Confused about Route Targets with MP-BGP / MPLS VPN

Humper

Hi guys, long time no post

I have been reading about MPLS (VPN).

My question for you is this, and I have a hard time getting my head around this.

I understand the VRF is created and applied to the interface facing the customer. Within the configuration of this VRF you specify the route distinguisher (RD). My understanding is that the RD is a way to seperate IPv4 prefixes from other customers prefixes/networks, by adding a 64 bit field to the existing IPv4 prefix making it a VPNv4 prefix. Where I get confused is with the route target (RT). I understand that when the MP-iBGP is running between the two PE routers they need a way to identify which VRF to put the VPNv4 routes into. I started reading that the RT is just an extended community, and when routes are transported from one PE router to the next, the VPNv4 prefix, RD, and RT is carried. Why is the RT significant? I understand that we need to have a way to import VPNv4 routes into the VRF, but I thought this is what the RD was for? What does the RT do that the RD doesn't? Why can't I just import and export the RD?

dtlokee

The route target is what allows you to control the import/export process within the VRF. At first you may think it would only be necessary to have a single route target for each customer and just import that rout target into the remote router VRFs, but it is common to import other route targets into the customer VRF when a centralized firewall and Internet access is supplied on the MPLS network. In this case the each PE router connected to the CE router will be exporting the customer routes using a route target and that route target will be imported into the VRF of all other PE routers that connect to the customer. In addition to that all the PE routers will import the route target that the the provider firewall is connected to into the customer VRF.

The RD is just that, something to make the 10.0.0.0/8 of customer A different from the 10.0.0.0/8 of customer B. Usinng the RT is a more flexiable method of importing and exporting.

Humper

I understand their needs to be a way to control what routes are imported/exported from a VRF.

But please explain this.

ip vrf dtlokee
rd 1:1
route-target import 1:1
route-target export 1:1

Is the RD = RT?

keenon

Humper wrote:

I understand their needs to be a way to control what routes are imported/exported from a VRF.

But please explain this.

ip vrf dtlokee
rd 1:1
route-target import 1:1
route-target export 1:1

Is the RD = RT?

your route-target is what says import/export.. based on the target id will determine which routes you accept into the vrf routing table. the RD is locally significant to the PE router

pfillips

Just thought to add a little more to it.

A route can only have one RD, but it can have multiple RTs.

Scalability is the factor that brought in RTs in picture. Using RTs lead to one route as opposed to multiple routes using RD.
RTs can be eliminated by creating more routes using more RDs.