Configuring the ECN330-switch
3171553-KDU 137 365 Uen D 2006-06-16
VLAN frame received. The ingress and egress PE switches must have
the same MTU value. If single VLAN tagged packets are supported, the
MTU should be 1544 bytes when jumbo frame is not enabled.
• L2MPLS only supports VLAN packets that conform to IEEE 802.1Q
standard frames.
• The native VLAN of a trunk cannot be an L2MPLS VLAN, otherwise
untagged packets would be forwarded onto the L2MPLS VLAN.
• QinQ and L2MPLS mode cannot be supported at the same time.
• The tunnel label is based on the uplink port. Each GE port has a register
to store the tunnel label. All outgoing MPLS packets will use this register
value to construct an MPLS header. It is not VLAN or VC based. So only
one uplink port can be connected with one remote PE device.
• No VC redundancy is supported.
• L2MPLS link redundancy can be supported only if the entire network
supports STP or EAPS.
• A VC can be associated with multiple VLANs. But a VLAN can only be
associated with one input VC and one output VC.
• MPLS packets with Control Word cannot be received correctly. See the
following section.
• If a VC is associated with the VLAN (see “Binding a VLAN to an MPLS
Tunnel” on page 319), then the same VC cannot be configured as a
port-based VC (see “Configuring an MPLS Uplink Port” on page 320),
and vice versa.
• If the VC label is port-based (see “Configuring an MPLS Uplink Port” on
page 320), the PVID (port’s native VID) is 0. When an untagged packet
is received at the ingress port, the PVID is used to build a VLAN tag for
this packet. This PVID is then used to search the VC table. If an entry is
found, the ECN330-switch adds a VC label and tunnel label, and sends
the packet out the uplink port.
When an untagged MPLS packet (has MPLS header but no VLAN tag)
is received at the MPLS uplink port, the VID of the VC entry is used to
build a VLAN tag for this packet instead of the PVID. So if the VC is port-
based, incoming untagged MPLS packets with this VC are dropped.
To Next Page
Loading...
