I use two networks. One network with default threshold and the other with RSSI threshold as 0. Both the networks are sharing the network equally. Ideally the network with CS turned off should give much higher throughput right? I think the one with low RSSI is still competing fairly with the other network.
thanks sadeep On Mon, Mar 31, 2008 at 1:54 PM, Luis R. Rodriguez <[EMAIL PROTECTED]> wrote: > 2008/3/31 Sandeep Kakumanu <[EMAIL PROTECTED]>: > > Hi, > > > > Is it true that setting a really low AR5K_TUNE_RSSI_THRES value (0) in > > ath5k.h disables carrier sense? I find that doing so does not alter the > > behavior of my cards. > > > > I use linksys wmp55ag PCI cards (atheros chipset 5213). My setup is as > > follows. I set two wireless networks in the same channel (5.24GHz) using > two > > wireless APs. Two PCI cards on different machines using the ath5k driver > > work as clients. I send UDP traffic (iperf) from each of the clients to > > their respective APs. I get the same throughput through each network > (about > > 17Mbps on either network) irrespective of the different RSSI thresholds > I > > use. In particular I tried a scenario with one card having a threshold > of > > 255 and the other with 0. If setting the RSSI to 0 does infact disable > > carries sense, I should be observing a higher throughput in the network > with > > lower RSSI. > > This is only true if your previous throughput tests are in a busy > wireless network and therefore are sharing the medium. If you're the > only one on your network you shouldn't see any difference AFAICT. To > really determine if modifying the carrier sense threshold worked you > should see if you your node is always TX'ing even if other nodes are > present who should be competing fairly for the medium. > > Luis > -- Sandeep Kakumanu Presently doing PhD (Gatech Atlanta) http://users.ece.gatech.edu/~ksandeep/
_______________________________________________ ath5k-devel mailing list [email protected] https://lists.ath5k.org/mailman/listinfo/ath5k-devel
