With optimized IP multicast replication, a controlling bridge (CB) only sends one copy of the IP multicast traffic per receiver VLAN for a particular IP multicast group. For example, if group 225.0.0.1 has five receivers in VLAN v1 and four receivers in VLAN v2, the CB sends one copy to VLAN v1 and one copy to VLAN v2 and the bridge port extender (BPE) does the replication.
If the receivers of VLAN v1 (or VLAN v2) are spread across different native cascade ports, separate copies might be sent for VLAN v1 on each native cascade port depending on the best path to reach a receiver. However, when the receiver is an extended port LAG, the CB has to do the replication. Because extended port LAG members could belong to different BPEs, and because BPE cannot do the pruning, the CB has to send copies to the individual LAG member. For instance, if VLAN v1 membership includes a LAG comprising ports 101:1, 103:1, and another LAG comprising ports 102:1, 102:2, two copies are sent, one for the 101:1 LAG member and another one for the 102:1 LAG member.
Because replication and VLAN flooding both use multicast ECIDs, using optimized IP multicast reduces the number of VLANs with extended ports to 3,000.