[aprssig] Impossible to be digi+something else? (DWAIT=0)
trasukg at trasuk.com
Wed May 18 12:42:07 CDT 2016
Our club just put up a new digipeater this past weekend (VE3RSB) running APRX as the IGate, but with the TinyTrak4 running as the digipeater. One interesting thing I found was that it would digipeat perfectly, but then if I hit the internet services, like WHOIS or EMAIL-2, it would be inconsistent if I got a response. Rare, actually.
Turns out, we had APRX set up to do the “viscous digipeating” mentioned below for APRS-IS packets. So if it heard the IS return packet first from a different IGate, it didn’t send the packet. And since the digipeater is up on a hill overlooking Lake Ontario, it has a pretty good chance of picking up a packet from somewhere else along the lake that probably also heard the original weak station’s digipeated request. When it heard the packet on RF from somewhere else, it didn’t send it out to RF locally. So the net result was that you had almost zero chance of getting a response from the “good” digipeater, basically because the digi has a far better vantage point than the handheld on my dining room table.
So the moral of the story - viscous delay is good for fill-in, bad for a wide-area IGate. Thinking about it now, though, I kind of wonder if it even makes sense for a fill-in. Assuming there isn’t too much difference in latency between stations receiving IS packets for retransmission, it might be best to just let them all send simultaneously, same as local RF digis. I’m not sure how good that assumption is, though.
> On May 18, 2016, at 12:27 PM, Frank Knobbe via aprssig <aprssig at tapr.org> wrote:
> On Wed, 2016-05-11 at 09:21 -0700, Bill Vodall via aprssig wrote:
>>> I'm curious because I'm doing some research on an idea for improved
>>> digipeating, and it requires using an external program with a KISS TNC. As
>>> such, I'd like to be able to selectively decide whether my station is
>>> "polite" or "impolite" on a per-packet basis (assuming impoliteness is still
>>> preferred digipeater behaviour).
>> Perhaps related is the viscous digipeating currently supported by
>> APRX. While theoretically controversial, some users find it performs
>> well. The concept is for a low level digi to not repeat a packet if
>> it hears somebody else digipeat it. Otherwise, after waiting a short
>> period and not hearing a transmission - it digipeats the packet.
> Yup, that's a great feature of aprsx. My Igate runs aprx and also acts
> as a viscious fill-in digi. It only does this for directly heard packets
> (non-digipeated, but regardless of path -- WIDE1 or WIDE2). Once a
> packet has been heard directly, it will wait up to 8 seconds and see if
> another (high) digi is digipeating it.
> If the heard packet has not been digipeated (high digis didn't hear it),
> mine will digipeat it. If someone else (like an area high-digi) has
> digipeated it, mine will ignore it and not digi it again. It cuts down
> on congestion by avoiding unnecessary duplicate digipeating, while still
> "filling in" when high digis did not digipeat a stations packet.
> I think staying quiet, except when a directly heard packet has not been
> digipeated by someone else, is very polite.
> aprssig mailing list
> aprssig at tapr.org
More information about the aprssig