Owner's Manual

spanning tree protocols (STPs) used by legacy networks. Unlike STP, there are no blocked ports in DLF

protocols and all ports in the network are used for packet forwarding. The single shared tree is used only

by the broadcast, unknown unicast, and multicast (BUM) packets. The known unicast traffic always uses

the shortest path between two end points. DLF prevents data loops because all DLF traffic can only go

through the nodes connected by the shared tree.

NOTE: DLF provides node and link resiliency, but requires a longer convergence time than VLT. DLF

does not support L2 multipath with paths that span switches.

When identifying a topology consisting of a subset of switches connected in a VLT-compatible method

and using VLT, the controller uses the DLF tree protocol to connect the remaining switches and VLT-

compatible switch groups.

DLF Port Roles

DLF roles are assigned to the ports of all switches in the network, except for the root switch. All other

switches are analyzed and assigned DLF port roles. The defined port roles are:

• DLF_ROOT_PORT — A forwarding port on the switch that is closest to the root switch in terms of

path cost.

• DLF_DESIGNATED_PORT — A forwarding port on the switch that designated to forward the traffic

within the local area network.

• DLF_ALTERNATE_PORT — A forwarding port on the switch that is blocked using ingress port blocking

to prevent network loops. This port forwards standard unicast traffic but not BUM traffic.

DLF Topology

You can configure a DLF topology for traffic blocking. The DLF traffic travels through the head nodes and

returns to the leaf nodes if the endpoints are attached to leaf nodes. The DLF traffic uses the next hop

and the unicast traffic uses the shortest path between endpoints. The DLF blocking supports DLF and

unicast traffic requirements but does not affect unicast traffic.

Supported Technologies and Protocols