Order Tray | Contact Us | Home | SIG Lists

[aprssig] New n-N success in North Carolina

Wes Johnston aprs at kd4rdb.com
Sat Feb 12 22:20:11 UTC 2005


I love a good discussion....

> If I follow your example correctly, we're saying the exact same thing.
> I simply left out the WIDEn-n terminology in favor of hop count a)
> because it's simpler, b) it is the terms we should be thinking in, and
> c) it's the concept we should be shifting toward.

Well, yes, we are kinda saying the same thing.... what you are doing is putting
a cap on the number of hops a packet can take beyond you.... what I'm
suggesting is that we pass it along as-is it hasn't hopped too many times
already....

In either case, we need to also figure out a way to stop things like
relay,w7-7,w7-7,w7-7 from entering the network too...

But I'm gunna think aloud here... I've been looking at this from the city's
perspective with too long path packets streaming in from rurual america... the
other way around is a guy in a city TX'ing 7-7 packets... which escape the city
in two hops... and then pummel the rural areas with the 5 remaining hops... this
could happen using my method of hop limits and that's a bad thing... your method
would stop that for sure.  Your method would cap the hops of a packet
originating in the city to two and save the rural areas from those packets.

Maybe when/if a micro controller gets written it'll give both options to the
digiowner... or maybe it'll have a lookup table like digi ned does.

Jason, as I typed the table below, I have come to see more of the merits of your
idea... I'll really have to give some thoughts to the macro/micro network
implications of both methods.

Just to clarify the five philisophies we are talking about... KF4LID comes thru
town transmitting w7-7 every 60 seconds.  (I sure hope no one actually has the
call kf4lid hihi.)

Bob's idea of trap out for wide7-7 (i'm going to omit 6-6 and 5-5 for brevity)
digi hears   digioutputs
----------   -----------
w7-7         w7-7*
w7-6         w7-5
w7-5         w7-4
w7-4         w7-3
w7-3         w7-2
w7-2         w7-1
w7-1         w7-0*

Bob's/Wes' idea for cut-off solution - here we support only w2-2,w2-1,w3-3,w3-2
digi hears   digioutputs
----------   -----------
w7-x         no digipeat
w6-x         no digipeat
w5-x         no digipeat
w4-x         no digipeat
w3-3         w3-3*
w3-2         w3-2*
w2-2         w2-2*
w2-1         w2-1*

The next two ideas are not able to be implemented TODAY.. they would require
modification of the ROMs in KPC and Paccomm TNCs.  However, either of these
methods can be done TODAY with digined.

Jason's idea of asserting hop limits at the first digipeater.  In this case,
we're going to limit to 2 hops.
digi hears   digioutputs
----------   -----------
w7-7         w7-2
w7-6         w7-2
w7-5         w7-2
w7-4         w7-2
w7-3         w7-2
w7-2         w7-1
w7-1         w7-0*

Wes' idea of a per digi hop limit... in this case, we're also limiting the
packets to 2 hops.
digi hears   digioutputs
----------   -----------
w7-7         w7-6
w7-6         w7-5
w7-5         no digipeat
w7-4         no digipeat
w7-3         no digipeat
w7-2         no digipeat
w7-1         no digipeat

I can see that Jason's idea protects the network... Wes' idea protects only the
local digi.

Pete Loveall ae5pl came up with this idea in Packet Status Register winter 2005.
 The w7-7 is substituted by the digi's own callsign upon digipeating the packet.

digi hears   digioutputs
----------   -----------
w7-7         kd4rdb*
w7-x (x<>7)  no digipeat
w6-6         kd4rdb*
w6-x (x<>6)  no digipeat
w5-5         kd4rdb*
w5-x (x<>5)  no digipeat
w4-4         kd4rdb*
w4-x (x<>4)  no digipeat
w3-3         kd4rdb*
w3-x (x<>3)  no digipeat
w2-2         kd4rdb*
w2-x (x<>2)  no digipeat
w1-1         kd4rdb*
w1-x (x<>1)  no digipeat

>From there, all digipeaters who know kd4rdb is local to them will digipeat this
packet one time each.  They will keep a checksum of each packet they TX for
30seconds so they don't Tx the same one twice.



Wes




More information about the aprssig mailing list