MCT 1 sends LD PDUs on VLAN x on all the interfaces that are part of the CCEP, client edge port (CEP), and ICL interface.
If the Client has LD configured on the LAG interface, then it drops the PDUs and no loop exists. If there is a misconfiguration, the Client floods the PDUs and they reach MCT 2.
MCT 2 floods the PDUs back to MCT 1, where the loop is detected. With loose mode no information about the interface that transmitted the PDU is encoded in the PDU, so normally the receiving interface is shut down. Because in this case the PDU is received on the ICL interface, that interface is not shut down.
MCT 1 receives the loop detection PDUs on the CCEP interface as well, as the packets were flooded in the VLAN in the following sequence: MCT 1 > MCT 2 > Client > MCT 1. In this case the receiving CCEP is shut down to break the loop. For MCT 2 to forward the PDUs in this case it must be the designated forwarder (DF) for that VLAN.
MCT loose mode: Use case 2
Use case 2: LD enabled on VLAN
x on MCT 1 and MCT 2
Both MCT 1 and MCT 2 will flood the PDUs in VLAN
x on all the interfaces that are part of the CCEP, CEP, and ICL interface.
Assuming PDUs from MCT 1 take the path MCT 1 > MCT 2 > Node > Client > MCT 1, then the receiving CCEP interface is shut down. For MCT 2 to forward the PDUs in this case, it must be the DF for that VLAN.
Assuming PDUs from MCT 2 take the path MCT 2 > MCT 1 > Client > Node > MCT 2, then the receiving CEP interface is shut down
If PDUs from MCT 2 take the path MCT 2 > Node > Client > MCT 2, then the receiving CCEP interface is shut down.
Multiple interfaces can be shut down in this case, depending on the sequence in which loops are detected.
In addition, to avoid CCEP interfaces from being shut down over a CEP interface, the user can configure a CCEP port not to be shut down.
