"...to eliminate the chance of a collision with another (foreign) unlicensed transmitting device, in the same chunk of the spectrum ?"
Marden, what you are referring to here is the ability to reuse spectrum,
which is relevant but not what I was referring to in the context of
collisions. My reference was to MAC (media access) layer collisions due
to high levels of contention by individual users who are competing to get
through, and not to spectrum allocation, per se. Not in mid air, but at the
base stations, or even at the cell sites, depending on design.
Rather, they result from high levels of contention for limited resources and
they take place on the back ends of systems between the base station's
air interface and the first point of arbitration looking towards the "cloud,"
whatever that cloud might be (Internet, PSTN, Cellular, etc.).
The contention domain itself could be an Ethernet collision domain (or
multiple switched/bridged segments) that sits between the radio receiver
and a router which interfaces to the Internet, or it could be an ATM
switch fabric on the back side of the station gear, or it could even be
channelized paths (like discreet channels in a larger network element such
as a channel bank or a mux) which are "cut through" in nature. Thus,
unimpeded, except for when all channels are being used, and then the
customer does not get a proceed to send or ready indication, and is
advised to "try again later."
The latter is an example of a deterministic connectivity mode [you are
either turned on or you are off for the duration of a session, no in
betweens or doubts], whereas the former examples are simply best effort
for the most part, with the potential for variable performance on the basis
of how many others are seeking entry at the same time.
This is the syndrome that portends to wreak havoc on cable modem
segments which support high numbers of users if they do not get upgraded
to deep fiber fixes, reducing the number of homes served per segment. It
would still be contention based, but here the 'headroom' once again would
be perceived to be sufficient, until the next plateau of user bandwidth
demands are reached. It's a constant tuning process.
In order to be more specific on these matters would require pinning down
the type of application (mobile pcs, stationary Internet access, MMDS
used for video/voice/date, and so on), and what types of air interfaces
were used, and what form of network protocol is used. Is it switched,
routed, perhaps both? Or... VDMA? (Hi, Bernard!)
These were among the points I was referring to in an earlier message
today. Not enough is being discussed on these issues, IMO. And while I
don't want to minimize the work that has gone into the design of air
interfacing and topology designs, these other attributes can become even
more pivotal, in the end, than the slight differences that may exist between
two different types of air interfaces which have already made it to the
short list. To wit, I stated recently, elsewhere:
from:
techstocks.com
----begin snip"
"Can help be far behind?
"I'm beginning to envisage a time when radio transmitters and terminals will
be able to adapt to xyzFDMs, or whatever, on the fly. It seems a bit far
fetched right now, admittedly, but I think that the larger service providers
may demand it.
"What I'm suggesting (may already exist in some ways?) is in many ways
similar to what's happened in the VoIP arena, where gateways can now
translate myriad compression algorithms on the fly, owing to the enabling
characteristics and heavy dependencies on DSP technologies.
"In an earlier time, these algorithms were viewed as a possible area that
would cause a wide schism among vendors and service providers, alike.
Today, it's not exactly a no brainer, but it's highly manageable and
becoming less of an issue, as time goes by. Soon, VoIP algorithm
translation issues will fade into the background as an historical anecdote.
"Will this happen here, in the wireless [Internet access] domain, as well,
when consumer level purchasing of end point wireless provisions have
been relegated to commodity status? I suppose the same question could
be asked of cable operator CMTS and cable modem architectural
considerations."
-----end snip
Regards, Frank Coluccio |
