In response to reader requests for a refresher in basic wireless LAN operations, I'll finish up a discussion I began last time of 802.11 WLAN channel assignments and user access.
Some of you have asked if a "channel" and a "port" are the same thing. The answer is no. A channel is a segment of the total radio frequency spectrum in a given band available for communications to and from a given radio. A port usually refers to the connection between the WLAN and the wired network. This might be the jack on an Ethernet switch into which you plug an 802.11 access point (AP), for example.
Once an AP radio infrastructure is configured across multiple channels, as discussed last time, how do users access and share the network?
Users' client network interface cards (NIC) discover and associate with the "best" AP in a certain area for communication. This is usually the AP that has the strongest signal or is least congested. The NIC transmits and receives over the channel currently assigned to that AP. Users within range of a given AP at any point in time share that channel and contend for its bandwidth.
The way NICs access a channel is defined by the 802.11 media-access control specification, which is similar to the MAC in 802.3 wired Ethernet standards. The MAC protocol for 802.11 uses a scheme called carrier-sense, multiple access, collision avoidance (CSMA/CA). This "listening before talking" protocol avoids collisions entirely, rather than backing off the transmission medium upon detection of a collision, as with traditional wired Ethernet. The reason for collision avoidance is that it is difficult to detect collisions in a radio frequency network.
The PHY (Layer 1) portion of the 802.11 specification also plays a role in the CSMA/CA operation by sampling the energy over the transmission medium to determine whether the channel is clear.
To minimize the risk of collisions, the 802.11 standard includes a MAC function called request to send/clear to send (RTS/CTS). If data arrives at an AP destined for a wireless node, the AP sends an RTS frame to the wireless node requesting a certain amount of time for delivery. The wireless node responds with a CTS frame that it will hold off any other communication until the AP finishes sending.
Other "listening" wireless nodes hear this transaction and delay their transmissions, as well.