TABLE OF CONTENTS Multicast for IPv6 is based on the same principles as multicast in IPv4 or, for that matter, multicast in link layer protocols like Ethernet.
Nodes that are members of a multicast group receive packets transmitted to the multicast group address.
The original sender of a multicast packet sends it only once, but the packet will be repeated so as to be delivered to all group members.
When a multicast packet is transmitted on a multiaccess network, in which all nodes can detect all transmissions but ignore those addressed to nodes other than themselves, group members all process the same packet transmitted by the original sender.
When a multicast packet is transmitted on a nonbroadcast, multiaccess (NBMA) network like ATM or Frame Relay, an intermediary node accepts the multicast and then repeats it to all nodes subscribed on the network.
Although the fundamentals are quite similar to multicast in IPv4, IPv6 protocols depend to a great degree on the use of multicast for functions such as neighbor discovery (see Chapter 13), node autoconfiguration (see Chapter 16), mobile IPv6 (see Chapter 17), and more.
As part of the renovation and simplification of IPv4-generation protocols, and to improve the efficiency with which the functions just cited are performed, IPv6 multicast has also been streamlined. The Internet Group Management Protocol (IGMP) used in parallel with ICMP in IPv4 has been dropped for IPv6. Instead, IPv6 multicast recipients are detected with the Multicast Listener Discovery (MLD) protocol, which is actually...
