The scope field makes it possible to control the scope of the desired broadcast very easily. This was done in IPv4 with the TTL value. The scope field has a major impact on the allocation process as it has to be specified for every allocation. Using scoped addresses makes it easier to control address allocations than with global addresses.

The Remaining 112 bits of the multicast address hold the identification of the group. If the T-flag is set, the identifier is valid inside its scope only. It means that the same identifier may be used outside the scope or with different scope to address another group. If the address is not transient, the address is independent of the scope. Its meaning stays the same and the scope just limits the spread of the subgroup involved.

For example, the (permanent) group identifier 101 (hexadecimal) has been assigned to all NTP servers. If somebody sends a datagram addressed to this group, scope-differing addresses have the following meaning:

• ff01::101 – all NTP servers on the same interface
• ff02::101 – all NTP servers on the same link
• ff0e::101 – all NTP servers in the whole Internet

The scoping is an innovative principle providing a neat mechanism to restrict the multicast distribution. Datagram lifetime limitation (TTL) is used in IPv4 to solve the same problem. Address scoping is a much more exact tool. Using the appropriate part of the address you define the area of the network which the distribution may not exceed – the datagrams never cross the boundary of their scope. So for example you may be sure that the datagrams addressed ff02:something never leave the physical network to which they have been sent.

Some multicast addresses are reserved, having predefined meaning. Some others are simply prohibited, like the addresses containing all-zero group identifier and a zero T-flag. RFC 3513 defines the meaning of some multicast addresses, de facto replacements of former broadcasts. They allow the sending of packets to all nodes or all routers in a given scope:

• ff01::1 – all nodes on the same interface
• ff02::1 – all nodes on the same link (layer 2 network)
• ff01::2 – all routers on the same interface
• ff02::2 – all routers on the same link
• ff05::2 – all routers in the same site

Another group of multicast addresses with the common prefix ff02:0:0:0:0:1:ff00:0/104 is dedicated to identify the solicited-node during neighbour discovery. Some other RFCs specify various dedicated addresses for particular purposes. The current list of reserved multicast addresses is available at

http://www.iana.org/assignments/ipv6-multicast-addresses .