- 802.11g repeater repeating a repeater
- Posted by eug k on June 13th, 2004
hi!
Can a Linksys 802.11G AP in repeater mode repeat a
repeated signal?
i.e.
[AP] <---> [repeater] <---> [repeater] <---> [client]
i'd assume the speed would be halved once again if
it's possible...
thanks!
--
- Posted by Gregory Toomey on June 13th, 2004
eug k wrote:
Yes, it works. It the principle behind all the community wireless networks
that are springing up.
Speed is about the same but latency will be longer.
gtoomey
www.ausinvestor.com Australian Investor Forum
- Posted by Uncle Bully on June 13th, 2004
"eug k" <wired_au@change-this-to-ho.tmail.com> wrote in message
news:cag7u2$37l$1@bunyip.cc.uq.edu.au...
If it halved then it wouldn't be a repeater would it. The whole purpose of
a repeater is to pick up a weak signal and repeat it at full power.
- Posted by eug k on June 13th, 2004
In comp.sys.ibm.pc.hardware.networking Uncle Bully <wakeupcall@optushome.com.au.remove> wrote:
that is how a normal FDM-based [normally] passive repeater works. Receive the
RF signal, amplify it, and retransmit it at full power... no processing done.
wifi repeaters are active and TDM-based, which means that half the time
it's receiving and half the time it's transmitting on the same freq ranges
because there's only one transceiver. it stores the data after receiving,
and forwards it in the next frame. So the net result is, your speed is
halved.
(fdm = frequency division multiplexing, tdm = time division multiplexing)
--
- Posted by eug k on June 13th, 2004
In comp.sys.ibm.pc.hardware.networking Gregory Toomey <nospam@bigpond.com> wrote:
excellent, thanks!
hm, are you sure? how would it get around the inherent speed-halving
of the store-and-forward method?
--
- Posted by Gregory Toomey on June 13th, 2004
eug k wrote:
TCP/IP is a sliding window protocol. You keep on sending packets af full
speed, and the receiver sends ACK (acknowledgemets) back. There can be
quite a delay betweend sending a packet and sending the acknowledgment, and
many packets in the "pipeline".
See
http://www.efd.lth.se/~d92pa/SUB/Lin...er/node57.html
gtoomey
www.ausinvestor.com Australian Investor Forum
- Posted by eug k on June 13th, 2004
In comp.sys.ibm.pc.hardware.networking Gregory Toomey <nospam@bigpond.com> wrote:
yes, but APs only have one radio, so they can only transmit on
one channel at once. So in repeater mode, it receives it, stores
it, and retransmits it
Just found a site that explains it well:
http://www.cisco.com/univercd/cc/td/...g/br350ch1.htm
(or)
http://tinyurl.com/35e9s
check figure 1-4, Repeater Bridge.
What I wanted was confirmation that it works with the linksys
in particular. Looks like it probably will! 
Hooking up two APs back to back will solve the bandwidth
issue, but that'll double the costs.
thanks
--
- Posted by Yousuf Khan on June 16th, 2004
eug k <wired_au@change-this-to-ho.tmail.com> wrote:
But the halving of receive and transmit would occur whether or not you had a
repeater or not. That is to say if you had just a single base station, it
would spend half of its time receiving data and the other half transmitting
it. The halving of the bandwidth has already occurred at the first hop. Just
adding additional access points is simply going to add latency to the
signal, but it won't reduce bandwidth.
Yousuf Khan
- Posted by eug k on June 17th, 2004
In aus.computers Yousuf Khan <bbbl67@ezrs.com> wrote:
That would be true if all the additional access points did were to receive
the RF signal, amplify it, and retransmit it. But that is not what happens.
As you've said, wifi is half-duplex. The second access point receives a frame
from the first, stores it, switches to transmit, reads it, then transmits it.
The halving occurs on every hop. The net result is half the bandwidth
and increased latency at each hop.
This is well-documented and widely known, e.g. in the cisco link i posted earlier
in this thread.
one way to get around it is to use two APs back-to-back. But that would double
the costs. 
--