[ax25-layer2] More thoughts

[pete at ae5pl.net]

> -----Original Message-----
> From: Timothy J. Salo
> Sent: Saturday, September 17, 2005 6:13 PM
> To: ax25-layer2 at lists.tapr.org
> Subject: RE: [ax25-layer2] More thoughts
> 
> I presented two examples earlier:
> 
> o	Routing is generally regarded as a network-layer function.
> 	The AX.25 specification includes routing functionality:
> 
> 	-	It specifies how the source can specify the route
> 		through the network by listing the nodes through
> 		which a packet must travel.  Most people call
> 		this "source routing".  (IP even includes this
> 		functionality, although most routers are configured
> 		to not support it.)
> 
> 	-	It specifies how nodes [routers] should digipeat
> 		[route] source routed packets.

This is not routing.  This is an implementation done by some people
using repeaters which are not network nodes.  Digipeating is not
routing, it is simply repeating the packet onto the same or another
medium within the AX.25 link.

> o	End-to-end reliability is generally regarded as a transport-
> 	layer function.  To the extent that acknowledgments and
> 	retransmissions are end-to-end, not hop-by-hop, this portion
> 	of AX.25 looks much more like a transport protocol.

End-to-end reliability is not in the AX.25 specification.  Link endpoint
to link endpoint is and that is within the realm of the link layer.  In
fact, each layer can (and most do) support some form of reliability
capability.

> I'm talking about the protocol, not various implementations.  

As shown above, you are talking about implementations, not the protocol.
The key is to understand that provisions for repeating of layer 2
packets within the link is not routing.  And the end-to-end reliability
you speak of is simply layer endpoint to layer endpoint which is within
the realm of each layer's definition.

> Where does this number come from?  Or, is this largely the 
> number of APRS nodes?  I question including users that use 
> only the datagram functionality of AX.25.

Yes, there are over 25,000 APRS users in addition to many non-APRS users
which there is no way to accurately quantify.  However, to ignore AX.25
UI users is to ignore a portion of the specification.

Once again, I challenge you to find anywhere in the AX.25 specification
where it specifically address performing other layer functionality
_within_ AX.25.  As pointed out above, people have attempted to do so
via various implementations, but the protocol specification is very
clear that AX.25 does not address other layers.

73,

Pete Loveall AE5PL