| |
ZL3AI > APRS 29.10.00 15:59l 220 Lines 8075 Bytes #999 (0) @ WW
BID : 34162_ZL3AI
Read: GUEST
Subj: TAPR Digest, Oct 25, 5/6
Path: DB0MRW<OK0PKL<OK0PPR<OK0PPL<OK0POK<9A0YRB<HI8GN<KP4IG<7M3TJZ<IK1ZNW<
VE1DRG<WA7V<KB7FRV<VK4WIT<ZS6RO<ZL3VML
Sent: 001027/1039Z @:ZL3VML.#80.NZL.OC #:45150 [Chch-NZ] FBB7.00g25 $:34162_ZL3
From: ZL3AI@ZL3VML.#80.NZL.OC
To : APRS@WW
> APRS servers anyway, it seems like a convenient place from which to receive
> time synchronization).
> Of course, IGATES could also use GPS receivers for time synchronization.
This may not be practical in many installations. IGATES would have to purchase
or use an existing gps for this purpose.
> > If a time standard were adopted it should include a provision for an APRS
> > time message. This message could be originated by servers, and gated to RF
> > if desired. If APRS applications were modified to accept this message, it
> > would permit the synchronization of all stations to the same time base.
The
> > applications would have the responsibility of adjusting the system date
> > time if permitted by user settings.
> This sounds like a really useful idea. However, propagation delays are a
> bit harder to determine over the (multi-hop) APRS RF (or RF/Internet)
> paths. On the other hand, this still seems worth thinking about.
The time difference introduced by packet propagation delays would be in
seconds. This would be a vast improvement over the present situation where
station clocks are minutes, hours, or days off.
> It would also be nice if APRS clients used these on-the-air time messages
> to synchronize their local clocks.
That was the whole point of my post. "The applications would have the
responsibility of adjusting the system date time if permitted by user
settings".
Bill KC9XG
>
> -tjs
>
>
----------------------------------------------------------------------
Subject: RE: Next Layer of Standards
From: Bill Diaz <billdiaz@megsinet.net>
Date: Wed, 25 Oct 2000 15:25:39 -0500
X-Message-Number: 25
Alex,
By definition IGATES are connected to the Internet. As Tim pointed out, it
is a rather trivial matter to obtain accurate time via the internet. No extra
hardware is required. If the APRServers sent periodic time updates to
connected users and IGATES, all users could, if desired, sync to the same time
(with some acceptable propagation delay).
Bill KC9XG
On Wednesday, October 25, 2000 3:03 PM, Alex [SMTP:kg4ecv@amsat.org] wrote:
> I don't know if the hardware is readily available or not, but this would
> be a nice application for a digi with a DCF77 or WWV receiver (or
> perhaps a GPS receiver), rather than an Igate (which often first has to
> go through a digi to be heard). Especially since propagation times of
> packets are very hard to predict within acceptable margins of error, a
> client application should probably only accept packets heard from the
> Time Server within one or two hops. But that of course depends on the
> application. For most applications it would already be nice if we would
> know that we're talking about the _same_ time (utc, local...), but I
> think that that already is covered by the current standard if I'm not
> mistaken.
>
> 73s,
> Alex
> Bill Diaz wrote:
----------------------------------------------------------------------
Subject: RE: Next Layer of Standards
From: Timothy J Salo <tjs@tc.umn.edu>
Date: Wed, 25 Oct 2000 15:39:52 -0500
X-Message-Number: 26
> From: Bill Diaz <billdiaz@megsinet.net>
> Subject: [aprssig] RE: Next Layer of Standards
> Date: Wed, 25 Oct 2000 15:18:57 -0500
> [...]
> > IGATES could use any convenient Internet source for time synchronization.
>
> This is not being done by many IGATES now. Requires running a separate
> application periodically to set the clock.
> [...]
Time Protocol (RFC 868) is a trivial protocol. Presumably, it could
be implemented within IGATE software, rather than requiring an additional
application.
