Gpowell,
My point was simply that the early design criteria that were used in the
early to mid-Nineties for cable modem distribution design were just as
shortsighted with respect to today's requirements, as those for any other
venue. Only, the cable infrastructure folks really haven't budged in their
vision of the ultimate physical delivery system in the last mile where it
counts the most. Some rudimentary work was done to manipulate media
mixes more economically, but the fundamental limitations of the original
design still persist. And those are the frequency plan being used, and the
use of coax as a main ingredient to the home, because it is the coax that
sets the limits for all other maximum expectations.
They spent about four years between 93 and 97 examining shared
network designs that encompassed anywhere from 500 to 2,000 or more
homes on a single loop segment, and this is where they commenced
putting in live networks, at those levels of concentration. This, despite the
recent history they had at their disposal in 96 through '98 that told them
that these designs would become grossly inadequate, before long. All it
took was some trend analysis of bandwidth utilization to tell them that.
These large segment sizes, of course, portend higher levels of congestion
and waiting times with ever increasing uptake on the system, than was
previously factored in or even conceptualized as being possible. Previous
notions of how cable modem services would be used revolved around the
MSO's own interactive services [those quaint shopping cart
shop-at-home visions, distance learning, work-at-home telecommuters
(which, ironically, are now actually prohibited) etc.] have yielded to
WWW surfing and other forms of traffic profiling that have resulted in
dwarfing anything that was conceived of by the MSOs, previously. They
are not to blame for that, of course, but they should have done something
more proactively to become better prepared for it. Instead, we are now
seeing the beginnings of forced reductions in speed potentials, as we
discussed upstream.
When utilization and traffic payloads were low, segments of the size 500
to 2000 homes were deemed viable. As the uptake in Internet access
ensued, and the file sizes and formats became larger and more graphical,
nothing was done to compensate for the increase in traffic in the outside
plant part of the design, other than some token re-blocking of homes, and
in some cases reductions in segment sizes to what are still, IMO,
inadequate reductions in node concentration.
True, on the backbone side ATHM has taken measures to increase
capacity and introduce caching in strategic places, but that does very little
to offset the increased burdens between the operator's head end and the
subscriber's cable modem. In the last mile, starting from the head end and
working its way back to the sub, new frequency allocation plans and
media strategies are needed as band aid measures at this time (looking to
avert a year 2001 choke period), and none have been forthcoming, yet,
as far as I can see. Some mention has been made to increase the number
of downstream channels when they are needed, but this strategy can only
be leveraged so far, and for only so many ISPs use.
It then becomes a tradeoff analysis between which? The video content
channel for program services, or the Internet access service for
non-affiliated ISPs? Which one of these gets offered up at the alter?
Only recently has T taken the initiative to begin exploring scaling down the
segment sizing (in order to alleviate potential congestion) from the current
x00 to 2000 homes, to groupings of 75 homes, or thereabouts. And the
economics on this model have not proved yet, as far as I know. But this is
at least a confirmation of what I've stated above, and a step in the right
direction, but I don't think it will be enough, for the following reasons.
It could be another several years before this reduced segment design
eventually gets implemented, and by that time the bottleneck may very
well prove to be somewhere else, like in the very last thousand feet of
coax that runs to each of those 75 homes. Or it could once again come
down to not enough allocated 6 MHz channels to go around for
everyone. I say everyone, because by that time, if not much sooner, you
can rest assured that there will be more than one service provider
traversing that last mile along with the now incumbent provider (ATHM
and/or RR).
It's my belief that the principal MSOs should stop fiddling around and
trying to leverage the HFC design any further, much less six years into the
future, and begin targeting at this time to go after a time frame of 2001 to
2002 to begin running fiber directly to the home, in earnest.
Yes, it might take something on the order of a Manhattan Project to see
this through. But at some point, the MSOs need to look at this situation
for the war that it is, lest they watch the integrated fiber in the loop
(IFITL) architectures being established now by the ILECs sweep them
away, entirely. Forbearance on the part of the FCC and stupid legal tricks
can work for them to a certain extent in the short term, but these tactics
can only last so long. JMHO.
Regards, Frank Coluccio |
