[ale] Linux SuSe vs Windows 2000 Server

hbbs at attbi.com hbbs at attbi.com
Fri Mar 7 10:33:56 EST 2003


To a great extent, Geoffrey's sentiments echo mine.

Laurie, one thing about me that you should probably know about me when
evaluating anything I have to say about Linux-vs.-Microsoft decisions is that I
was once a *BIG* Microsoft booster.

I started my career working with VAX/VMS systems, first as a software engineer
and then as a system administrator.  As far as home computers go, I started with
a mere Sinclair and I picked up an Amiga 1000 in late 1986.  That last detail is
important in that unlike the Apples, Ataris, and IBM PCs of the day, the Amiga
ran an honest-to-goodness multitasking OS with a surprisingly capable
interprocess communication capability and a windowing GUI.  

>From that standpoint (i.e., VMS and Amiga), Win3.1 machines appeared a little
bit crude, but when WinNT came out, what I perceived was an OS that was almost a
spiritual combination of both VMS (this is no accident) and AmigaDOS for Intel
hardware.  I had a production NT Server system running in 1995 when I was
working for the US Department of Energy.  I built and bought a progressively
heftier series of NT servers and workstations and by the time I left, I had
created an NT-based infrastructure that supported an entire DOE field office.   

I should also add that WinNT pretty much filled the feature set of Banyan VINES,
a network server OS that I had worked with extensively in the early 1990s.  

During that time, I knew Linux existed but it wasn't until the mnnths before I
left that I really started to care.  I had become closely associated with WinNT
and had done all kinds of things with it - videoconferencing servers,
multimedia, RDBMS, vertical-market client-server apps - but I was starting to
get disillusioned.  

Running and working with VAXen, when you got into a jam, you called DEC tech
support and people who were sharp as a tack would be on the line iwth you in a
couple of minutes.  Get into a jam with NT, and MS would prefer to send you to
your hardware vendor (what if you built your own server?).  Hardware vendors
tended to be ineffectual and ineffective, which bugged me to no end.  However,
what bugged me even more was MS' even more standoffish attitude, as though their
attitude about things like hardware drivers were "release and forget;" something
about your problem might show up in the Knowledge Base or maybe not, and in any
case, hardly anyone seemed to be able to actually help you with a problem and
you couldn't even *buy* your way out of that.

The perception I had had of MS being a benevolent force that was going to lead
us away from six-digit mainframe support costs began to curdle.  I began to see
non-standard ways of doing what had been standard things (DNS/WINS, IP/NetBEUI)
and a seeming rejection of "fitting in" with enterprise networks (although
certain pieces were friendlier, like printing to LPD printers).

Fast-forwarding to the present day, I know too much to ever recommend an all-MS
enterprise IT infrastructure with a clear conscience, and that's *without*
taking into consideration that MS has been twice convicted in the US courts of
running an illegal monopoly and aiming its giant cash gun wherever it needs to
even today (witness their attempts to bribe the Indian government away from Linux).

I do not see Linux as being THE answer; I think that the BSDs deserve notice and
consideration as well.  But, more to the point, I have come to see the value of
the whole Open Source Software concept, regardless of exactly which OSses are in
play.  

Let me give you a real-world example (albeit one that never actually came to
fruition although the point is every bit as valid) that shows how this affects
me as an IT manager.  I was researching the design of a document image
management system that would have utilized very serious scanning hardware -
think high-end Fujitsu units.  And there would have had to have been a lot of
them, because the number of documents being scanned in one engagement would have
been on the order of 10^6.  Naturally, if you go buy one of those scanners,
you're going to get Windows drivers and some operating software on CD.

Now, because I need these scanners and the computers that run them to be part of
a larger business operation, I might need to have some say-so over just what
it's like to run those scanners on a real job.  My scanner operators are likely
to be only "lightly skilled;" I need a lot of them and I don't want to have to
pay them a fortune or have to spend days training them.  Also, I want them to be
able to scan docs very quickly because I'd be paying them by the hour; having
them futz about with a mouse and a keyboard doesn't help my cause.  

At this point in my story, just look up the two roads that fork ahead, one Win
and one Lin, and think of what my life will be like on each.  For purpose of
argument, suppose that the scanner-model-specific driver in both cases (under
TWAIN for Win and under SANE for Lin) doesn't quite work right for what I need
(either buggy or doesn't behave like I need it to).

The software that comes with the scanner with invariably be Windows-only and it
will make assumptions about the operator and his/her knowledge and capability
that may be at odds with my purposes.  Likewise, Linux' xsane app will do the
same thing.  BUT, what are my problem-solving paths foward?

  Win:  Buy Win app devel tools/compiler; write new TWAIN front end
  Lin:  Use app devel tools/compiler already on hand; modify 
        existing xsane 

Either way, I need good programmers.  However, the barriers to actually getting
work done the Lin way are far lower.  Suppose I don't even want to manage a
programmer (they consume all the Dew in the office anyway :) ); what do I do?

