May I ad this little edict that "Buggy/Unreliable OS's" is just a bad
wrap that the community has labeled to MS "WINDOWS"... without much
explanation of WHAT was the "unreliability".

Indeed, Security is a different story, and there is plenty of "reliable"
argument to that topic... but to constantly perpetuate this argument
that "Windows" is unreliable and buggy makes me ill.

What makes the story of it being buggy/unreliable has always been
related to device drivers that sloppy-a$$ programmers whip out for
sub-standard chipsets on the Intel platform running "Windows".  Not to
mention, the OS's that have been the most unreliable/buggy have been the
desktop OS's- NOT the Server platform Windows NT.  If you think that you
should use that Windows 98 box as your company's Server- it's your own
stupid fault for all the headaches that are derived from therein.

I've worked predominantly in the NT environment for over 8 years, going
through the NT 3.51, 4.0, and now Windows 2000 version of the server
platform, and I ONLY have had servers crash when a vendor-specific
device driver was updated (ahem, Intel ironically was the culprit, and
they were supposed to be the other half of the "Win-tel" agreement).
I've also maintained a fair share of different-flavored *nix boxes that
performed similar functions, for which they suffered the same ailments-
bad drivers for add-on hardware, whether it be NIC's, RAID Controllers,
Telephony boards, or power failure.  One thing for sure, the NT box
didn't spend 30 minutes spewing INODE errors all over the place once
power was restored... unlike the AT&T Unix brothers did... And yes, I
know, NT uses a journaling file system as opposed to the file system
Unix uses.  But for heaven's sake! The DB application on the *nix box
should have the corruption issues to worry about, NOT the OS!

Most of these Windows NT Servers under my command were Computer
Telephony systems, a.k.a., IVR's.  They ran like a champ for several
years without a reboot... the ones that ran for shorter periods were
maintenance reboots for Service Packs or because of Power Failure to the
location the box was residing.  These servers were both DEC Alpha's and
Intel-based OEM and Clone machines.

As I said before, just as much as it is a problem for the *nix platform,
the "things" that make the OS unreliable is the cheap hardware and
sloppy device drivers that are applied to the system.  Proper
installation, and hardening of the OS for the specific purpose it is
supporting (read don't use the same machine you've set up as your server
as your desktop too, installing all kinds of non-server related programs
on it like "free-ware" and demos of programs found in the center or back
of some periodical you got in the mail), and the Windows NT / 2000
Server will be just as stable as the next implementation of Solaris on a
Sparc station.

And again, as Chuck pointed out, if the Applications developed to run on
the Windows NT / 2000 platform were developed properly, than the servers
would be reliable in that respect too.  I'm not a programmer by any
means, but from what I've observed, you can have just as many crashes
for building crappy DLL's as you can from improper handling/use of C
library modules on a *nix box.  Not to mention, both types of
programmers need to know how to program for Memory Address handling.


But who am I to argue... the whole slamming of "Windblows" is probably
just because some bull-headed *nix lackey is just pi$$ed off he can't go
rebuild the kernel half a dozen times to "tweak" the system on
"Windows".

And as a final note, I do maintain the argument that ALL of the OS's out
there have their own place in the industry; there isn't just ONE O.S.
that addresses all the use/needs of any particular business (keeping
Support in mind).

Now- Back to our regularly scheduled commentary on Cisco Studies.

-Mark

-----Original Message-----
From: nrf [mailto:[EMAIL PROTECTED]] 
Sent: Thursday, July 25, 2002 12:07 AM
To: [EMAIL PROTECTED]
Subject: Re: CCIE Security Lab [7:17848]

Buggy/unreliable software is indeed the same anywhere.  But when
combined
with buggy/unreliable OS's, now we're talking about a solution that is
REALLY buggy and unreliable.  For example, if your software is only
guaranteed to run at 3 9's, and your OS is also only guaranteed to run
at 3
9's, then overall we're talking about a less-than-3-9's of a solution.

You can actually run packetized voice very reliably, and not just for
toll
bypass (although it is definitely true that toll-bypass  is the easiest
and
most mature kind of packetized voice to do).  The key is that you have
to
design things in  a certain way to maximize your reliability.  Many
carriers
like SBC use packetized voice with soft-switch signalling in certain
parts
of their network, and then you have packetized voice wholesalers like
Ibasis
that have massive available voice capacity and a good reputation for
reliability.  There was a huge amount of serious talk after 9-11 for
Verizon
and other carriers to contract for backup voice capacity through
somebody
like Ibasis in case their voice switches got destroyed again - as during
9-11, people saw that while traditional voice service was severely
affected,
packet networks like the Internet were still functioning, so in these
kinds
of circumstances, you could say that packetized voice might actually be
more
reliable than regular voice.    But again, it takes very careful design
to
achieve this kind of reliability.


