PIM SSM Range Note
The PIM-SSM range uses the embedded prefix flags, which to me is slightly confusing.
As a reminder, the IPv6 multicast address is broken down into these fields:
R is set for Embedded RP
P is set for Prefix-embedded
T is set for Transient. When set to 0 it is a well-known address.
The P and T flags are set for two purposes:
Embedded RP
In this case, the R flag is also set, so all embedded RP addresses are FF7X::/12
Unicast-embedded unique multicast range
This allows an organization to embed its globally unique unicast block to produce unique multicast group addresses.
Ex. 2001:db8:1::/48 gives you a multicast group block of FF3X:0030:2001:db8:1::/96
What is slightly confusing is that SSM uses the P and T flags as well. The idea is that SSM uses these but the prefix length and prefix are set to zero. This shortens down to FF3X:0000::/32 where X is the scope and 0000 is the RIID and prefix length. This is not FF3X::/16, because you must set the prefix length to 0.
IOS-XE and IOS-XR routers will pre-map these ranges to the SSM range. These do not have an RP. You will see many listings, because one /32 is needed for every possible routed scope. (Every scope ID except for 1 or 2, as those are not routed).
It does not appear that you can change the IPv6 SSM range, as you can with IPv4. There shouldn’t really be a need to anyways, because there is so much available address space in these ranges.
When you want to use a unicast-prefix-embedded range for PIM-SM (not SSM), you can simply match the default FF00::/8 range, or specifically map this to an RP. Since the unicast-prefix-embedded range will have a non-zero value for the prefix length, it won’t match the FF3X::/32 ranges.
Last updated
