When users roam from one AP to another, you want them to make that transition as seamlessly as possible. They'll probably keep the same IP address (renewed by DHCP when connectivity is reestablished). They'll also continue to use the same wireless quality of service and security measures, no matter which AP they're connected to. Depending upon your security policy and WLAN hardware, this could even mean avoiding re-authentication and using the same encryption keys to speed handoff.
An SSID (technically, an ESSID) is simply a unique name associated with all of these network characteristics. As signal strength wanes, each client scans the air, looking for other APs with the same name (SSID) that offer better signal. If you don't use the same SSID for both APs, the client may be forced to disconnect and (depending upon client configuration) may not connect to the other AP without user intervention.
Conversely, if APs are incorrectly configured with the same SSID but different QoS/security/LAN parameters, users may be able to connect to one AP but not the other, producing connections that appear to flip-flop up and down. If you really need two different QoS/security/LAN combos, assign each combo its own SSID and configure both APs to beacon both SSIDs.
Finally, why should your APs use different channels? In a typical micro-cell WLAN, adjacent APs should always use non-overlapping channels (e.g., 1, 6, 11) to avoid co-channel interference. The only cases where this is not true are products that use single channel architectures (e.g., Meru, Extricom). In those WLANs, a controller coordinates how airtime is used to avoid interference between APs and speed AP-AP handoff.
This was first published in June 2009