreeBSD will gain from that as well.
“I have NEVER see a end user create a database, write a macro or use a high end graphics tool. What there needs to be, however, is the ability to batch convert large numbers of templates, macros and documents over to OpenOffice.org format ”
Really? There are several people where I work that if they lost their macros it would decrease their productivity by almost 90%, and I’ve also have users who are looking to create a database to hold all of their expense data.
So you’re telling ME that there are users out there who write VBA macros in their spare time? the same users who don’t know what a right mouse button click is, the same users who are confused about using the internet?
OpenOffice needs to create an OpenOffice document format that others can look at and implement.
Already exists. The OASIS File Format Standardization and Inter-Office Cooperation – http://www.oasis-open.org/committees/tc_home.php?wg_abbrev=office
Microsoft was given an offer to join but they declined. Members include Arbortext, Corel Corporation, National Archives of Australia, Society of Biblical Literature (SBL), Sun Microsystems and The Boeing Company
Anyway, my biggest thing with *ixs in general is that they don’t inegrate well. We have several Samba servers running, but in order for people to get to them they have to reset their password to gain access. It’s a hassle, and the password sync program our SysAdmin created generally doesn’t work.
Does any one know if there is an *ix equivalent to Active Directory? What do I have to do to get an Unix domain running? I mean is that even possible? This is just out of curiousity. I may want to do it in the future.
Completely incorrect. One could do either of these things:
1) Wait till SAMBA 3.0 is released which includes support fo LDAP which is supported by Active Directory.
2) Install eDirectory ( http://www.novell.com/products/edirectory/ ) which is available for Linux and since it is LDAP certified, it should work nicely (in theory) with Active Directory.
This is ridiculous. Sure, Linux is taking the place of ailing comercial Unix derivatives like Irix or AIX, but the BSDs are formidable opponents that don’t seem to be going anywhere. Furthermore, the author uses money earned to determine how well Linux is doing. Hello? Linux is free remember? What’s more, compatability is not such an issue for servers. No one wants to run the original WordPerfect or a Dos game from the 80’s on servers, hence, compatability is a virtual none-issue. What’s more, neither Linux nor any other OS in question has this “dearth of applications” he accuses Linux of. They all have the key ability any server needs: to serve, be it webpages, CGI apps, or whatever else. The only compatability issue would be if MS .NET were actually going anywhere. It seems almost as though this author read a string of headlines and first paragraphs from other publications and tried to form a coherent opinion on the matter. He clearly lacks the understanding necessary to provide intelligent commentary, much less predictions in this domain.
“I have NEVER see a end user create a database, write a macro or use a high end graphics tool. What there needs to be, however, is the ability to batch convert large numbers of templates, macros and documents over to OpenOffice.org format ”
With regard to databases the experiance in my company (Canadian subsidiary of a medium size global corporation) is that there is a lot of user created databases out there. We have a large number of Filemaker Pro databases (running on local file print servers) that were created by users at the departmental level and used to a certain extent horizontally across departments. Probably more commonly in other companies databases of this type.use MS Access but we are using Filemaker for historical reasons.
It is now corporate policy to migrate all these local databases to Oracle databases administered by MIS at the company level. Currently we are know using Oracle on Linux. The more relevant and important databases are in process of being migrated. But how we do this over the long term with all the dozens of departmantal databases no one is quite sure.
Though I am not so sure about macros, I expect there are a lot of Word macros created by secretaries to partially automate production of documents in the corporate style