Re: NEW TOPIC! : What about an AX.25 Bridge?

[Phil Karn <karn@unix.ka9q.ampr.org>]

>I know there will probably be idealogical arguments against bridges but
>they are a useful tool and might provide some of the functionality we
>require! :-)

I considered them a long time ago but concluded that although bridging
is a very useful tool in a LAN environment, they have serious problems
in a random-access RF environment that make them less than workable.

The main problem is that bridges assume full bidirectional
connectivity within each of their subnets. That assumption is inherent
in the way a bridge works; if it hears a packet from an interface
that's destined to another host on the same side, then it takes no
action because it assumes that all hosts on that subnet can hear each
other directly. This assumption clearly doesn't work for simplex
multiple access radio.

Now if you wanted to establish bridges between full duplex repeaters
they would probably work well.  But I'm kind of down on repeaters
because of their low spectral efficiency (in terms of bits/sec/hz/sq
km).

Phil