[aprssig] NETROM Throughput?

[Robert Bruninga]

Ancient Packet Question:

For a long string of NETROM nodes operating on alternating A/B BANDS (full
duplex), would throughput be efficient or would ACKS from the next digi
collide with incoming traffic from the previous one?

Consider a string of 9 nodes in a row:

--1--  --2--  --3--   --4--  --5--  --6--  --7--   --8--   --9--
TXA   RXA  TXA   RXA  TXA   RXA   TXA   RXA   TXA
RXB  TXB   RXB  TXB   RXB   TXB   RXB   TXB   RXB

If I remember conventional traffic, a sending station will send up to seven
packets in a single transmission and, on a half duplex channel,  the ACKing
station will ack the last one he got correct..  But if it is a full duplex
channel, would the single incoming transmission of 7 packets be generating
a continuous stream of all seven acks since the receiving node does not
have to wait for the channel to clear?  What setting tells it to wait to
see if there are more packets before originating a single ack?

It would seem that for only a few packets (all fitting in a single
transmission) that the bulk of traffic might ripple quickly along the
chain.  But as soon as more traffic was pending at the source, or a
continuous feed of traffic, that things would clog up quickly.  Then this
might be no better than all TX/RX on a single frequency half duplex anyway.

Or is max throughput only possible with every link having a clear frequency

---1---   ----2----    ----3----    ------4-----   ----5----
-----6-----    ----7-----
TXA     RXA/RXB     TXB        RXB/RXC     TXC        RXC/RXA      TXA
   RXC    TXC       RXC/RXA      TXA        RXA/RXB     TXB         RXB/RXC

I guess I never ever operated conventional packet over full duplex links
and so have no basis for thinking it through..

Bob, WB4aPR
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.tapr.org/pipermail/aprssig_lists.tapr.org/attachments/20180303/175ab833/attachment.html>