Knowing that AP's operate at half duplex, what in the wireless technology prevents it's operation at Full duplex in where the sending station (client) sends/receives on a channel and the receiving station (AP) sends/receives on matching channels? Much like the way a switch operates?
I have not participated in discussions with Wi-Fi designers but I have read the standards multiple times. Here is what I think.
A full duplex Wi-Fi could be built but it would be prohibitively expensive compared to the Wi-Fi we have. Each station would have to have redundant radio resources and a management protocol to assign their use. The matching of acknowledgment frames to data frames would become much more complex. More than three stations interoperating would be especially challenging as there are only three 2.4GHz non-overlapping channels to be used simultaneously in any one space.
Some might say that a "full duplex" bridge link can be created between two locations with two pairs of stations, directional antennae, and two channels. Nonetheless each pair of stations is operating in half duplex and together are using a lions share of the available channel spectrum.
As for the comparison to wired switches, I believe RJ45 has 2 individual pins for Rx and Tx each, which allows full duplex. Replicating this in Wireless would be as described in the example above.
I am not a standards writer, only a standards reader. Here is what I think.
Yes, 10BASE-T and 100BASE-TX Ethernet each devotes one pair to transmit and one pair to receive. Whether the cable is used in full or half duplex is up to the transceivers at the two ends. A similar case can be made for 10BASE-F, 100BASE-FX, and 1000BASE-F with fiber strands dedicated to transmission in one direction. Ethernet scales with massive numbers of closely packed cable segments.
These are analogous to two dedicated Wi-Fi radio channels in a hypothetical full duplex Wi-Fi. Unlike Wi-Fi radios each Ethernet cable brings with it additional dedicated communications channels.
1000BASE-T uses all four pairs to transmit and receive simultaneously and there are no half duplex 1000BASE-T transceivers. Relatively sophisticated circuits compare the difference between what is being transmitted and received in order to detect and decode the relatively weak receive signal.
In 2004 the Ethernet standard was amended to include descriptions of Subscriber Ethernet, also known as Ethernet in the First Mile (EFM). The several additional physical layer alternatives are full duplex only and include single pair category 3 copper and single strand single mode fiber implementations. Because any one transceiver only receives from one other transceiver ever, modern Ethernet still scales. (One of the EFM PHYs involves a multi transceiver topology and a TDMA based access protocol, but lets not go there now.)
These modern Ethernets are analogous to a single Wi-Fi radio channel being used simultaneously in a hypothetical full duplex Wi-Fi in a topology limited to two stations only. As soon as more stations joined the topology the complexities of separating signals would be orders of magnitude more daunting than modern Ethernet complexities.
The fundamental Wi-Fi limitations to full duplex operation are that there are precious few Wi-Fi radio channels available for simultaneous use in a given space, and using the same radio channel simultaneously for multiple messages in the same space is not inherently limited to two stations.
I beleive that the SDR stuff has the most promise for making this happen.
Do you all think that 3G will provide a better platform for what lacks in the 2.4Ghz band? In comparison to 10Mbps/half-duplex - 100Mbps/full-duplex)I'm comparing oranges and tangarines.
I don't think so, 3g is a stop gap.
Wimax(800lb gorilla) will either be a hit or miss. 11n could be something for the wisp gear makers but who knows it changes so fast every 6 months. But it is exciting.
I agree with what was written above.
In fact, to listen, the environment has to be
quiet. To speak, and for the other party
to be able to hear the transmission, the
environment has to be reasonably silent.
Now, somone cannot listen and talk at the same
time, without complicating the process.
In fact, if there was a 2.4 ghz frequency to send,
and 5 ghz frequency to receive on the same AP
or wireless router, it would probably make
the task feasible, but more expensive.
However, the lag time, of half duplex is neglible
compared to the advantages of potential full duplex -
at least as far as wireless is concerned.
In my ongoing research, I was contacted by a rep from Dragon Wave on their top line product, "AirPair, a 50 and 100mbps full duplex Ethernet radio solution that may not only be an ideal option to extend a service offering to a customer base, but can also act as a highly reliable, carrier grade backhaul solution for various network and LAN applications. We have three flavors to our radios, 18 and 23 Ghz licensed, and 24 Ghz unlicensed, which in some instances can deliver service at 99.99% availability over 20 miles."
Love to have a demo!
I have installed some Proxim (Tsunami) 54mb full duplex radios. PTP 8miles no problems