3com 3031 Instruccion De Instalación
748
C
HAPTER
55: MPLS L2VPN C
ONFIGURATION
In addition, L2VPN service can be provided by configuring static VC labels rather
than signaling. CCC is one approach for statically configuring L2VPN.
than signaling. CCC is one approach for statically configuring L2VPN.
Introduction to CCC
Circuit Cross Connect (CCC) implements L2VPN through static configuration.
Compared with the conventional MPLS L2VPN, CCC uses one tier of label for user
data transmission. Therefore, its use of LSP is exclusive, and a user must manually
configure two LSPs for each CCC connection, with one for each direction. These
two LSPs can only be used for transmitting the data on this CCC connection and
cannot be used on other L2VPN connections. Besides, they cannot be used on
BGP/MPLS VPN or for carrying IP packets.
Compared with the conventional MPLS L2VPN, CCC uses one tier of label for user
data transmission. Therefore, its use of LSP is exclusive, and a user must manually
configure two LSPs for each CCC connection, with one for each direction. These
two LSPs can only be used for transmitting the data on this CCC connection and
cannot be used on other L2VPN connections. Besides, they cannot be used on
BGP/MPLS VPN or for carrying IP packets.
The major benefit for CCC mode is no special signaling is require to transmit L2
VPN information as long as MPLS forwarding is available. In addition, QoS is
guaranteed for the LSP is privately leased.
VPN information as long as MPLS forwarding is available. In addition, QoS is
guaranteed for the LSP is privately leased.
MPLS L2VPN in SVC mode
Static Virtual Circuit (SVC) mode is in essence a static implementation of Martini
mode. The difference is that in SVC mode the VC labels are configured manually,
while LDP is not used as the signaling to transmit L2VC and link information.
mode. The difference is that in SVC mode the VC labels are configured manually,
while LDP is not used as the signaling to transmit L2VC and link information.
MPLS L2VPN in Martini mode
Martini mode uses extended LDP (Martini draft extends LDP and adds VC FEC) as
the signaling to transmit VC information. This mode identify a VC using
VC-TYPE+VC-ID, where VC-TYPE stands for link-layer encapsulation type and
VC-ID must be unique in a PE. A PE connecting two PEs uses LDP to exchange VC
labels and binds together the corresponding CEs via the VC-ID. After established,
a VC can transmit L2 data for two CEs.
the signaling to transmit VC information. This mode identify a VC using
VC-TYPE+VC-ID, where VC-TYPE stands for link-layer encapsulation type and
VC-ID must be unique in a PE. A PE connecting two PEs uses LDP to exchange VC
labels and binds together the corresponding CEs via the VC-ID. After established,
a VC can transmit L2 data for two CEs.
Unlike CCC mode, Martini mode does not support local switching function.
Besides, it also owns no privately leased LSP, but a LSP shared by many VCs.
Besides, it also owns no privately leased LSP, but a LSP shared by many VCs.
In comparison with Kompella mode, it detects faults in a faster way, for it is not
based on regular refreshing mechanism.
based on regular refreshing mechanism.
Martini mode is more applicable to a low-density L2 connection, for example, star
connection.
connection.
MPLS L2VPN in Kompella mode
MPLS L2VPN in Kompella mode is similar to the L3 BGP/MPLS VPN. Like BGP/MPLS
VPN, the PE uses IBGP session to detect the nodes in L2VPN and transmit VPN
information. In terms of label distribution, MPLS L2VPN in Kompella mode makes
use of label block to assign labels for multiple connections at a time. The user can
specify a local CE range to indicate the number of CEs to which the current CE can
be connected. The system will assign a label block equivalent to CE range for the
CE at a time. This approach allows the user to reserve some extra labels for VPNs
for future use. While causing the waste of label resources, an enormous benefit of
this undertaking is the reduced configuration works performed in VPN
deployment and expansion. Once the size of the label blocks is determined, the
system can calculate labels required in each link with a specific algorithm and
forward packets with MPLS LSP. MPLS L2VPN also (in the same way as BGP/MPLS
VPN) differentiates VPNs with VPN-target, which enable flexible VPN networking.
VPN, the PE uses IBGP session to detect the nodes in L2VPN and transmit VPN
information. In terms of label distribution, MPLS L2VPN in Kompella mode makes
use of label block to assign labels for multiple connections at a time. The user can
specify a local CE range to indicate the number of CEs to which the current CE can
be connected. The system will assign a label block equivalent to CE range for the
CE at a time. This approach allows the user to reserve some extra labels for VPNs
for future use. While causing the waste of label resources, an enormous benefit of
this undertaking is the reduced configuration works performed in VPN
deployment and expansion. Once the size of the label blocks is determined, the
system can calculate labels required in each link with a specific algorithm and
forward packets with MPLS LSP. MPLS L2VPN also (in the same way as BGP/MPLS
VPN) differentiates VPNs with VPN-target, which enable flexible VPN networking.