[ntp:questions] Re: tinker step 0 (always slew) and kernel time discipline

[Joseph Gwinn]

In article <IJSdnV6hOrl5cYjYnZ2dnUVZ_uWdnZ2d at megapath.net>,
 hmurray at suespammers.org (Hal Murray) wrote:

> >> The telephone companies tend to be very aware of time and timing.  The 
> >> time division multiplexing of T1 and T3 lines requires splitting the 
> >> second very precisely.  Cellular phones also require very precise 
> >
> >That can be true of the public network, although even then it may
> >be more so for the more engineering oriented layers, like bearer
> >synchronisation, than the more commercial oriented layers, like
> >call detail recording.
> >
> >However, it is very definitely not true of most PABX systems, which
> >typically have wristwatch and eyeball set times and run in local
> >time, with no automatic daylight saving switch, and have no high quality
> >frequency standard.
> 
> There are two separate issues: time and frequency.
> 
> Anybody know if I can get a good frequency off a DSL line?
> Assume I'm willing to hack a wire into my modem/router.
> 
> If so, it might be a nice/cheap way to get a stable clock
> to use with a NTP box.  (Handwave, PLLs and such.  Not a hard
> problem.)

I assume that DSL follows a public standard, and that this standard will 
specify how good the clocks must be.

For comparison, broadband cable modems follow DOCSIS 2.0, available from 
<http://www.cablemodem.com/specifications/specifications20.html>.

My understanding is that both are timed protocols of some kind, but I 
don't know the details.

Joe Gwinn