MAC sublayer protocol

[barry@dgbt.doc.ca (Barry McLarnon DGBT/DIP)]

Glenn writes:
>Well, as one trying to whack hardware into bit pumps over radio I have a
>slightly different view.
>Full duplex is no cakewalk.. particularly for wideband/faster systems.
>Not only is the spectrum coordination issue more confusing but the
>hardware itself is *considerably* more difficult.
> [good discussion of the problems deleted]

I'll grant that it's no walk in the park; on the other hand, folks should
not be unduly scared of trying it.  As you point out, the most critical
element is the site: in general, the more desirable the site, the more
likely the RF problems resulting from shared usage.  However, the
application of full duplex I was addressing is CSMA MANs with restricted
coverage.  They don't need to be (and shouldn't be!) on mountain tops or
TV towers, so the possibilities of finding a reasonably clean RF site are
greater.

Although the spectrum coordination issue may be more confusing, the use of
crossband full duplex allows some room for creativity in dodging the
problem that Fred mentioned, where all of the coordinated in-band duplex
pairings tend to be locked up by the voice repeater crowd.

One final point, going back to the specific example of a 56kb crossband
repeater: there are some interesting upgrade possibilities.  Suppose the
usage of your 56kb MAN increases to the point where the performance falls
off noticeably - how do you expand the capacity?  You might upgrade to,
say 250kbps, assuming the gear is available, but this means wholesale
changeover of hardware, possibly including a change of bands to accommodate
the increase in bandwidth (if, for example, only 100kHz slots are available
in the existing band(s)).  You might put up a second 56kb MAN... assuming
for the moment you are using half-duplex nodes, and you don't want to go
and find a new site, then you have the problem of co-siting the two nodes.
This involves duplexers (assuming a pair of suitably spaced 100 kHz slots
is available, which may be unlikely), or going to a new band.  If you ever
needed to add a third node at the site, it could turn into a nightmare.
And it would be difficult for users to have access to more than one of the
MANs (i.e., select the one offering best performance, or switch if one goes
down) without investing in a lot of extra hardware.  Now suppose the initial
node includes a crossband repeater, and consider the possibilities.  A new
node/MAN can be added by slotting in a second crossband repeater, using the
same bands as the first one.  The new repeater can be multicoupled to the
existing receive and transmit antennas, and no duplexers are required.  In
fact, in the case of the DSY modem, it appears quite feasible to do the
multicoupling at the 28 MHz IF.  In this case, no new RF hardware would be
required!  Of course, things are never quite that simple - a higher-gain
omni transmit antenna or a linear amp might be needed to maintain the
coverage area, since the transmit ERP is now shared between the two
repeaters.  Sufficient 100kHz slots have to be available in the two bands,
and the IF filtering in the DSY receivers would have to be improved if the
slots were closely spaced.  Given the available slots, though, it would be
relatively straightforward to extend this "channel bank" concept to add
several more repeaters/nodes, with minimal expense for RF hardware.  And it
would be fairly easy for users to switch channels, by simply switching
crystals in their DSY modems (a simple synthesizer design might even be
feasible).  Intriguing concept, no?

Barry