NTP is a fairly complex protocol. Any reasonable person would run a
separate, existing application to provide this functionality.
APRS servers presumably would want to run NTP, (presumably as an additional
application, as you indicate).
Alternatively, IGATES could use APRS messages to synchronize their own
time, but this solution is less general for Internet-connected machines
(it would allow IGATES to synchronize their clocks only with APRS servers,
rather than zillions of machines throughout the Internet).
-tjs
----------------------------------------------------------------------
Subject: RE: Next Layer of Standards
From: David VanHorn <dvanhorn@cedar.net>
Date: Wed, 25 Oct 2000 15:42:27 -0500
X-Message-Number: 27
>
>IGATES could use any convenient Internet source for time synchronization.
>Alternatively, the APRS servers could all offer time synchronization
>services to the IGATES (since the IGATES have to communicate with the
>APRS servers anyway, it seems like a convenient place from which to receive
>time synchronization).
This does seem pretty straightforward, and would elmiminate any
complidations due to daylight savings or other political wierdness. Leaving
the PC clock alone is a good idea, since there may be reasons that it has
to be set to some other value (don't know what, but principle of minimum
interference..) The PC-ware can read the hardware clock, and calculate an
offset that is then used for getting the time, until the next "noon cannon"
from the APRS server.
>Of course, IGATES could also use GPS receivers for time synchronization.
Gaaak.. Expensive, and yet another cable, serial port, and thing to find a
place for. I have no other need for GPS at an Igate, since I know where I am.
>It would also be nice if APRS clients used these on-the-air time messages
>to synchronize their local clocks.
Definitely better than nothing, and with most TNCs running software clocks
that only approximate correct time, an improvement.
----------------------------------------------------------------------
Subject: RE: Next Layer of Standards
From: Alex <kg4ecv@amsat.org>
Date: Wed, 25 Oct 2000 17:55:24 -0400
X-Message-Number: 28
Hi Bill,
I'm not as concerned where the time information comes from, be it from
the net, gps or wwv, as I am about the propagation of these time sync
messages on RF. Propagation delays on 144.39 are enormous compared to
the internet (although I doubt we have any time critical (as in split
second) applications on APRS). Another problem is how a client
application can verify that the time sync message it just received is a
valid one (how do they expire?). Perhaps the earlier mentioned RFC can
help us there.
73s,
Alex
Bill Diaz wrote:
>
> Alex,
> By definition IGATES are connected to the Internet. As Tim pointed out, it
> is a rather trivial matter to obtain accurate time via the internet. No extra
> hardware is required. If the APRServers sent periodic time updates to
> connected users and IGATES, all users could, if desired, sync to the same time
> (with some acceptable propagation delay).
>
> Bill KC9XG
----------------------------------------------------------------------
Subject: RE: Next Layer of Standards
From: Bill Diaz <billdiaz@megsinet.net>
Date: Wed, 25 Oct 2000 17:23:25 -0500
X-Message-Number: 29
Alex,
On Wednesday, October 25, 2000 4:55 PM, Alex [SMTP:kg4ecv@amsat.org] wrote:
> Hi Bill,
>
> I'm not as concerned where the time information comes from, be it from
> the net, gps or wwv, as I am about the propagation of these time sync
> messages on RF. Propagation delays on 144.39 are enormous compared to
> the internet (although I doubt we have any time critical (as in split
> second) applications on APRS).
I believe that propagation delays < 1 minute would be acceptable and would
result in a big improvement in Time accuarcy. This would eliminate or greatly
reduce the timestamp problem experienced with Mic-E packets.
>Another problem is how a client
> application can verify that the time sync message it just received is a
> valid one (how do they expire?). Perhaps the earlier mentioned RFC can
> help us there.
A mechanism must be put in place to ensure that only current time records would
be accepted.
If any APRS station could send a time message/record we could not rely on the
accuracy. If the APRS protocol were modified to include a time message or
Read previous mail | Read next mail
| |