Talk:Route distinguisher

Meaning of RD

As per my humble understanding the RD has no meaning. The same VPN cannot have 2 RDs. BGP does not allow one to have 2 possible paths to the same destination/prefix. So for example is RD1:IPv4 is 1 prefix and RD2:IPv4 is the other prefix pointing back to the same host, then the RD has no meaning.

Seems to me it means a lot. You said "the same VPN", but what if the VPNs are different? Then RD1:IPv4 and RD2:IPv4 can (probably have to) designate different destinations (when IPv4 is the same value). That's exactly what's being distinguished by the RD. -R. S. Shaw 21:28, 19 June 2006 (UTC)[reply]

So why not "RT" Who cares about the RD anyways? eitherways if the RD does change, then the the RD has no meaning, The RT from the importing side should be sufficient to identify things —Preceding unsigned comment added by 220.227.207.194 (talk) 08:15, 18 December 2007 (UTC)[reply]

It means nothing, just a pseudo error made by knocked up coders. if the RD can change for a VPN, then the bases of BGP is violated. or can you say that RD is the same across the VPN?

Either the RD should be the AS number, or the RD has no meaning, if the AS number exists too.

Let's check the facts first

Sorry folks, but this discussion of whether or not an RD means anything is irrelevant when it comes to an encyclopedia. RD definitely does have a specific meaning and it is definitely different from an RT (route target). The IETF has several documents explaining RDs and much more, and you can find at least one of them in the references section. May I suggest that you not just rely on what you learned in class, from your friends and from on the job training. Take the time to find and read the relevant IETF RFCs and you will understand how it all fits together.

Also note that not all networks use every feature of every IETF protocol. In my employer's network we use lots of RDs, one for each primary and each secondary path for every VPN. They send traffic to the correct VRF (tunnel entrance) on the PE router. And we use lots of RTs too, at least one primary and one secondary for each VPN bundle that a customer has. In other words customer A has 7 VPNs, A1, A2, A3, B1, B2, C1, which are in three bundles. Some sites, like data centres, have all three bundles, A, B, and C. For each bundle there is a pair of MPLS VPNs, one using a primary CE-PE pair and one using a secondary CE-PE pair. The RT identifies a bundle, and the RDs keep the individual VPNs (A1, A2, A3) in a bundle distinct from one another. Customer pay us to do this so don't ask why it is so complex. They have their reasons and the cash to back it up. —Preceding unsigned comment added by Wavetossed (talk • contribs) 17:30, 11 February 2010 (UTC)[reply]