Brocade Multi-Service IronWare Switching Configuration Guide (Supporting R05.6.00) - Manual (Page 705)

Multi-Service IronWare Switching Configuration Guide 675
53-1003036-02
MCT for VPLS
18
Local switching with MCT
With MCT-VPLS support, local-switching refers to traffic switched between end-points of both the
MCT nodes of the cluster.
Identical vpls-local-switching configurations are required on the two MCT nodes for either
supporting local-switching or disabling local-switching.
When local-switching is disabled, packets from the Active MCT node end-points are not sent to
MCT-Spoke PW.
Packets from standby MCT node are always sent to MCT-Spoke PW.
CPU protection with MCT
CPU protection is supported with MCT VPLS.
CPU protection can be turned on independently on each MCT node of the cluster. When the MCT
VPLS is enabled for a VPLS instance, it will use cpu-protection support with MCT.
Auto-discovery with MCT
VPLS auto-discovery is where the VPLS peer’s are discovered by BGP and added into VPLS instead
of manually configuring the peers.
If the MCT peer node is discovered as a VPLS peer by BGP, then it will not be added as VPLS remote
peer
NOTE
This special handling is only for “MCT enabled VPLS Instances” and will not affect non-MCT VPLS
instances.
Cluster-peer verses vpls-peer
Cluster-peer PW is the peering session between the two MCT nodes of the cluster. It is called
MCT-Spoke PW.
• MCT-Spoke PW is different from the regular PW.
• MCT-Spoke PW bring up will be triggered only when the configuration sync between the two
MCT nodes succeeds.
• MCT-Spoke PW signaling is triggered even though there are no local end points that are up or
configured.
• MAC learning is disabled in software and hardware for the data traffic received on MCT-Spoke
PW.
Graceful Restart and Upgrade
You can gracefully restart a node. If the node which is being brought down for maintenance has an
“Active” role for any VPLS instance, that node will become the standby. This can be done using the
client-interface shutdown command.