Order Tray | Contact Us | Home | SIG Lists

[aprssig] Local Event using RELAY?

Henk de Groot henk.de.groot at hetnet.nl
Wed Mar 30 19:44:11 UTC 2005


AE5PL Lists schreef:
> APRS is NOT ALOHA!!!  Read what the ALOHA project was and then we can
> discuss this.  AX.25 UI packets are transmitted using CSMA with a
> varying probability (depending on station types, power, location,
> terrain, etc.) that one station will try to transmit at the same time

Yes, I know. But when you are unable to detect any carriers of the other 
APRS stations because they are too weak or hidden from you then, in those 
cases, CSMA just sees a clear channel and hence transmit whenever there is 
something to tranmsit (with a short delay because of the persistence and 
slottime setting but otherwise pure ALOHA behaviour).

> another station is without being able to hear (at all) the other
> station.  In fact, the numbers you quoted are based on invalid
> assumptions and are of little value other than to mislead people about
> operating a "deaf" station.

Maximum APRS channel througput is much closer to this 18.4% (as calculated 
by Norman Abramson of the University of Hawaii for ALOHA when using 
variable length packets and no time-slotting) than to anything you would 
expect from even 1 persist CSMA. This suggests APRS behaviour is more like 
ALOHA than like CSMA. For CSMA the maximum channel load before collisions 
kill it should be much higher.

Assume your PERSIST setting is 128, this will equal to 0.5 persist CSMA 
(since there is a 50% chance you will transmit if the channel is clear). 
Under these conditions maximum channel load should be 70%, not 18%. (Note 
that for low values of persist the maximum channel load reaches near 100%. 
Sounds cool but the price to pay is a huge delay...).

But... even if you could hear and detect ALL your neighbours, then still 
CSMA is flawed for HAM radio. This is because you listen, then draw a 
random number for PERSIST and if you are clear to go the TRX takes anohter 
200 ms to change from reception to transmission and to generate a signal 
that your neighbours could detect. In those 200 ms your station is not 
able to listen anymore and this makes CSMA on HAM radio not even close to 
what it should have been in theory. This is one of the reasons why DAMA - 
which is a master/slave protocol - performs so much better.

Anyway. CSMA with all the hidden transmitters like in APRS (and in 
combination with the TX delays that HAM equipment needs) has only a little 
gain over having no carrier sense at all. Whether you like is or not, the 
carrier sense stuff in APRS saves very little and my feeling is that its 
more there to keep your conscience quiet than for any sensible technical 
reason. The odd packet this will save is not worth the effort.

Kind regards,

Henk.





More information about the aprssig mailing list