""Chuck""  wrote in message
[EMAIL PROTECTED]">news:[EMAIL PROTECTED]...
> sure. ok. agreed.
>
> OTOH, buggy / unreliable software is the same, no matter whose
platform it
> runs on. A long time ago in a galaxy far away I was able to
successfully
> crash Sun Unix boxes several times through sheer ignorance. one was in
the
> Sun Sys Admin training class I was taking, the rest were Sun boxes
that
Big
> Brokerage Firm had installed in the office where I worked. Proof that
there
> ain't no such thing as "foolproof" because this here fool can break
just
> about anything ;->
>
> BTW, you have just ht on the major reason for NOT doing packetized
voice.
Or
> maybe just limiting it to toll bypass, while keeping your PBX.
Sometimes I
> think the only real selling point for AVVID is that is "kewl"  The
biggest
> selling points for Windows way back when were the screen savers and
the
> games. MCSE = Microsoft Certified Solitaire Expert
>
>
> ""nrf""  wrote in message
> [EMAIL PROTECTED]">news:[EMAIL PROTECTED]...
> > Well, this kind of thing cuts both ways.  A reseller I know is
trying to
> > sell AVVID and is on dangerous ground precisely because CM is on
Windows
> and
> > the potential customer has had some very bad experiences with
Windows
> > servers due to reliability issues and so forth.  The customer is
deciding
> > whether to go with AVVID or a traditional PBX, and the fact that
AVVID
is
> so
> > Windows-centric is a significant minus, and in fact could be the
whole
> basis
> > for losing the deal, because the customer has to know that his phone
> system
> > is going to always be up without any dithering around.  Yes, yes,
you
can
> do
> > things like clustering to improve the reliability of CM, but the
simple
> fact
> > of the matter is that Windows has a well-founded reputation for
> > unreliability when compared to UNIX, and when you're talking about
phone
> > systems, unreliability is definitely something that a potential
customer
> > does not want to hear.  Not at all.   This is why you rarely see any
> vendors
> > of enterprise software (like DB's, ERP, CRM, SCM, etc. etc.) that
don't
> > offer a UNIX version - because just like a phone system, these are
crucial
> > applications that just have to reliable.
> >
> > ""Chuck""  wrote in message
> > [EMAIL PROTECTED]">news:[EMAIL PROTECTED]...
> > > for whatever reason, Cisco and Microsoft are partnered for a lot
of
> > things.
> > >
> > > Call Manager for *nix??? hahahahahahahahahaha
> > >
> > > IIRC the last Cisco management software presentation, just about
> > everything
> > > is on NT or Win2K boxes these days.
> > >
> > > I believe it's called Market Share - there are far more Microsoft
> > certified
> > > folk than *nix certified folk. Try selling AVVID when you also
have to
> > tell
> > > a customer that he has to hire a *nix capable individual or
retrain
his
> > > existing Microsoft capable people. Same for the management
platform.
> same
> > > for any of the security related software products.
> > >
> > >
> > >
> > > ""sergei""  wrote in message
> > > [EMAIL PROTECTED]">news:[EMAIL PROTECTED]...
> > > > Makes sense, - M$ products need it more...
> > > >
> > > >
> > > > DAve Diaz wrote:
> > > >
> > > > > Remember cisco have no money just $21 Billion dollars in the
bank,
> so
> > > no
> > > > > new hardware for a while, no unix in a security lab that is
absurb,
> > > > >
> > > > > Dave
> > > > >
> > > > > >From: "markh"
> > > > > >Reply-To: "markh"
> > > > > >To: [EMAIL PROTECTED]
> > > > > >Subject: Re: CCIE Security Lab [7:17848]
> > > > > >Date: Tue, 5 Mar 2002 00:33:52 -0500
> > > > > >
> > > > > >really?
> > > > > >--
> > > > > >
> > > > > > >I have an official statement from Cisco that says that
there
will
> > be
> > > no
> > > > > > >UNIX, only NT.
> > > > > > >I was there and it's true.
> > > > > > >
> > > > > > >MS
> > > > >
_________________________________________________________________
> > > > > Get your FREE download of MSN Explorer at
> > > http://explorer.msn.com/intl.asp.




Message Posted at:
http://www.groupstudy.com/form/read.php?f=7&i=49627&t=17848
--------------------------------------------------
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]

Reply via email to