You may have heard the term “MPLS”, but not know much about it. If so, it’s time to learn more!

Imagine that you’re a WAN service provider, and that you make your money by leasing bandwidth to customers. A customer’s goal is to reliably connect its sites together so that its applications run correctly, while spending as little as possible on WAN services. Your goal is to make as much profit as possible while keeping your customers reasonably happy. To illustrate the concepts, we’ll need an example topology.

As you can see, the WAN provider has a Customer “A” with six sites that must be connected together. For good performance of the customer’s VoIP application, a full-mesh topology is required. One possible solution would be for the provider to run an actual physical wire, fiber or wireless link (terrestrial or satellite microwave) from one customer site to another, as shown

As you can imagine, when dealing with many sites, running a full-mesh of physical links from one site to another would be cost-prohibitive. To solve this problem, years ago WAN providers devised another solution:  Let’s build a fault-tolerant physical topology within the provider’s cloud using Layer-2 switches

We’ll assume that the customer is running a LAN protocol (typically Ethernet) at their sites, and uses an edge device to convert between the LAN protocol and WAN framing. The device that does this (typically a router) is referred to as a “CE” (Customer Edge) device.”

A CE device connects a customer site to a provider’s POP (Point of Presence). The POPs are distributed throughout the provider’s service area. Located at each POP will be one or more “PE” (Provider Edge) devices, which connect the provider to the CE devices at the customer sites. The PE devices also connect to “P” (Provider) devices, which form the core of the provider’s network.

Next time, we’ll look at this solution in more detail.

Author: Al Friebe