  Win:  Call Japan to find out who you have to pay to get a modified 
        TWAIN front end; you'll likely never speak with the programmers 
        (not that they'd necessarily speak English); you'll have to pay
        whatever they quote you
  Lin:  Untar the xsane source code, get the programmer's name, call him 
        up or e-mail him (could be a her, sue me), offer him $2500 to code 
        up a list of changes

Now, you tell me how you would rather conduct your working life!  In the Win
way, everything about the external relationships you have to transact within is
on the other party's terms (who you don't know and probably can't speak to) and
you have to pay them an awful lot to get them to raise their amount of caring
off of the zero peg.  In the Lin way, terms are arrived at mutually and to
*some* extent, the programmer is incentivized to do your bidding *just because
you are interested in his abilities and gave him a challenge*!  Add real money
to the equation and you're very likely to get what you need done!  Personally,
I'd rather not go around, hat in hand, ready to be shaken down at every street
corner.

Laurie, I can just about flat-out guarantee you that if you go the Linux route
instead of the Win2K route, even if you wind up in the same place functionally
speaking, you will know more, understand more, and be able to accomplish more at
the end of the day.  I also contend that your overall mental bandwidth will be
less directed at the OS and more directed at the apps you run on it and how you
can actually improve things at your place of work.  

I have a Linux/Samba file server at home that Just Works, much like others have
said.  Its uptime appears to be limited by its hardware (specifically, fans),
its electricity supply (it's on a UPS), and the need to move it between houses.
   It blows even large files (>500MB) around itself and other machines on the
network flawlessly, even though it's running ReiserFS from back when it was
considered a bit dodgy.  It's even using a crappy Gateway-2000-OEM Slot A
motherboard.  I've worked in business that didn't have an IT resource like this
available to it, and the box probably doesn't have quite $500 in parts in it.  

I can take other people's castoff here-just-take-it JUNK and create real working
business IT infrastructure out of it for no money.  I can make a print server
out of a 486 that isn't any worse in the job than a P4.  There are LOTS of
businesses that are just dumping server hardware with nice hot-swap disk bays
that can be refitted very easily.  If you work in a place that utilizes some
kind of sequential repetitive processing due to the nature of the business,
imagine being able to make a Mosix cluster out of JUNKERS that you can build and
expand for nearly no hardware cost.  

I could go on and on about aspects of the difference between working in WinWorld
and LinWorld but I hope you get my point.  It's not JUST the money and it's not
JUST the ethics - it's the ability to make choices and create what you need. 
It's about not having the software being developed, released, and limited in
such a way to continually vacuum money out of your company (to the point where
that seems to be as much its purpose as anything else it's actually designed to
do).  

It's about having apps that seem to be a bit twittery on purpose.  Lots of
businesses create Access databases and inevitably seem to hit that wall where,
either due to its inherent limitations (usually having to do with locking) or
from being just plain buggy, a whole marketing machine awaits to take you to MS
SQL Server Win2K Server, and a suite of MS-only app design and deployment tools.
 It *serves MS' purposes* for Access to be limited and buggy; their customers
have to buy it ALL to move forward (Access included, which almost always would
have been bought and/or licensed PER SEAT as part of the $$$ MS Office suite). 
Who wins?  Who loses?

- Jeff
> Greg,
> 
> I find your whole post quite appropriate and I thank you for providing, 
> probably the most balanced input available on this list.  But.... :)
> 
> I would suggest that one must consider the future and try and make a 
> difference.  By continuing to invest in the Microsoft monopoly machine, 
> we continue to reduce our own options.  We all must make the difficult 
> choice to change the future.  To make it better.  To that end, I'd 
> suggest that Linux is not only the best solution, but the only solution 
> when the options are Linux or a Microsoft solution.  Grant it, if you 
> have some hardware that is not supported by Linux, then you're pretty 
> well stuck.  But, we should all build on the future.  Until we create a 
> competitive environment in the world of computers, it won't happen. 
> Microsoft certainly isn't going to do it, and they are in control.
> 
> I never purchase any hardware anymore until I've done my research and 
> verified that it will work with Linux.  That's not possible if you've 
> already got the hardware in place as Laurie has indicated.  Still, I'd 
> suggest that the last resort is a Microsoft solution.  And only then, if 
> proper plans are made to migrate to a more open solution in the future. 
>   Your systems environment is only going to get better and your options 
> grow.  Something we don't have a lot of these days in a world with 
> Microsoft....
> 
> 
> Laurie, take a long hard look at your decision.  Going with a Linux 
> solution may well be more difficult to start with, but in the long run, 
> you're going to be better off.  Going with a Microsoft solution is only 
> going to continue to restrict your options and future.  Move in the 
> direction of growth, not the restricted, costly solutions that are all 
> Microsoft.
> 
> -- 
> Until later: Geoffrey		esoteric at 3times25.net
> 
> The latest, most widespread virus?  Microsoft end user agreement.
> Think about it...
> 
> _______________________________________________
> Ale mailing list
> Ale at ale.org
> http://www.ale.org/mailman/listinfo/ale
_______________________________________________
Ale mailing list
Ale at ale.org
http://www.ale.org/mailman/listinfo/ale






More information about the Ale mailing list