Hey Burt.. did you get your head around this? I was jsut reading some old posts and thought ide give my 2 cents
basically when you order a service on our MPLS network your going to be given your own VRF..
every customer has their own VRF and the beauty of the vrf is that you can use whatever ip addressing you want and it wont affect other customers as each vrf is like your own private network.
so from a provider standpoint it is very scalable.
Typically the provider will put a managed device at your location and its called a CE.. Customer Edge..
your local access circuit will go back to the providers PE (you guessed it Provider Edge) router.
So again depending on the service but this connection will typically be running bgp. From that point the PE router will have a minimum of 2 trunks into the providers core network. the core is built up of many routers which we call P routers.. they have no customer connections and are purely labele switching routers (MPLS). They will be GSR routers at minimum and for us i saw somewhere that we are starting to put in the new CRS routers which are rediculous.
so as mentioned above the beauty of this MPLS is that it is all QOS aware and so the qos models can be implemented from end to end. which is really critical these days to ensure the voice and video data streams are given the priority queuing they require all across the network.
A real benefit is also due to the design you can achieve total meshing of your sites by default without having to build pvcs between them.
The goal for the telcos is to get all legacy frame, atm etc customers cut over to the mpls backbone.
anywayslet me know if you have any other questions maybe i can answer.. but i think i forget your original question already
