cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
1051
Views
10
Helpful
3
Replies

When are RD and RT attributes sent

I'm having a mental block with this. I understand that RD will make routes unique when you have two customers:

Cust A:

RD: 1:1

1:1:10.0.0.0/8

 

Cust B:

RD: 2:2

2:2:10.0.0.0/8

 

I understand that RT will tell the router which route to import into that VRF and can be used a way to leak routes from one VRF to another. 

 

I do not understand WHEN the router sends these attributes. Does it only send for a VPNv4 neighbor with send-community-extended? If I have configuration for ipv4 VRF CUST_A, the RD and RT attributes, will not be sent between these peers, correct?

 

In this case, I can have configuration like this:

vrf CUST_A

RD 1:1

(no route-target)

 

vrf CUST_B

RD 2:2

 

and each VRF will get their appropriate prefixes from the neighbor, as long as each vrf BGP peer IP is different.

2 Accepted Solutions

Accepted Solutions

Harold Ritter
Level 12
Level 12

In the context of MPLS VPN, the RD is prepended to the actual prefix being announced to create the VPNv4 prefix. The route target is used to control the prefix import/export process, as you mentioned. You need MP BGP VPNv4 to be configured for this to append. 

 

If you can have VRF locally configured without the BGP VPNv4 component. This would generally be referred to as VRF lite.

 

Regards, 

Harold Ritter
Sr Technical Leader
CCIE 4168 (R&S, SP)
harold@cisco.com
México móvil: +52 1 55 8312 4915
Cisco México
Paseo de la Reforma 222
Piso 19
Cuauhtémoc, Juárez
Ciudad de México, 06600
México

View solution in original post

Martin L
VIP
VIP

You will need to set up BGP to do RD-to-VPN prepending. I think this is called MP-BGP VPNv4 prefixes. I will attach sample soon but it is just another MPLS label. ISP will route those MP BGP VPNv4 based on RD labels attached.

View solution in original post

3 Replies 3

Harold Ritter
Level 12
Level 12

In the context of MPLS VPN, the RD is prepended to the actual prefix being announced to create the VPNv4 prefix. The route target is used to control the prefix import/export process, as you mentioned. You need MP BGP VPNv4 to be configured for this to append. 

 

If you can have VRF locally configured without the BGP VPNv4 component. This would generally be referred to as VRF lite.

 

Regards, 

Harold Ritter
Sr Technical Leader
CCIE 4168 (R&S, SP)
harold@cisco.com
México móvil: +52 1 55 8312 4915
Cisco México
Paseo de la Reforma 222
Piso 19
Cuauhtémoc, Juárez
Ciudad de México, 06600
México

Martin L
VIP
VIP

You will need to set up BGP to do RD-to-VPN prepending. I think this is called MP-BGP VPNv4 prefixes. I will attach sample soon but it is just another MPLS label. ISP will route those MP BGP VPNv4 based on RD labels attached.

Awesome, thank you guys! Yes, I know when the traffic is transported across the MPLS backbone, it will get to the 3 or POP and then underneath will be the VPN label, I have seen this in wireshark. I understand now that the router will ONLY send append the prefix with the RD and send the RT extended community to an address-family vpnv4 (or v6) neighbor. So, configuring RT for a VRF that will peer over address-family ipv4 vrf CUST_A is pointless.

 

Thank you both so much, I feel very confident now! :)