Your understanding of the "leaking" effect is the only thing that is
incorrect in ANY of your statements. The route-targets are added to "LOCAL
ROUTES" that have been introduced into the VPN, not routes that have already
been imported or "learned".
This is also why default route propagation can be a real pain when
implementing Internet Access VPN's.
Everything else was very well explained, even your explanation of RD.
Paul
-- Paul Negron CCIE# 14856 CCSI# 22752 Senior Technical Instructor > From: Keller Giacomarro <keller.g_at_gmail.com> > Reply-To: Keller Giacomarro <keller.g_at_gmail.com> > Date: Sat, 17 Dec 2011 22:45:54 -0600 > To: Cisco certification <ccielab_at_groupstudy.com> > Subject: MPLS - RDs, RTs, and Import/Export > > Hi all, > > I'm having some trouble wrapping my head around how VRF RDs and RTs are > being handled by the router. > > First, some things I know (I think...): > - A VRF holds a set of routes, and is a copy of the standard routing table > - A RD is set per-vrf and its sole purpose is to create a way for BGP to > distinguish between identical prefixes from different VRFs > - Routers do not use the RD to make import/export decisions into VRFs > - An export RT on a VRF means that if the router advertises a route from > that VRF into BGP, it will attach the extended community specified in the > export statement > - An import RT on a VRF means that if the router receives a prefix from BGP > that has the specified community set, it will import it into the VRF where > the import statement is configured > > It came as a revalation to me that the RD doesn't have ANYTHING to do with > importing/exporting routes -- so much of the docs I've read seem to imply > that it does! > > Okay, so here's what I don't get. Say we're doing a central services MPLS > VPN. Two customers, one central services VPN so there are three VRFs in > play. Assuming only one PE per customer/server site to simplify a bit. > > Exports: > Customer 1's routes are exported from their PE into MP-BGP with a RT of 1 > (not valid, but this is an example). > Customer 2's routes are exported from their PE into MP-BGP with a RT of 2. > Central services's routes are exported from their PE with a RT of 3. > > Imports: > Customer 1 imports routes with a RT of 3 to receive central services routes. > Customer 2 imports routes with a RT of 3 to receive central services routes. > Central services imports routes with an RT of 1 and 2 to receive Customer 1 > and 2's routes. > > Based on the reading I've done, at this point Customer 1 and Customer 2 > should be able to communicate with Central Services but not with each > other. But that doesn't really make sense to me! Wouldn't central > services *import* routes from Customer 1 with an RT of 1 and then > *export* those > same routes with an RT of 2? Wouldn't that mean that both Customer 1 and > Customer 2 are receiving each other's routes, with the Central Services PE > acting as a route bridge? > > Any lead on what I might be missing, as well as confirmation of what I > think I know, would be very much appreciated. > > Thanks, experts! > > Keller Giacomarro > keller.g_at_gmail.com > > > Blogs and organic groups at http://www.ccie.net > > _______________________________________________________________________ > Subscription information may be found at: > http://www.groupstudy.com/list/CCIELab.html Blogs and organic groups at http://www.ccie.netReceived on Sun Dec 18 2011 - 12:39:13 ART
This archive was generated by hypermail 2.2.0 : Sun Jan 01 2012 - 08:27:00 ART