Forum

For 802.11g, I believe the figure for CCA is -76 dBm with a valid decode, although I have been told by one large manufacturer that their G receiver used a different figure. I needed the information for a WLAN design. This is looking just at the situation for 802.11g, without getting into energy detect without decode for 802.11a. Since the stated sensitivity for a 24 Mbps connection for a Cisco 802.11a/b/g card is -84 dBm, this is starting to look interesting.

If you had tight control of the client radios, you could measure and estimate and control a client collision domain. Without control of the client radio, your controller could know the collision domain for clients as it pertained to AP radios. You could estimate client-client interference by calculating the position of each client using multiple receivers. With enough CPU power, you could probably infer client-client interference.

How much difference would there be in throughput for a number of constant bitrate clients as opposed to HTTP or letting other TCP protocols fill in the blanks?