Re: [tlug] FTTH ISP recommendation

[Curt Sampson <cjs@example.com>]

On 2008-08-07 16:26 +0900 (Thu), Edward Middleton wrote:

>> As for the "SNR" stuff, well, err...I'm also not sure where even to
>> start. You are aware that a piece of metal doesn't have a "bandwidth in
>> Hertz", right? And that only communications channels, not wires, have a
>> signal-to-noise ratio?
>
> Lets make it simple. A phone line is designed to support 3.1 kHz
> bandwidth with an acceptable SNR.

In a word, no.

If you're going to talk about what things were "designed for," try this on
for size.

Entire local loops, from the jack through the customer premises wiring
through the building's distribution frame and demarcation point
through copper all the way back to the CO, which might well be several
kilometers or even tens of kilometers away, were designed to have
electrical characteristics compatable with doing analogue modulation
of voice signals within the bandwith you specify, possibly with the
addition of coils for longer runs.

That was up until about the sixties. T1 modulation was introduced in
1961 and worked just fine on many local loops already (if any existing
coils were removed), giving twenty-four channels each capable of
carrying something approaching 4 KHz bandwith with an S/N ratio close
to 50 dB. And since that time, most local loops were designed to handle
this as well, not that the design needed a lot of changing.

Since then, better modulation schemes have come along, with some very
nice advances in the 1990s and 2000s. Combine that with the fact that
the VDSL we're talking about is running over a few dozen meters of cable
rather than a few thousand, and you'd expect several orders of magnitude
more bandwith.

And keep in mind that the minimum spec and how the plant (wires,
connectors, etc.) actually comes out are quite different things; in
essence, with modern technology, telephone lines are far overspecified
for a single voice channel already, and are generally far overbuilt as
well.

On 2008-08-07 10:12 +0200 (Thu), David Santinoli wrote:

> You are aware that it is perfectly legal to define the bandwidth (in
> Hertz) of a (couple of) wire(s), right?

Sure. For a given modulation scheme. We've come up with a lot of new
ones in the past twenty years, and even fifty years ago we had a fair
selection that gave dramatically different bandwidths.

On 2008-08-07 16:26 +0900 (Thu), Edward Middleton wrote:

>> He's not objecting to how the utility pole is used; he's objecting to
>> new things attached to his building.
>
> Yes, well he probably objects to the hours and company you keep too, but 
> clearly not enough to reject you as a tenant.  The point is not that you 
> would do it without the landlord consent but *if* you are entitled to do 
> it, and you let your landlord know this fact the ball is in his court to 
> resolve your internet issues.

Which he did, as far as he's concerned, by allowing mansion fibre to be
installed.

It seems to me that a landlord is probably morally, as well as legally,
within his rights to dictate what sort of prominent things you can
attach to the outside of his building for all the world to see.

cjs
-- 
Curt Sampson       <cjs@example.com>        +81 90 7737 2974   
Mobile sites and software consulting: http://www.starling-software.com