Hello Thomas,
the term "VC merge" is used in cell based MPLS. There ATM cell headers (VPI/VCI) are used as labels. Basically LVCs are setup through f.e. LDP. The main issue VC merge does address is the comparably low number of VCs supported per interface in ATM routers and switches. Assume 100 routers attached to an ATM network - this could mean roughly 10000 LVCs.
VC merge means, that an ATM switch has several incoming VCs (VPI/VCI) mapped to one outgoing label (VPI/VCI). Thus on the outgoing interface towards one router R1 only one VC is created and not 99 (a separate one per other router).
In cell-based MPLS you can configure this through the command "mpls atm vc-merge".
The detailed description can be found at: "MPLS over ATM: VC Merge"
http://www.cisco.com/en/US/tech/tk436/tk798/technologies_configuration_example09186a00801c2d73.shtml
In frame-based MPLS the downstream router distributes a single label per destination network. So all upstream neighbors will use the same label. This could also be called "merging of LSPs". As this is the default, there is no specific configuration required.
Hope this helps! Please rate all posts.
Regards, Martin
