Forum

This is from the SpectraLink 2006 "DEPLOYING NETLINK WIRELESS TELEPHONES: BEST PRACTICES" document:

The push-to-talk (PTT) mode of the NetLink i640 Wireless Telephone uses
SpectraLink?¡é?€??s proprietary SpectraLink Radio Protocol (SRP) ADPCM encoding. If a PTT broadcast is active (i.e. a user presses the PTT button), the feature will use the bandwidth as indicated in the table above for ?¡é?€??Push-to-Talk transmit?¡é?€?? for the AP with the transmitting NetLink i640 handset. All other APs in the system will use the bandwidth in the table for ?¡é?€??Push-to-Talk receive?¡é?€?? regardless of the number of Wireless Telephones using the AP. The data rate used for PTT depends on the AP?¡é?€??s settings for multicast traffic. Because the PTT mode uses IP multicasting, all APs on the subnet will transmit a PTT broadcast. This can be limited to only the APs that are handling one or more PTT-enabled handsets by enabling the Internet Group Management Protocol (IGMP) on the wired infrastructure network.

Knowing this, take a peek at Cisco's v3.1 implementation of LWAPP multicasting starting on page 6-1 of the Enterprise Mobility Design Guide 3.0. After saying to yourself, "holy smokes batman - using 3.1 or earlier code in a multicast-rich environment (like with spectralink push-to-talk), my network would be killed by this stuff", then consider how much worse it would be if the multicast frames were exceeding the MTU of the link between the controller and APs (so that each multicast had to be fragmented). Hideous. Moral of the story: upgrade to 3.2 or 4.0 airespace code QUICKLY if you have a push-to-talk, multicast-rich environment. yikes.

Devinator