SEARCH
0-9 A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
Prev | Current Page 991 | Next

Abdul Kasim, Prasanna Adhikari, Nan Chen, and Norman Finn

"Delivering Carrier Ethernet: Extending Ethernet Beyond the LAN"

It also centralises VPLS functionality in the PE in contrast
to some designs where MAC learning has been separated from MPLS signalling or
designs such as ATM LANE where dedicated servers have been used for replication.
Each PE maintains a separate MAC FIB for each VPLS instance??”known as a Virtual
Switching Instance or VSI??”where the destination for each MAC address may be either
a local attachment circuit or a pseudowire to a remote PE. Since there is only one VSI
per VPLS instance, and since MAC addresses are unique per VSI, there is typically only
one VLAN mapped to each VPLS. This is known as qualified learning and is an implementation
of ???Independent VLAN Learning??? as per 802.1Q. If multiple VLANs are
mapped to one VPLS, then the MAC addresses may not overlap between the VLANs.
This is known as unqualified learning and is an implementation of ???Shared VLAN
Learning??? as per 802.1Q.
One key decision made in the LDP VPLS design was to reuse the Ethernet pseudowire
encapsulations without modification. This was not without controversy. The benefit
of using the existing encapsulation, of course, is that no changes need to be made
to pseudowire forwarding for VPLS.


Pages:
979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003