[aprssig] ax25 v 2.2

[John R. Ackermann N8UR]

Guys, just so you understand -- the AX.25 2.2 spec was done in 1997. 
Its digi limitation derived from the widely accepted view that for 
general packet use, digipeaters are a bad idea and should be discouraged 
in favor of more efficient network implementations.  Because other 
packet networking approaches such as NetROM and TCP/IP were becoming 
widely deployed at the time version 2.2 was written, the authors 
deliberately limited the number of digis the protocol would support.

Without igniting a religious war, I think it's safe to say that APRS 
uses AX.25 in a way that was simply not anticipated or intended by the 
authors of any version of AX.25.  Bob simply implemented a very, very 
clever hack on top of the protocol.  Digis are good for APRS, but they 
are not good for the idea of packet radio networking as it was 
envisioned by the AX.25 authors.

Over the last 8 years, there's been little interest in implementing the 
version 2.2 protocol -- to my knowledge, only the Linux AX.25 stack has 
implemented any part it, and there its use is optional.  Also, it was 
never "officially" adopted, as no one actually owns AX.25 -- TAPR 
supported the working group that developed it, and published the spec, 
but there's nothing out there that says version 2.2 shall rule over 
version 2.0.

In short, I don't think there's much point in getting exercised over the 
changes in version 2.2.  It's not exactly a steamroller coming down the 
road.

John
----


Robert Bruninga wrote:
>>Remember that the limit isn't a "two-hop" limit, 
>>but a 2-digi limit. If your packet digi path is: WIDE4-4, 
>>then you have 1 digi, but get 4 hops...
> 
> 
> The WIDEn-N system is *not* in AX.25 because it is an APRS
> specific APRS enhancement by some MFRS and APRS 
> software supporters.  In fact, we have made MANY 
> enhancements that the TAPR AX.25 team has ignored 
> in their new spec.
> 
> We must get them to open their eyes to recognize that
> APRS is a well established AX.25 application and they
> are being short-sighted in their new spec... (or at least
> the one I saw a year or so ago)...  de WB4APR, Bob
> 
> 
>>>>josecanuc at gmail.com 1/14/05 12:43:45 PM >>>
> 
> 
> I'm not certain, but the 2-digi limit appears to be a push to limit
> source routing and let the digis decide how to route. Don't know how
> many digis do this, certainly no APRS digis. It would seem to be in
> the frame of mind of real "networks" of digis, like the networks of
> the Internet, with router configuration protocols, etc.
> 
> -Lance KJ5O
> 
> On Fri, 14 Jan 2005 12:36:08 -0500, Robert Bruninga <bruninga at usna.edu>
> wrote:
> 
>>>AX25 version 2.2 limits the number of digipeaters to 2.
>>>Has anyone implemented 2.2?  What effect would the
>>>2  digipeater limit have on APRS?
>>
>>The people who are writing 2.2 do not seem to recognize
>> APRS or other forms of UI communications.  We agree
>>completely that more than 2 hops for connected protocol
>>is not wortth trying on a typical shared channel.  But
>>the SPEC should not be used to try to overcome dumb
>>uses of the hardware.
>>
>>I can imagine several scenarios where multiple level 2 hops
>>over 2 may be useful and may be quite reliable.  Yes, these
>>scenarious are RARE and unusual, but the spec should
>>allow them!
>>
>>Further, although APRS works best at about 2 hops in very
>>dense areas of 60 to 100 stations, it can easily use more hops
>>in many low density applications.  Again, the hardware and firmware
>>of a TNC should allow greater hops for RARE or unusual
>>applications.  Besides one-way UI packets like APRS are
>>about equally reliable over FOUR hops as a connected
>>packet is at 2 hops.  THus if these spec writers are hard
>>set on a 2 hop limit for connected protocol, then they
>>must allow a 4 hop UI as a minimum equivalent limit for UI..
>>
>>Who do we get this common sense to?
>>de Wb4APR, Bob
>>
>>
>>Mike
>>
>>_______________________________________________
>>aprssig mailing list
>>aprssig at lists.tapr.org 
>>https://lists.tapr.org/cgi-bin/mailman/listinfo/aprssig 
>>
>>_______________________________________________
>>aprssig mailing list
>>aprssig at lists.tapr.org 
>>https://lists.tapr.org/cgi-bin/mailman/listinfo/aprssig 
>>
>>
> 
> 
>