Linux Myths by Micro$oft    

Others:
   daviddennis (david@amazing.com): The big advantage of that 30-year development is that we know it works, and works well. It's been honed to the point that there are minimal amounts of bugs. That's a tremendous advantage over NT, an operating system that was created under strict deadline pressures unimaginable in the Unix and open source worlds.
The foundation works. There's little point to changing it, especially since it outperforms NT as it is in many if not most situations.

   Anonymous Coward: That's right. Unix _has_ been around for 30 years.
Unix _has_ proven itself in the mission critical enterprise sector.

"Linux was not designed from the ground-up to support symmetrical multiprocessing (SMP)"
"Windows" wasn't initially designed to run on servers. So what?! That means nothing. It has SMP now right?! What's your point?

"These architectural limitations mean that as customers look for a platform to cost effectively deploy scalable, secure, and robust applications, Linux simply cannot deliver on the hype."
"Architectural limitations"? You better re-read what "open-source" means because you haven't got a clue. "Secure"? Microsoft _dares_ to even mention the word "secure" with a straight face?! Robust = Mission critical = Mainframe, AS-400, Unix, etc... Robust? You mean like SP2, SP4, SMS 2.0? (...ad nauseum)

  Anonymous Coward: When was NT written from the ground up to support SMP

   Dwonis (dlitz[IBoycottSpam]@cheerful[spamsucks].com):First, it's worth noting that Linux is a UNIX-like operating system. Linux fundamentally relies on 30-year-old operating system technology and architecture. Humans fundamentally rely on billion-year-old architecture in their genes, so they should be scrapped and reinvented.
Linux started out sleek, lean, and unbloated, and then evolved into a viable OS. In contrast, NT started out with bloated features you don't need in a server (running a GUI 24/7), and can barely support its own weight (finding bugs in NT is like finding a blade of grass in a cubic acre of hay).

   Anonymous Coward:
in the 70's.

• GUI.
• OOP.
• Networked personal computers with email etc.
• Videoconferencing.
• UNIX.
• VMS.
• KeyKOS.

• Don't believe study made by company bought by M$ (ZDnet).
• Nothing is black'n'white. In certain situations it most certainly does ( Linux performs better than Windows NT). And in other situations NT is better. (Anonymous Coward)
• Sweeping statements about performance are unenlightened, ignorant of reality and quite frankly juvenile. (My car's faster then your car) But hey if you pay a company to prove yours is better you usually get what you pay for! (Mindcraft) (BTW, that's such an appropriate name) (Anonymous Coward)
• Of course put Linux on a Sun server, consult some SAMBA pro's and let's see who wins that one.
• And there are _other_ published (Non-Microsoft sponsored) results by other publications that show Linux the winner on single processor systems. Why aren't those results listed here as well? What do you have to lose in such a _clear_ situation?

  " Linux performance and scalability is architecturally limited in the 2.2 Kernel. Linux only supports 2 gigabytes (GB). Linux SWAP file is limited to 128 MB."

  Bryan_K (bryan_k@liNnOuSxPsAtMart.com):"The largest file size Linux supports is 2 GB versus 16 terabytes (TB) for Windows NT 4.0" Does NTFS even support partitions this large, or is this just the theoretical limit of the operating system? I know NT has large file system support, but didn't think it could scale this large.

   JPelorat (pelorat@netscape.net): _A_ Linux swap file may have a limit of 128MB, but you can have _more than one_ swap file, which is probably on a completely different paradigm than that of the Microsoft developers.

  Anonymous Coward: NTFS has greatly increased the size of files and volumes, so that they can now be up to 2^64 bytes (16 exabytes or 18,446,744,073,709,551,616 bytes).

But note that they also say...

NOTE: Underlying hardware limitations may impose additional partition size limitations in any file system. Particularly, a boot partition can be only 7.8 GB in size, and there is a 2-terabyte limitation in the partition table.

  Kintanon (sleffer@hotmail.com): Well, gee, let me see, I'll just hook up this 20 terabyte drive I have laying around and test it...
I mean really, you can't refute a claim like that because it's just ABSURD... NO ONE PERSON has 16 terabytes of anything... Yeesh, they can say this with complete security since it will be at least 2 years before anyone can actually test this.

