A couple of thoughts here:

1) I know at some sites there is an external, shared, GPS antenna which is
run through a distribution amplifier to clients.   Worth checking into just
in case it exists and they forgot to offer it to you.

2) Do you have any specs on what you need for the TDM clock?   If you don't
have GPS or any other reasonable way to discipline your local clock you
could conceivably get an accurate freerunning clock and use that.  However,
if this is even possible is going to be based on the accuracy/precision
needed for the clock.   The spec should be something like 10E-11 or
something like that, possibly with jitter or other specs specified as
well.   The more accurate, the fewer options you will have and the more it
will cost.    If you only need 10E-6, you can do this dirt cheap.  If you
need 10E-13, you're going to need a Cesium clock which will set you back a
good five figures (and then some).

3) Do you have spare, dark fiber or perhaps even a WDM color to somewhere
you can get GPS?  Copper might even work depending on the needs of the
switch. The thought here is if you have a stable, non-packetized link to
somewhere with GPS you then have quite a few options for transferring time
back to the site.

I agree with you that NTP time transfer isn't probably accurate enough by
itself to discipline a clock for TDM...  of course depending on the exact
needs.


On Thu, Jul 11, 2019 at 9:25 AM Mike Hammett <na...@ics-il.net> wrote:

> I'll look into Meinberg.
>
> I recent thread mentioned high-sensitivity receivers often allow GPS to
> work inside. Obviously "inside" has a lot of definitions.
>
> I will need this facility for the TDM timing signals. It's a central
> office, not a datacenter.
>
> I don't know that Internet-based NTP would be accurate enough for the
> timing signals that I need. Maybe, maybe not.
>
>
>
>
>
> -----
> Mike Hammett
> Intelligent Computing Solutions <http://www.ics-il.com/>
> <https://www.facebook.com/ICSIL>
> <https://plus.google.com/+IntelligentComputingSolutionsDeKalb>
> <https://www.linkedin.com/company/intelligent-computing-solutions>
> <https://twitter.com/ICSIL>
> Midwest Internet Exchange <http://www.midwest-ix.com/>
> <https://www.facebook.com/mdwestix>
> <https://www.linkedin.com/company/midwest-internet-exchange>
> <https://twitter.com/mdwestix>
> The Brothers WISP <http://www.thebrotherswisp.com/>
> <https://www.facebook.com/thebrotherswisp>
> <https://www.youtube.com/channel/UCXSdfxQv7SpoRQYNyLwntZg>
> ------------------------------
> *From: *"Majdi S. Abbas" <m...@latt.net>
> *To: *"Mike Hammett" <na...@ics-il.net>
> *Cc: *nanog@nanog.org
> *Sent: *Thursday, July 11, 2019 9:54:26 AM
> *Subject: *Re: Time and Timing Servers
>
> On Thu, Jul 11, 2019 at 09:29:46AM -0500, Mike Hammett wrote:
> > There were a lot of NTP threads several weeks ago, but I didn't get an
> answer to my question amongst all of the other chatter.
> >
> > I'm looking for a device that can receive GPS inside a building without
> the
> > assistance of an external antenna (Frontier says they no longer allow
> > external antenna), will provide traditional NTP services, and will
> provide
> > a timing signal that my Metaswitch can work with.
>
>         Unfortunately, L band satellite signals are incredibly weak by
> the time they reach the surface.  It's very unlikely this is going to
> work for you (unless it's a wood framed single story building.)
>
>         Generally, I try to ensure that a GNSS antenna is built into the
> contract, to avoid games like this.
>
>         You have two options:
>
>         A) Find a new colocation provider.  This may already be on your
> to-do list for other reasons.
>
>         B) Rely on the Internet for timing, using NTP or PTP from
> another location to backfeed the site, and use a box with a good
> stable oscillator to keep time (this can actually be a commercial
> time server with decent holdover characteristics.
>
>         If you're just looking for alternatives to Microsemi, I highly
> recommend talking to the fine folks at Meinberg.
>
>         --msa
>
>

-- 
- Forrest

Reply via email to