For those operators who need absolute control of customer traffic from source-to-destination, a consistent protocol is needed across the network. In these cases, the most widely chosen architecture is MPLS across the network infrastructure.
The ubiquitous use of MPLS allows network operators to provision bandwidth profiles and prioritization of traffic in all segments of the network. Coupled with the advanced traffic engineering in MPLS, hard QoS parameters can be applied to customer traffic; SLA’s can be guaranteed across the network.
The drawbacks to this design are simple – cost and complexity. Requiring MPLS enabled switch/routers at all end-points (edge locations) can increase network cost significantly. MPLS enabled edge devices are frequently double the price of their simpler Ethernet cousins.
In addition, the configuration model becomes more complex. Provisioning
pathways (trunks) solely across the core involves only a limited number of
network elements and variables. Provisioning pathways from edge-to-edge of
the network is much more complex – the number of end points in a network
greatly increases the provisioning complexity.
Tools to manage MPLS networks are becoming more sophisticated and easier
to use. Additionally, familiarity with the technology allows operators
the comfort to assume more complexity in their design. As MPLS grows
in acceptance and ease of use, more operators are seeing the advantage of
a complete MPLS network architecture.