  James M Rogers (jrogers@visnetinc.com): The Linux filesystem does support larger files than 2GB but your applications need to use special libraries to access above 2GB. Or you could just use a raw partion anyway you want. Sorry, wrong again, we can use 2GB swap partitions and even when we were limited to 128MB swap we could run 16 of these for a total of 2GB of swap space. And Linux does have asynchronous I/O and fine-grained kernel locks. Linux also has real time support at the kernel level, does NT? Linux also supports at least a dozen different file systems.

  Abattoir: Point - Microsoft Windows NT 4.0 has been proven in demanding customer environments to be a reliable operating system. Companies such as (blah blah blah). Rebuttal - Never heard of BellSouth? Berlington Coat Factory? GENERAL MOTORS?!
Point - Linux lacks a commercial quality Journaling Filesystem. Rebuttal - Where's the journaling filesystem for NT? And apparently Microsoft isn't keeping up with SGI... XFS under GPL? I think I heard a rumor about that...

  dirk: As for the reliabilty of Linux, it is very reliable, but there are no real studies (that I've seen). This is a valid point. I run an NT network where I work, and it runs well. I'm not going to say there is never a problem, but there are no major distasters on a weekly basis that many people would have you believe happen. I have also run Linux and think it's about as stable as they come. But anecdotal evidence will never be able to compare to real-world studies. Just because my network runs well doesn't mean all NT networks do, just as your Linux system running well doesn't translate to all Linux systems.

   Anonymous Coward: No myth there. Call it anecdotal. Call it perception. Call it anything you want, but when organizations make it a part of "Standard Operation Procedure" to reboot their NT boxes at a prescribed time interval. And Unix 'uptimes' are interrupted only by hardware upgrades. Well... _you_ figure it out.
Microsoft availability? There's a joke signature that's gone around for the last few years that reads:

"Your mouse has moved, your system must be restarted in order for the changes to take effect".
Pretty telling I think.

Windows NT 4.0 has been proven in demanding customer environments to be a reliable operating system. Customers such as Barnes and Noble, The Boeing Company, Chicago Stock Exchange, Dell Computer, Nasdaq and many others run mission-critical applications on Windows NT 4.0.
It _can_ be done with NT on the _right_ hardware. Certified by a server manufacturer like Dell, HP, IBM, etc... But generic run of the mill hardware seems to be more stable with Linux. And it takes far less hardware resources to run Linux then NT. (A bonus for small businesses)

  slykens:I don't know about most people, but I don't need PC Week or Ziff-Davis to tell me Linux is more reliable than Windows NT. I refuse to run NT on my network, and my servers have uptimes exceeding 100 days, and in a few cases, 180 days. Our test NT servers never made it past 7 days.
Then this, which is disinformation (as our ceo calls it):
Linux lacks a commercial quality Journaling File System
What about the fs that SGI is contributing?

  jeremy f:There are no OEMs that provide uptime guarantees for Linux, unlike Windows NT where Compaq, Data General, Hewlett-Packard, IBM, and Unisys provide 99.9 percent system-level uptime guarantees for Windows NT-based servers.
Why do we need OEM guarantees? With a properly maintained linux system, 100% uptime can be achieved.

  Clairvaux (boss@bigiron.org): As someone once pointed out on a builder.com thread, 99.9% uptime sounds impressive until you actually calculate what it really means in hours of downtime. There are 8760 hours in a year. 0.1% of that is 8.76 hours. So let's turn that statement around to say that Compaq, Data General, HP, IBM, and Unisys provide 9 HOURS OF DOWNTIME PER YEAR GUARANTEES FOR WINDOWS NT BASED SERVERS. ^_^ Yes I know that statement is logically faulty but it sounds funny, and you get the message

   ult: 2) Myth: NT is more reliable than Linux,
I keep hearing stories from NT admins about how they dream of that 99.95% uptime but just can't achieve it, at least my newest Linux box (6months old) has never once crashed and has only ever been down because I kicked the power plug out :)

  Anonymous Coward: It has? Were these _studies_ also done by Mindcraft? (Microsoft bought and paid for?) And as both a NT (MCSE) and UNIX administrator I'd have to say administration of NT _is_easier on the surface, but when something goes wrong it's _more_ difficult to troubleshoot _and_ it lacks the configurability and remote administration capabilities built into every copy of UNIX. As far as how this can be measured in cost? I'd call it _even_ since UNIX generally requires _less_ administration once setup.

   mikemacd: It really is amusing that they cite a document which compares the costs Sun Servers to WinNT and then conclude that Linux servers cost more than WinNT. By their logic Linux must therefore be equivalent to Sun. Wow. What a compliment.
Seriously though I don't think that any of us will disaggree that sun servers can be expensive. They are, and rightly so. But to equate Sun servers with Linux servers does an injustice to both. They are each designed for different markets even though they can do similar tasks.

