Leigh Porter wrote:
I have to ask WHY!? Surly it would be far nicer to download the Linux
source from somewhere since it already has pretty decent networking code
and use that?
What is the point of putting NT on anything but, well, but somebody's
workstation or a server at a push? We already have tons of memory in our
routers without having to add any more 
The reason is simple. While we sit here bitching about this "minor"
issue and that "minor" issue, like the state of the net, the S&M
(Sales and Marketing, but sometimes I wonder) people are out there
from Mickeysoft promoting the "No one ever got fired for buying M$"
message. NT is a >corporate< product. This, in my view, means that
you don't care if it works as long as you can hire someone else to
fix it and your budget is maintained. Hell, if we started using "free"
products, we wouldn't have license fees to pay and my budget to run
a 100 node NT network would disappear. Budgets = Power in that sad
world.
A new vendor on the market will us "Our product runs embedded NT."
as a plus in a sales pitch. Note the full stop. Their target audience
believe them. A different vendor says "Our product runs on xBSD/Linux
..." and then has to spend the next hour justifying why theis
selection is a valid one.
Oh, yeah. I intend to do something about it. In my own little way.
After all, someone has to combat M$ thinking they can buy the ISP
marketplace.
Regards,
You are wrong a little. The difference is:
- if you use MS and it don't work, your boss blame to BILL GATES.
- if you use FreeBSD (or Linux, through FreeBSD is better for the
networking) and it don't work, your boss blame _guess, who? - YOU_ .
That's a matter.
The reason is simple. While we sit here bitching about this "minor"
issue and that "minor" issue, like the state of the net, the S&M
(Sales and Marketing, but sometimes I wonder) people are out there
from Mickeysoft promoting the "No one ever got fired for buying M$"
message. NT is a >corporate< product. This, in my view, means that
you don't care if it works as long as you can hire someone else to
fix it and your budget is maintained. Hell, if we started using "free"
products, we wouldn't have license fees to pay and my budget to run
a 100 node NT network would disappear. Budgets = Power in that sad
world.
...
No, your boss will blame you in -both- cases, because you chose the final
problem solution. It doesn't matter if the core cause was faulty software,
poor installation, or a compatibility issue: in any case, it was -your-
responsibility to be aware of the problems before putting a live system in
place.
You, however, can either:
a) blame Bill Gates/Linus Torvalds/Theo De Raadt/etc., in which case you
probably won't have a job for long, since passing the buck doesn't keep
you employed very long.
b) fix the problem, either by adding/upgrading software (the only real NT
way), or by tracking down the source of the problem and fixing it
yourself.
The "your boss will be mad if something goes wrong" argument is FUD,
frankly. If a solution you put together goes bad, it won't matter what you
built it on...it's still -your- solution.
And what does any of this have to do with NANOG anyway?
You are wrong a little. The difference is:
- if you use MS and it don't work, your boss blame to BILL GATES.
- if you use FreeBSD (or Linux, through FreeBSD is better for the
networking) and it don't work, your boss blame _guess, who? - YOU_ .
Nope, buying IBM/M$ to secure your job is a canard. In both cases, if you
are an employee you will lose credibility (lose enough of that and your job
follows), as an owner of a business you lose customers (lose enough of
*that* and you'll have to get a real job). If an employee causes sufficient
loss of business through bad decisions, they get canned (back to
credibility again).
We run Caldera here, with WinNT workstations. This is not simply because we
are a Caldera VAR. Linux servers are much more reliable than WinNT,
although we have *one* of those also, no choice and it's constantly in
Intensive Care (Netscape ES3.51 locks up regularly or NT auto-boots and we
have to MANUALLY enter the VeriSign password. We are working on an
Linux/Apache-SSL/PostgreSQL/self-certifying solution). BTW, Caldera does
not yet have an Enterprise Server solution, but there's one on the way.
Come on guys.... not on nanog.
Can't you smell a clever market research plot from Microsoft when
you see one? Whoever started this is just testing the waters for
embedded NT...
Dirk
PS: Yes, Linux will kill NT as server platform. No doubt about it.
This list has a different purpose.
Yup. The question was originally posted by by someone from Wired to get
info on an article for someone else from Wired.
