Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!ucsd!ucbvax!ucsfcgl!cca.ucsf.edu!wet!cmilono
From: cmilono@wet.UUCP (Carlo Milono)
Newsgroups: comp.dcom.lans
Subject: Re: shielded twisted pair
Message-ID: <1600@wet.UUCP>
Date: 4 Oct 90 15:33:41 GMT
References: <4200008@ux1.cso.uiuc.edu>
Reply-To: cmilono@wet.UUCP (Carlo Milono)
Organization: Wetware Diversions, San Francisco
Lines: 45

In article <4200008@ux1.cso.uiuc.edu> andys@ux1.cso.uiuc.edu writes:
>
>I would like to get some feedback on what people think
>about installing SHIELDED twisted pair.  

Awful!  If you ground *both* ends, prepare for meltdown on some runs...I am
serious; the differential ground potential (ground isn't ground) can run
into sufficient amperage to melt copper.

>as a computer research facility.  High speed networks will
>be required.  Each room will have UTP, coax or a  some 
>combination of the two.

UTP would be a must - coax is only necessary if you want to support
broadband with the attendant, relatively complex electronics.  What
about installing fiber?  AT&T now has a single-sheath wire that has
a pair of 62.5 micron fibers and two four-pair copper.  You can leave
the fiber unterminated and implement as you see fit...That will provide
the "F" of FDDI.

>
>But, I'm getting the feeling that FDDI will be supported on
>shielded twisted pair.  800+ runs of anything will be 
>expensive and I am willing to bite the bullet, but which
>bullet?
>
FDDI - if by that you mean 100Mbps transmission with token technology,
then I would say that you are correct, but it *is* doable under UTP.
The problems with 'porting' FDDI to copper media is FCC regulations, with
either STP or UTP; the harmonics are such that you would need to triple
or quadruple the shield in every exposed (not in conduit) piece of cable.

There are committees working on the *correct* way to achieve interoperability
between FDDI and what has been called CDDI, but the encoding scheme will
change, so that a simple media converter will not suffice, but rather a
coding-converter will be required as well.
>
>Andy Schmidt
>UofI


-- 

                           Carlo Milono                                 
                   UUNET: cmilono%wet.uucp  (via claris/well)