  Spurius: "How many certified engineers are there for Linux"?
And this means.... what? How does a certificate make you better qualified? If you can fix a problem, you're qualified to fix a problem. How many solaris-certified engineers are there? AIX-certified? VMS-certified? MacOS-certified? Be serious. Only Microsoft feels the need to pass out pieces of paper to make people qualified.

  =w= (teges@bigfoot.com): Thank you Bill for the information on TCO. My boss and I will be buying mail server and name server software for NT Server. That will be one NT Server license mail server license and a name server license plus the hours it will take for me to learn how to use these tools, versus Linux dist. (free) sendmail (free) and bind (free) plus the time that it takes for me to learn it.
I am sorry Bill I am not a math whiz how is the NT Server solution TCO less than Linux TCO?

  grae: It does if "when we speak of free software, we speak of freedom, not price". Unless the latest version of NT comes with working source code, I can't tweak the OS, or any underlying components to my liking.
Microsoft doesn't care about free software, and many corporations don't either. They care about how much money they're spending. The point MS was trying to make here is that TCO is less for NT than for UNIX. Since Linux is about as easy to administer as any other UNIX-ish system, it follows that it costs about as much money to hire someone to maintain your servers.
I think it's important to address the points that they bring up, rather than talking about the benefits of free software. We all believe that free software will cost less in the long run, because people will find (and patch) the bugs in it. Unfortunately, I don't think there have been any studies done to quantify this...

   Tau Zero (spherethis@youknownottoincludethis.yahoo.com): Of course, the entire page refuses to address NT's TCO, like:

• Total system re-installs when something hoses the registry.
• Endless waits on hold when you have a problem that requires a tech-support call.
• The wasted time when the tech-support information turns out to be wrong.

  fart_face (tdurbin@nospam.dnai.com): Hey. Not like I need to point this out to many people, but there is some stuff in there that is plain wrong. I can speak from experience that Linux ( or BSD ) do provide a lower TCO for development organizations and ISV's. There's no disputing the fact that hiring one competent admin for a Linux or BSD solution is still far cheaper than what it will cost keep all the licenses kosher for WinNT WS with MS Office and an Exchange Server group for an office of say, 200 people. You may have to spend some contracting money up front to get it implemented correctly, but after that it's all maintenance ( paying the admin's juicy salary ).
Sorry Micros~1, but your TCO argument's bogon emissions are dangerously high, like - it's glowing blue.

  dclydew (dclydew@interhack.net): Next we have some FUD about TCO... they don't actually show any numbers... just point in the general direction of classic UNIX and say, it's probably like that!!!

   bhurt (bhurt@visi.com): The only TCO study I've ever seen showing that NT is cheaper than Unix (which was included in the MSDN, BTW) among several other humorous assumptions, assumed that anyone using Unix on the desktop also needed to have a Windows box on their desk as well for office apps, email (Unix doesn't have email, don't you know...) etc. In other words, TCO(NT) = TCO(Unix) + TCO(NT). Pay no attention to that man behind the curtain!

  Anonymous Coward: Oh dear, they resorted to comparing TCO with *UNIX* not Linux and then saying there is no reason to believe that Linux is significantly different. So the fact that Linux is free and most other UNIX Operating Systems cost quite a bit seems to have escaped them. Microsoft must think people are very naive. I hope this page backfires on them, as it should.

  drix: Really? How bout the fact that Linux is free, morons. Oh yeah, and it runs on non-proprietary hardware. This means that, while the hardware available won't be as incredible as that offered by Sun, you don't have to pay for a proprietary system and operating system. This is, in my opinion, a blatant lie. Just a paragraph before they long for "Real World Proof Points Rather than Anecdotal Stories," but then throw out this crud without a single bit of real world data. Hypocrisy - ain't it a bitch?

  Anonymous Coward: The cost of the operating system is only a small percentage of the overall total cost of ownership (TCO).
I agree. The cost of upgrading your computer every two years to keep up with WinNT system requirements is much higher than the cost of NT itself.

  Spurius: LAF and the main difference-causing factor in their TCO study?
The cost of a solaris system! LAF ; ) Most of the TCO difference in THEIR OWN STUDY was the system cost. And they seem to be pulling the rest of their numbers out of thin air, there are no references, just vagueness.

