Multi-Protocol Label Switching (MPLS) transport is a funny thing -- what it means depends on who you are talking...
to at the time. If you are talking to an engineer who is responsible for designing and developing MPLS services for a carrier, he or she will more than likely discuss MPLS in terms of MPLS backbone transport.
MPLS backbone transport is analogous to both frame relay and ATM WAN circuits, in that MPLS, frame and ATM all use the concept of virtual circuits. Frame relay uses permanent virtual circuits (PVCs) between the WAN routers, ATM uses VPI/VCIs, and MPLS uses label-switched paths (LSPs).
So, understanding a little about how the carrier transport has changed, it only makes sense now to discuss how this affects the access (transport) links that connect to the MPLS backbone. The access circuit now becomes the transport between your sites and the carrier's MPLS router sitting on the MPLS backbone.
Interestingly enough, you can (in theory) connect your sites with multiple types for WAN access circuits because a router sits between the access circuits and the MPLS backbone. For example, let's say you have a three-site WAN where each site has a local circuit option to the MPLS cloud. You could provision one circuit as ATM, another as frame relay and another as Ethernet, and each of the sites could talk to the others over the carrier's MPLS core using IP. Legacy WANs required like interfaces because the virtual circuits were built from customer router to customer router, not from customer router to provider router. You could do this, but it is not recommended. The point is that the transport options remain the same for legacy WAN and MPLS WANs, but one is carried over routers (MPLS) and the other over frame and ATM backbones.
Choosing your MPLS transport
Why is the choice of MPLS transport important? As mentioned above, you can use ATM, frame and Ethernet, as well as private line, Broadband DSL and others. The choice of one or the other depends on multiple factors, and an analysis is required to determine the best fit.
- First of all, it makes sense -- if possible -- to utilize the same WAN technology you use today. This may not be possible, but it should be the first criterion analyzed. Using the same transport alleviates problems such as having to change interfaces on routers and avoids migration from one technology to another for local access. If the routers do not support the interface type, complete router upgrades may be required, at significant cost.
- A second factor is bandwidth. The fact that a customer is migrating to MPLS makes feasible the converging of voice, video and data onto a single WAN backbone. Voice and video will increase the bandwidth requirements significantly. It is a good idea to provision the right size of access transport or to use a transport with flexible upgrade paths to higher bandwidth.
- Another factor is carrier reach. ATM and frame are widely available almost anywhere in the world, and the costs are well known. With MPLS, you can get the ATM and frame local loops, but the MPLS PoP that these terminate on may be hundreds of miles away from your location. Ethernet may not be available in some areas. The costs of backhauling one access over another should be considered. Most carriers recommend one access only, although some may support multiple access types for customers.
About the author:
Robbie Harrell (CCIE#3873) is the National Practice Lead for Advanced Infrastructure Solutions for SBC Communications. He has more than 10 years of experience providing strategic, business and technical consulting services. Robbie lives in Atlanta and is a graduate of Clemson University. His background includes positions as a principal architect at International Network Services, Lucent, Frontway and Callisma.