Why "bgp redistribute-internal" not required in MPLS L3VPN

Guys,

Anyone knows why "bgp redistribute-internal" is not required for BGP VPNv4?

Because they don't have this documented that "even without bgp redistribute-internal", iBGP VPNv4 routes get redistributes into IGP

 

http://www.cisco.com/en/US/docs/ios-xml/ios/iproute_bgp/command/bgp-a1.html#wp4270480859

Thanks,

Amit

Comments

  • Hi Amit

    Very good question.  I too would like to know purpose of command "bgp redistribute-internal" within "address-family vpnv4 unicast". 

  • There are several changes to BGP when it comes to using it in BGP VPNs. Not only the one that you mention but another "modification" to the behavior is how next hops are handled. In normal BGP (AFI 1 SAFI1) when a bgp update come from an eBGP neighbor and then gets passed to an iBGP neighbor, the next hop is not changed. However, in a BGP VPN environment this is not the case. The next hop get changed to the advertising iBGP peer's peering address. 

    You may also want to think of why this command was ever needed in the 1st place. This bgp behavior (not redistributing internal) exists because in normal designs some time ago, redistributing from iBGP into an IGP was something you'd likely stay away from doing to prevent issues (such as killing your IGP among others). However, a VPN environment totally relies on doing this very same thing (redistributing iBGP into an IGP). The circumstances are much different. 

     

Sign In or Register to comment.