Route Distingusiher - Mystery?

[KATTUBAVA E.S]

Hi Experts?

In MPLS L3VPN, VRF in PE routers to distinguish routes from different custmers with same ip subnets and RD is used to distinguish routes from different custmers with same ip subnets on Core (P) routers.

My doubt is why RD needed in MPLS core as MPLS P routers wont have any customer routes installed in their routing table as BGP is not running in MPLS core. All it does is swapping transport labels from ingress PE to Egress PE to reach the BGP Next-hop.

So, what is the use case of RD in MPLS L3VPN core??

Thanks
-KB

[manvsing]

Hi Kattubava,

RD is used to distinguish overlapping subnets in SP environment and it is PE(provider edge routers where a customer connects to) where you define RD value(not on P routers).

Now RD converts an ipv4 prefix to VPNV4 prefix and these prefixes are exchanged by PE routers only. PE routers can either use full mesh BGP or can use route reflector to exchange these vpnv4 prefixs.

P routers only exchange routes(via IGP) and labels(via LDP) for BGP next hop ip associated with every vpnv4 prefix

[KATTUBAVA E.S]

Thanks Manvsing for your quick response.

\\RD is used to distinguish overlapping subnets in SP environment and it is PE(provider edge routers where a customer connects to) where you define RD value(not on P routers).\\

Why do we need RD here to achieve this, on PE router VRF can help us to do this. What is the exact use case of RD in MPLS L3VPN, can you explain me with a scenario??

thanks

KB

[manvsing · Accepted solution]

                                                route reflector

       route target 1:1(both)                 |                                                    route target 1:1(both)

   10.1.1.0/24                                     |

Cust1(a) -------------> PE1--------------------->P--------------------PE2------------------>Cust1(b)

                                   ^                                  \                          

    10.1.1.0/24             |                                     \                         

Cust2(a)----------------- |                                         PE3             ---------------------> Cust2(b)

route target 2:2(both)                                                                                route target 2:2(both)

>> PE1 and PE2 have BGP peering with route reflector.

Here you have two different customers which are using 10.1.1.0/24. At PE1 VRF tables associated with respective customers will keep both network separate.

Lets say there is no RD associated with them. when these prefixes reach route reflector then RR will choose one of them best and forward that prefix to its clients. 

if route with route target 1:1 is chosen as best then only PE2 will be able to add install this prefix into respective vrf for CUST1 .

Now what about CUST2 connected with PE3 ?

since PE3 have no site associated with CUST1, it will not receive 10.1.1.0/24 with rt 1:1 into its BGP table. subsequently connectivity between different site for CUST2 will not be there.

Had we associated unique RD with prefixes from both customers, RR would have treated both vpnv4 prefixes as unique and forwarded both prefixes to its clients.

[Nagendra Kumar Nainar · Accepted solution]

Hi,

CE1

 |

iPE----P----ePE

|

CE2

Assume iPE is connecting CE1 and CE2 which are in same address range 10.1.1.x/24. On iPE you will have VRF to differentiate. When iPE advertises those prefixes to ePE, we need a way to differentiate those two prefixes in BGP as it will appear as same prefix. In order to make it unique, we will append 64 bit RD to 32 IPv4 prefix to make it as unique 96 bits VPNv4 address. Since RD will be unique for each VRF customer, 10.1.1.0/24 from CE1 and CE2 will be now be unique VPNv4 addresses.

HTH,

Nagendra