shaver@atanasoff.cs.iastate.edu (Dave Shaver) (10/02/89)
Everyone's tired of this subject, so I'll try and keep it short:
We (finally) have a need for some twisted-pair ETHERNET (TPE)
products. Over the last year or so I've gathered the following general
opinions on TPE products, some from the net, others elsewhere: (PLEASE
contest if this is wrong!)
1) The two-wire solutions were ``good'' if you have the wire
in place and don't have the time and/or money to replace it
with additional wire. They don't provide the amenities of
the four-wire solutions, but are good for *lightly loaded*
quick-and-dirty short hops.
2) The four-wire (ie: Cabletron, TCL, et al) solutions are to
be preferred in a big way over the two wire solutions,
mostly because they provide the necessary ``brains'' to
handle the the ``poor environment'' that unshielded
twisted-pair provides. Further, they can collect statistics
that can help if (when) a problem occurs. (That is, unlike
the two-wire solution, the are non-passive => expensive.)
3) For heavy loads (ie: lots of NFS and/or diskless stations)
the four-wire solution should be used. The ``passiveness''
of the two-wire does not lend itself well to peak loads.
4) Since the 10baseT standard will use the four-wire solution,
the vendors using four-wire solutions today should have a
``jump'' on the two-wire vendors. (The Cabletron sales
force goes further to say that they will have a 10baseT card
for the MMAC that will talk to the any modules which we buy
today.)
The problem is this: I've specified that we need about $10.5K worth of
Cabletron gear and the bureaucrats are saying ``prove the need'' for
the four-wire solutions v. the two-wire solutions. (They only see that
the two-wire solution will cost 1/3 as much as the four-wire solution.)
To further complicate things, several sites on campus are already using
the two-wire solution in a *lightly-loaded* manner, so the claim from the
people with the money is that it will work for our needs, too. Current
use is for extremely short hops with very minimal traffic.
Thus, YA request to the net is born. 8-)
What I need is hard ``evidence'' (ie: test reports, white papers, trade
journal write-ups, almost any sort of ``documented proof'' other than
sales brochures) to back up my claim that we need four-wire products.
Do we really need the four-wire solution? I think so, but I don't have
any experience with TPE. If you think we don't, call me an idiot, but be
able to back it up with some of the ``proof'' the bureaucrats are looking
for!
Also, I would like case histories from people using either two- or
four-wire TPE products. I don't care which you say is better, just
express your opinion.
If anyone has saved past comp.dcom.lans discussion about TPE, feel free
to send that to me, too. (I'm sure I've missed many postings.)
Finally, actual product recommendations would be wonderful.
Of course I'll summarize and probably put all the responses up for FTP
if that would be helpful to others.
/\ Dave Shaver -=*=- CS Systems Support Group, Iowa State University
\\ UUCP: {hplabs!hp-lsd, uunet!umix!sharkey}!atanasoff!shaver
\/ Internet: shaver@atanasoff.cs.iastate.edu
...In stereo where available...