  Spurius: The line that says "Linux Security is all or nothing" is blatantly wrong. Its either a lie or they know nothing about privilages.

  JPelorat (pelorat@netscape.net): And what the heck are groups for, if Linux security is all-or-nothing? Hehehe..

   Dwonis (dlitz[IBoycottSpam]@cheerful[spamsucks].com): Oh sure, Linux security is all-or-nothing. I guess my home system doesn't exist.

  ryanr: Linux only has protection for file system... umm, yeah, that's all there is. :) No registry, guys. Says Linux is all or nothing for delegating administration. Go do a web search on sudo. States Linux doesn't meet any of the "key security accreditation standards" Like NT. Says NT meets C2, or british equiv. C2 isn't anything to be impressed by. Says Linux people must spend time understanding security issues. NT users can go to one spot for patches. Boy, I'm glad I don't have to actually *understand* anything when using NT. Says NT is easier to configure properly for security, using the SCE. You must be stoned. Have you looked at Sutton's NSA paper on what it takes to secure NT?

  jd: Linux security -isn't- all or nothing. Groups are your friend. Simply make an admin group, for a specific set of responsibilities, and chown the necessary software to that group. If necessary, use suid to grant the programs necessary privilages, but use sparingly. Actually, NT 4.0 failed the FIPS security standard, thus failing C2 accreditation. No Linux products? That's sweeping. Actually, all Netscape products for Linux are approved for Government use in the US. (Netscape's SSL is, to the best of my knowledge, the only approved SSL system. NT's IIS is NOT an approved SSL system. )

  This horse has been beat to death. (BTW, show us the C2 certification for NT-4.0, where is it?)
But since you brought up security and how it _is_ a problem. Shall we enumerate all the Microsoft security holes and patches that have been discovered in 1999 alone?
Shall we mention that the US Army was so fed up with this that they moved their important systems to Sun's Solaris?
Shall we discuss what a security nightmare everyone thinks ActiveX is?
Or that NT was just mentioned in a _major_ trade magazine article as having security problems to such an extent that numerous companies are rethinking deploying Windows 2000?
Or the Navy vessel that had to be towed home. Or.. Or.. Or..
This was a topic best left untouched by Microsoft.

  Sesse (remove.this.please.sgunderson@bigfoot.com): They claim Windows NT has had C2 security since 3.51. In that case, somebody should fix the C2 standard. Easy example: take the MTF (Microsoft Tape Format) specs -- I believe it's somewhere on Seagate's web site. Throughout the document, there are multiple provisions for"C2-level security". And their password protection? Oh, that's a byte to or from. That's right:
FLIP ONE BYTE, and ALL security on the tape goes away... Great! Go NT security!

  TheKodiak (kodiak@flail.com): "The National Security Agency (NSA) lists the following configurations as having been evaluated for U.S. C2 certification: Windows NT 3.5 with Service Pack 3 on the Compaq ProLiant 2000 and ProLiant 4000 Pentium systems, and on a DECpc AXP/150. The LSEL diagnostics software was used as part of the configuration. The systems were certified only in a stand-alone configuration (no network). No other version of Windows NT has received C2 certification on any hardware platform (Windows NT 3.51 was C2-certified in October 1996, but only in the United Kingdom)."
http://www.zdnet.com.au /zdnn/content/zdnn/0923/2140612.html
"To date, Microsoft has not obtained C2 certification for any release of NT beyond version 3.5, the company acknowledges."

  demon (dpates@DONT.SPAM.ME.dsdk12.net): Actually, it can only qualify as C2 secure with no network connection and no removable media whatsoever (that includes floppy, CD-ROM, Zip, Jaz, or any other removable-media storage device).

   Pathwalker (hungerf3@cse.msu.edu): To the best of my understanding, normal Linux systems would fail C2 certification as chown allows root to give ownership of files, not just take ownership of the file. C2 requires that no one can give away ownership of a file, you can only take ownership of it from someone. (this is to prevent someone taking ownership of a file, reading it, then giving ownership of it back)

  Spurius: Their claim of NT supporting 39,000 systems and devices sounds rather strange... I'll have to look up their compatability list (also strange they didn't link it). As far as I was aware, NT has been very shoddy in its hardware support.
Their claims that linux doesn't support Plug&play, USB, and APM are all completely wrong.

