IP multicast flows can be designated by these notations: - - (S,G) A unique shortest path tree structure between the source and the multicast destinations, pronounced "S comma G." S is an IP unicast source address, and G is the IP multicast destination address or group.
- - (*,G) A common shared tree structure, where a multicast rendezvous point (RP) accepts multicast traffic from the source and then forwards it on to the destinations, pronounced "Star comma G." The star or asterisk (*) represents the RP, because it is a wildcard source that accepts input from any real multicast source. The G represents the IP multicast destination address or group.
IP multicast or Class D addresses begin with 1110 in the most significant address bitsAddresses within the range 224.0.0.0 to 239.255.255.255. Hosts anywhere in the network can register to join a multicast group defined by a specific multicast IP address. Registration is handled through the IGMP. IP multicast addresses 224.0.0.1 (all hosts on a subnet) and 224.0.0.2 (all routers on a subnet) are well-known and don't require registration. You can find other well-known multicast addresses listed in Appendix B, "Well-known Protocol, Port, and other Numbers." Multicast also uses Ethernet or MAC addresses beginning with 01-00-5e. (The least-significant bit of the high-order byte is always 1.) The multicast IP addresses must be translated into multicast MAC addresses in this fashion, following the structure shown in Figure 9-1: - The 25 most-significant bits in the MAC address are always 01-00-5e. - The 23 lowest-significant bits are copied from the 23 lowest-significant bits of the IP address. - The address translation is not unique; 5 bits of the IP address are not used. Therefore, 32 different IP addresses can all correspond to a single multicast MAC address.
Figure 9-1. Multicast Address Translation
|