Home > Articles > Cisco Network Technology > General Networking > Traffic Engineering with MPLS

Traffic Engineering with MPLS

Chapter Description

This sample chapter offers a brief look at the DiffServ architecture and its interaction with IP and MPLS packets, the MQC, and DS-TE.

DiffServ and MPLS Packets

One thing you might have noticed is that there are 6 DSCP bits and only 3 EXP bits. Because only eight possible EXP values and 64 possible DSCP values (21 of which are currently defined) exist, how do you offer DSCP services over MPLS?

On a frame-mode network (as opposed to a cell-mode network), you are stuck with the 3 EXP bits; you need to map multiple DSCP classes to these EXP bits. However, this has operationally not yet proven to be an issue in production networks, because hardly any QoS deployments offer services that can't be provisioned with the 3 MPLS EXP bits.

Work is being done to define something called L-LSPs (Label-Only Inferred PSC LSPs) that will help alleviate this problem. The basic idea behind L-LSPs is that you use both the EXP bits and the label to define different service classes. This is actually how label-controlled ATM MPLS mode with multi-VC works. However, this book doesn't cover cell mode, because, as of this writing, there is no MPLS TE cell mode implementation available.

L-LSPs for frame-mode MPLS aren't discussed in this chapter because they're not yet implemented or even fully defined. See RFC 3270, "Multiprotocol Label Switching (MPLS) Support of Differentiated Services," for more details on how L-LSPs work.

5. Label Stack Treatment | Next Section Previous Section