  Anonymous Coward: So NT supports USB and Power Management? Which Service Pack was that?
Because my TechNet subscription didn't include _that_ one! Are you saying because the _next_ iteration will include it? (How Redmondian of you) Guess what? The next iteration of Linux will address those issues as well.

The Linux operating system is not suitable for mainstream usage by business or home users.
Guess that must be why it's usage keeps growing and that 17% of all new servers are now Linux based.
Amazing isn't it? And Linux _is_ suitable for mainstream usage just as soon as there's a replacement for MS Office. StarOffice is pretty darned close I'd say.

  Dwonis (dlitz[IBoycottSpam]@cheerful[spamsucks].com): Linux does not support important ease-of-use technologies such as Plug and Play, USB, and Power Management Sad, my Linux box seems not to exist.

   DroolArt (DroolArt@hell.com): Ok USB isn;t supported to my knowledge, but NT does not support it either. Plug and Play under NT? ha! By deafult it is turned off, and the ONLY thing I ahve ever seen it work on is a soundblaster 16. No Power Management? RH 6.0 seems to manage power just fine on my Thinkpad 600, didnt even have to tell it it was a laptop, it just did it. Yet more half-truthes and lies told by the big softies...
I was once accused of being wrong. The bodies have been hid.

  jeremy: Lie lie lie, lie lie lie. Pnptools, USB as a kernel module, power management as a kernel modules (yes, I do believe that when a system goes to init 0, it CAN automagically power off). If developers would release drivers for their products in linux and stop cutting corners in production cost (winmodems, winprinters, etc...), any differences in the range of supported hardware would dwindle to a big fat 0.
You know, I do have respect for Windows as an operating system. However, any respect for Microsoft as anything more than a hardware device manufacturer has just went down to nothing.

   DunbarTheInept (SPAMBLOCKmadings@bmrb.wisc.edu.SPAMBLOCK): Windows doesn't run on Macs, nor on Sparcs, nor on Alphas (not anymore), and so on. If you remember that fact, and start a phallus-waving contest of comparing the number of pieces of hardware that are supported, Windows won't look so hot anymore. Windows only supports more hardware if you limit your comparasins to the Intel PC. (And even then WINDOWS DOESN'T SUPPORT THE HARDWARE!!! THE HARDWARE SUPPORTS WINDOWS. Microsoft isn't writing those drivers, the hardware manufacturers are.)

  Abattoir: The complexity of the Linux operating system and cumbersome nature of existing GUI's would make retraining end-users a huge undertaking... Rebuttal - This is about the most valid (albeit limited at that) statement they make in the whole article. It isn't the complexity of the operating system though. The operating system (in the UNIX world) does not have bearing here. The OS doesn't control what left mouse click and Control-Alt-F4 do. The fault lies in INCONSISTANT (!!) user interfaces. Windows isn't consistent all the time everywhere, however, it does provide a lot of consistency in places such as Menus and short cuts. A user in the Windows world can almost be garunteed that Control-C will copy and Control-V will paste. That the File menu is next to the Edit menu. That pressing F1 will get (semi weak in my opinion) a Help Dialog. Sure, KDE and GNOME are trying to fix this, but they use different conventions. And not everyone uses KDE or GNOME. They're bloated. Yup, I said that. It costs less system resources to ignore a desktop environment and just use a slick window manager (Be it blackbox, icewm, window maker, or twm). But what consistency that is present in the Desktop environments is lost and you're in a worse problem. Blah blah blah, I could go on this for quite some rambling, but I won't. This topic has been beaten to death on every forum newsgroup mailing list and IRC channel known to the Linux community.

  CNN April 6, 2000 (1436 GMT) by Joe Barr: What does Microsoft's legal loss mean for Linux? ...Regardless of the eventual outcome or the length of the legal process, Linux has proven that it can gain market share in the server space faster than Microsoft, even with Microsoft's unregulated monopoly in place.
Can Linux do the same in the desktop space? I don't believe that it can, at least not until the market is truly open to competition. But Linux can and will continue to expand in server the market, with or without justice being done in a timely manner in the Microsoft case. Linux has been doing the impossible for a while, and carving out a bigger and bigger slice of the pie for itself based on its growing utility and popularity. When the day comes that the playing field is level, Linux will be there waiting, stronger than ever.

   Mr. Young from Red Hat: " My goal is not to make our company as big as Microsoft, but to make Microsoft as big as Red Hat. "