Most likely they are doing a review of a) berkeley networks product or b) a
review of the new book "Gigabit Ethernet Handbook" by Stephen Saunders.
This book devotes a whole chapter (only about 20 pages) to Berkeley
Networks "NT Switches", describing them as a "new class of switched
internetworking platform ... to transform today's networks from passive
data conduits to application-aware communication and control
infrastructures" (pg 180). BTW, "application-aware" means the ability to
block Doom traffic. (their example).
Although the book is allegedly technical, this chapter is nothing but
marketing BS. (And full of warmed over ideas, recast as "brilliant
inovations", which is typical of MS partners, to boot) This chapter was
written by 4 people from Berkeley Networks, including the VP of Marketing.
This chapter clearly belongs in a marketing brochure, not a book on gigabit
ethernet. So, I'll pan both the book, and judging from the chapter on
Berkeley Networks "NT switch", their product as well.
I really don't think that (note the quotes) "NT-Switches represent a
fundamental shift in the business model of networking vendors." (pg 201,
Conclusion: The Future of NT-Switches)
The chapter makes it sound like everyone has one, and thats all they use.
Unless the world changed a lot since Friday, I don't think this is the
case. In fact, does *anyone* actually have any of these?
--Dean
This strategy worked fairly well for IBM selling mainframes. Its the way to
sell to morons. They aren't going to understand the technology well
enough to make an informed, intelligent choice, so you convince them they
will be "safe" buying your product. Of course, IBM was selling to business
managers who didn't know anything about computers. Does Microsoft think
they will be selling this kind of networking gear to people who don't know
anything about networks? (They've made such mistakes before. MS Bob comes
to mind, here)
Somehow, I think networking is and will remain hard enough that there won't
be very many such idiots in the networking marketplace. I concede that
this might be an unreasonable expectation however.
--Dean
Somehow, I think networking is and will remain hard enough that there won't
be very many such idiots in the networking marketplace. I concede that
this might be an unreasonable expectation however.
This is already demostrobly wrong. Look at the Fortune 500 - or any group
of "large corporations" - and see who makes these type of decitions. I
wouldn't hesitate to give you 10 to 1 odds that well over half of these
individuals couldn't tell ethernet from token ring.
Now, the good ones have "smart" people helping them, making recomendations
or some such, but they still make the final decision. And a lot of times
the marketing/sales guy can get in and make their mind up for them before a
recomendation is made. Or the sales guy plays golf with the president or
something and does the deal that way - with absolutely zero technical input.
Marketing is designed to let the clueless user think he knows what he's
talking about, that he's covered all the bases, and that he understands the
technology. All while touting the particular product's strengths and
completely ignoring its weaknesses. This makes it hard for a technical
person to explain why the manager is making a bad decision - he thinks he
knows enough to make it himself 'cause the sales guy told him all he needs
to know.
Too many times have I tried to explain why something was bad to someone in
"authority", only to have them pick the product against which I was
recommending because they were a bigger company or had more history or some
other factor which has nothing to do with technical competence. Oh, and
price comes to mind too.
BTW, this does not mean there are not any good managers out there. Just
that there are more than enough bad ones to make a company very successful
off their incompetence.
--Dean
TTFN,
patrick
No, its not a matter.
Either :
1- Educate your boss.
2- Set things up such that the "other-ware" or "free-ware" _will_ work.
Our boss has us jumping into NT, and still blames _us_ when it goes
t$#s up. I dont even give him a chance with the unix boxes anymore.
Morgan
A unnamed company runs a ccMail infrastructure with ~3,000 mail accounts and
at least ~2,000 of those on a Novell 4.11 machine on a Pentium 166. That box
can handle ~1,700 concurrent users with no problem. You can figure the cost,
we could put that together with spare parts nearly!
It WILL be replaced with MS Exchange. Cost to create the same infrastructure?
About 1.2million and ~8 boxes including a DEC Alpha ~300-500mhz.
A Linux/FreeBSD with IMAP,SMTP,LDAP etc. or even Netscape solution would be
much cheaper too
Its not even a question. Its been a corporate directive. Go figure. I gave up
on common sense in corporate America a long time ago!