Conference 7.286::fddi

659.0. "CDDI. ANSI X3T9.5 and the context of standardized mediums" by RCOTLW::WRIGHT (You can't argue with a sick mind!) Wed Jul 29 1992 19:45

Maybe I'm missing something here, but I'm puzzled and confused over the limited
context of cable-type transport mediums ANSI X3T39.5 is addressing.  My under-
standing is that our RG-58 implementation is just that...ours....not covered
by ANSI UTP (MLT-3 or NRZI) or STP standards.  

I've been told the thinking behind ANSI's limited cable type context is "market
demand" potential.

Folks, it's going to be, in short order, a "multimedia-crazed" market, and there
is over 1,000,000 (that's right "over a million") miles of RG-6 and  RG-11 type
cable serving the biggest market there is....the consumer.  Some folks who should
know believe we can achieve greater distances for CDDI over RG-6/RG-11 than with
our RG-58 (Thinwire).  

We have a CATV-IBU now.  Most of the major MSO's are upgrading their cable plants
to very fiber-rich hybrids of a combination of what I'll call "star-branching"...
which is a redundant star-topographied fiber ring serving branching tree bi-
directional service areas where RG-6/11 goes the final distance.  We believe (and
seem to be betting our livelihood) that the world wants standards-based solu-
tions.  (I agree with this to a large extent.)  Our major competitors are eye-
balling the cable space with lusty thoughts of exploiting this huge bandwidth, 
soon-to-be-bidirectional infrastructure (2-3 years), with multimedia dominance in
their minds.  IBM is pushing their Fiberchannel architecture, supported (and 
likely looking quite attractive to debt-heavy cable MSOs) by overtures of a 1/2 
billion $ investment to make it all happen.

Perhaps now is the time to secure a little insurance against ibm's strategy by
introducing ANSI XT39.5 to the reality of the installed cable base, with some-
thing other than Fiberchannel as the preferred technology. I don't see any sense 
in letting big blue establish the de facto when we could complicate their game 
considerably with the strategically-defensible standards posture.  Any comments
from any of you thinkers out there?

Terry

I think one problem with CATV is that it isn't just cables, but also active
components (amplifiers, etc.).  It's not a baseband system.

Consider Ethernet: support for RG-58 was a trivial mod to the original thick
orange hose stuff.  Support for CATV installations was very much non-trivial
and (as far as I know) not much of a market success even though it apparently
did work.

It's also not clear to me how you'd establish a ring network on top of a
branching tree cableplant...

	paul

Paul:

The "ring" on branching trees will be accomplished via the fiber-rich physical
cable plant upgrades cable operators are now doing.  Most are running fiber to
sub head-ends and on to feeder nodes associated with a particular service area
of coverage.  Feeder nodes will be looped to sub head ends, which in turn will 
be looped to central headend.  A FDDI/CDDI std for the physical rg6-rg11 cable 
types would introduce possibility of "last mile" FDDI through looping of final 
copper segment back to service area feeder node via reverse channel capability.

Terry
 

Thanks, that sounds a bit more plausible.

The obvious question is: apart from cable type, would our Thinwire FDDI adapter
and concentrator work?  If so -- in other words, if the topology fits --
then it would be a matter of tweaking the design for the different cable,
which sounds like a minor change.

We haven't pushed the standardization of the coax PMD; I believe the reason
is simply that we didn't think there would be much interest.  If a CATV variant
would be sufficiently interesting, perhaps we should propose such a thing.
Or alternatively, perhaps we should work with other interested parties and
let them do some of the legwork.

	paul