MPLS L3 VPN END to END Routing Flow

MPLS L3 VPN, END to END Routing Flow

We will see how END to END routing process takes place in MPLS L3 VPN

To make it simple, We will classify the process:

1.    CE to PE

2.    PE to PE

3.    PE to CE

CE to PE

CE to PE a IGP is configured, Routes are properly exchanged between CE-to PE with any IGP(RIP,OSPF,EIGRP,etc) and stored in respective VRF routing table in VRF.

PE to PE

To manage the dedicated routing for each customer, we configure routing in each customer vrf and then the received routes are properly stored in respective vrf. Then as the other fields are configured in vrf (RD and route-target,etc), they are responsible to process the routes.

Before PE router advertises those routes in MPLS cloud, respective RD is pre pended on each vrf routes to make it globally unique in MPLS Cloud (MP-BGP). This time customers IPv4 routes (32 bit IPv4) are converted in VPNv4 routes (96 bit VPNv4). As the routes are converted in VPNv4 routes then RT is attached to these routes (for VPN membership identification). Then PE will propagate routes to MPLS Cloud (via MP-BGP).

This time the update is MP-BGP update which contains lot of information:

96 bit VPNv4 address, MP-BGP communities like route-target and SOO (site of origin),Labels used for VPN route forwarding, Any existing BGP attributes (AS-path,local pref, Med,etc)

Then Downstream PE router (other end PE router) will receive those routes and they will store the routes as per configured parameters.  First they will match import route target and import the routes in correct vrf. Then the RD is removed and then the IPv4 routes are stored in VRF routing table.

PE to CE

After the routes received by PE router, It will converge with other end CE site with appropriate routing same as CE to PE process. In this way end to end routing flow works in MPLS L3 VPN’s.

The reverse routing flow will be vice versa.

For any queries, Plz write comment or mail me…