On Wed Mar 05, 2003 at 05:12:58PM -0500, Jim Hubbard wrote:

(You need to fix your reply-to)

> You know, I understand that developing software is hard.  I know I
> can't do it.  And not to make light of developer's efforts (many of
> them volunteers!), but to someone who's choosing a server OS, it
> doesn't matter even a little bit how hard the developer's job is.  I
> bet none of them could do my job either because it's hard too.

Oh, I'm not sitting here hoping for sympathy.  =)  I was attempting to put a
little perspective to the thread.  You're absolutely right.. the end user
shouldn't care how long it takes someone to do a job provided it gets done.
That isn't their concern.

However, people are asking for reasons and motivations for doing what we
did.  That's where the perspective comes in.

> Put yourself in everyman's shoes and try answering this little
> questionare as honestly as you can.  Make a little chart with these
> OS's at the top: MS Windows Server, Mandrake Corporate Server, Redhat
> Advanced Server, and Debian.  Along the side, put these categories:
> Product lifetime, Hardware compatibility, 3rd party application
> compatibility, and finally, Total projected cost (including training
> if necessary) over 5 years.  Now tell me, which OS looks better on
> paper?  THAT, my friends is what objective system administrators see.
> Linux is more stable?  I think so too, but how much more stable?
> Linux is more secure?  I agree, but how much more secure?  If you
> can't put a number on it, then it's not part of the equation because
> it is an opinion.  And it doesn't matter if millions of the world's
> brightest minds have been through hell and back to develop this
> software either.  All that matters is the bottom line.

Well, I can answer this honestly although I don't have time to make
comparisons.  There is absolutely no way in hell I would use Windows as a
server.  My experience with Windows is horrific.  Yes, I've had crashes and
burps and really bad problems with Linux, but these are few and far between.
When I can have a Linux box running for 6mos without a burp, and reboot due
to a kernel upgrade, I'm happy.  These are webservers, BTW, that run
proftpd, qmail, vmailmgr, djbdns, apache, MySQL, postgreSQL, and other stuff
(but those are the biggies).  Let's contrast this to my Win2k box that
*only* does EverQuest; and is only used maybe 2-3 times per week.  Just last
week I had a "burp" on that box that killed the partition table on my Win2k
drive.  Needless to say, for a box that's sole purpose is to play EverQuest
to do something like that is absolutely ridiculous.  Every 6mos, when I
reboot my Linux boxes, I'm *reinstalling* Win2k.  This is supposed to be
MicroSoft's server offering?  Good grief.  It's next to useless as far as
I'm concerned, but I enjoy playing EverQuest as time allows.  That's the
only reason it's there.

So to the stable part, I'd take Linux, of any flavour, over Windows.  Any
day of the week.  For security?  I can't really be objective because I've
never tried to secure a Windows box.  I have no clue how difficult it is.  A
Linux box can take some time to fine-tune and secure, yes, but the end
result is a rock.  Can anyone make the same claim for Windows?  I can't;
I've never tried.

As far as TCO goes, I can again only speak for myself.  But you have to
understand I have a unique background.  I went from DOS to OS/2, then Linux.
I skipped Windows entirely.  The only reason I tried Windows in the first
place was for games.  For anything serious, I've always preferred OS/2, and
more recently Linux.  So, to me, the cost of training would be more
expensive for a Windows box than a Linux or *BSD equivalent.  Heck, I'd not
used a Mac in years until last year and I felt right at home in OS X.

Can I say the same about everyone?  No.  Do I know everyone's situation?
No.  Will the TCO of linux be too much for some?  Sure, if they're
short-sighted.  Microsoft is hell bent for leather to suck their customers
dry.  The future will prove this; their plans for things like Paladium prove
it now.  That's a different issue, but should also be on an astute's
person's mind.

Would I pay $1500USD for Corporate Server?  I wouldn't.  But I don't need
the support.  Would I pay $700USD for it?  I probably would.  I've spent
more on other software I no longer use.  Three years of use out of a product
on critical servers?  Sure.

You also fail to factor in licensing.  There is no "per-seat" license for
Corporate Server.  I can purchase it and install it one computer, or a
hundred.  So here's another twist to your equation.  How big is the install
base?  Let's take two groups; one with two servers, one with a hundred.  For
the sake of argument, we'll make them redundant with the same software on
each.

CS2.1 @ $1500USD x 2 = $1500USD
CS2.1 @ $1500USD x 100 = $1500USD
Win2k @ $1000USD x 2 = $2000USD
Win2k @ $1000USD x 100 = $10000USD

(I'm guessing on the cost of Win2k... I couldn't be bothered to find an
exact price so I'm making a generalization that I hope proves a point).

Already with two machines, the base cost of CS2.1 is cheaper than win2k.  At
100 machines, the base cost of CS2.1 is a small percentage of that for
Win2k.  Throw in 100 MS SQL licenses, 100 IIS licenses, and so on and watch
that win2k TCO explode.  Per-user licenses?  Bring 'em to the party!  These
are per-seat licenses, remember... doesn't MS like per-users licenses as
well for concurrent access to a service?  Add 'em up.

I suppose my point is you can argue it all you like, the TCO for Win2k comes
out on top.  Again, if you look at a base OS, with no additional services on
the Win2k side (they all come with Linux), with your 100 servers you've got
$98000USD to spend on training.  You don't think you can educate someone
well enough with that kind of money?  Hell, a fellow with half a brain could
buy a whole bunch of books for a fraction of that cost and educate themself.

> Linux distros are being compared with Microsoft, and until vendors are
> prepared to play the game on their level, Linux won't go much farther
> than it already has.  Personally, I believe in Linux and the
> open-source movement, and I'll go out of my way to use some version of
> Linux for years to come.  But to get more administrators to use Linux,
> vendors are gonna have to inspire a heck of a lot more confidence than
> they do now.

Well, as per your argument, aren't we playing their game already?  I mean,
some think CS2.1 is too expensive (isn't this part of M$' game?).  Some
think upgrade sales are too aggressive (EOL, etc.), but isn't this, again,
part of M$' game?  My dad is the IT manager for a non-profit organization
that's quite large and uses a lot of MS software.  Ever since XP has come
out, MS has been attempting to strong-arm them into upgrading every machine
to XP, Office to the latest version, etc.

Anyways, regarding this "inspired confidence" comment...  MS inspires
confidence?  How laughable is that?

I'd write more, but I've spent too much time on this already (how *do* I get
suckered into these threads?!?).  I likely have little more to contribute to
this thread, especially now that we're comparing apples and oranges.

-- 
MandrakeSoft Security; http://www.mandrakesecure.net/
Online Security Resource Book; http://linsec.ca/
"lynx -source http://linsec.ca/vdanen.asc | gpg --import"
{FE6F2AFD : 88D8 0D23 8D4B 3407 5BD7  66F9 2043 D0E5 FE6F 2AFD}

Attachment: pgp00000.pgp
Description: PGP signature

Reply via email to