Re: BitTorrent swarms have a deadly bite on broadband nets

On Tue, October 23, 2007 4:06 pm, Jack Bates wrote:

> Errr, 8 pairs per customer? Even 4 is a step backwards.
If we're going to
> do construction at that level, might as well drop in
fiber. We're still
> enjoying the fact that ADSL runs on 1/2 a pair while
the customer's
> phone service is out.

Doing (or getting the incumbent to do, where the last mile
is a monopoly)
a little bit more of what you already do seems to be an
awful lot easier
than doing something completely different.  Certainly in the
(admittedly
all European) countries where I've seen it done, getting 4
or 8 copper
pairs from a customer site to the exchange is an order of
magnitude or
more difference in both cost and lead time to doing anything
at all with
fibre.

Every house I've lived in has had 4 pairs already going from
the house to
the first street cabinet, with just the first pair connected
for voice,
and getting a second line has always just needed a patch
onto a spare pair
from the cab to the exchange.  Obviously if *everyone* wants
4 or 8 pairs
to their house, there's going to need to be a lot more
copper between the
exchange and the street cabs.  It's not clear that
*everyone* wants
upstream though, and 2M to 5M on a single pair (depending on
distance /
quality) is quite possible if you wanted to think in terms
of ubiquitous
symmetric service.

I take it that getting spare / new copper in the US is more
painful?

Regards,
Tim.

Tim Franklin wrote:
> Doing (or getting the incumbent to do, where the last
mile is a monopoly)
> a little bit more of what you already do seems to be an
awful lot easier
> than doing something completely different.  Certainly
in the (admittedly
> all European) countries where I've seen it done,
getting 4 or 8 copper
> pairs from a customer site to the exchange is an order
of magnitude or
> more difference in both cost and lead time to doing
anything at all with
> fibre.
> 
Sorry, I am the incumbent. ;) I was just thinking of the
copper necessary to do 
such a task on a massive scale. It's definitely not in the
ground or on a pole 
at this point in time. One reason DSL was so desireable for
many small ILECs was 
the recovery of copper from dual phone lines caused by
dialup.

> Every house I've lived in has had 4 pairs already going
from the house to
> the first street cabinet, with just the first pair
connected for voice,
> and getting a second line has always just needed a
patch onto a spare pair
> from the cab to the exchange.  Obviously if *everyone*
wants 4 or 8 pairs
> to their house, there's going to need to be a lot more
copper between the
> exchange and the street cabs.  It's not clear that
*everyone* wants
> upstream though, and 2M to 5M on a single pair
(depending on distance /
> quality) is quite possible if you wanted to think in
terms of ubiquitous
> symmetric service.

In newer homes, most of the ILECs I work with use 6 pair
drops. This is 
relatively new, though. There are times that it was just 2
pair. It's amazing 
how things change back and forth over 50+ years.

> 
> I take it that getting spare / new copper in the US is
more painful?

Depends on locale and quantities. I know of several rural
ILECs which are 
currently undergoing a 3 year process to recover copper.
This involves replacing 
bad peds and boots, cutting off the bad copper in them, and
either pulling up 
some slack or splicing in some fresh copper (splicing 600
pair in a bucket truck 
is rough on the legs ;). We're shocked DSL has worked at all
in some of this plant.

The cost to recover and repair what we have is far less than
throwing anything 
else into the ground, but no one considered needing as much
copper as it would 
take to bump everyone from DSL to a 4 pair system. I won't
even discuss RBOC 
mentality when it comes to rural plant (including the entire
state of Oklahoma).

Jack Bates