First and foremost, domain tagging is a mechanism designed for legacy OSPF implementations that did not support the down bit [1] for type 5 LSAs. It has been established that route tagging is not to be implemented in OSPFv3[2] because v3 specifies the mechanism of operation in the down bit.
Essentially, when a prefix is received from a BGP speaker and redistributed into the OSPF running as the PE-CE protocol, the OSPF process for the VRF is given a domain tag. In the event that single customer site does attempt to readvertise the prefix to another PE and the other PE uses the same domain tag, the route tag will be matched and the prefix will not be accepted by the 2nd PE for redistribution into the BGP. Refer to [1] for specifics from the spec. Now, domain-id is NOT the same as domain-tag. Domain-ID serves a completely different purpose and is better explained in RFC 4577 section 4.2.4. Essentially, Domain-ID affects the type of LSA originated (whether type 3 or type 5/7). HTH -- William McCall [1] http://tools.ietf.org/html/rfc4577#section-4.2.5.2 [2] http://www.ietf.org/mail-archive/web/ospf/current/msg05751.html On Fri, Aug 13, 2010 at 9:28 PM, Jeferson Guardia <[email protected]> wrote: > Thanks a lot man, anyone else have anything else to add? I also agree with > him that is a mechanism to prevent route loops when redistributing from ospf > to bgp and vice versa.. but how exactly does it work? could anyone give an > example? I have read that link for a few minutes and could not clearly > understand the concept behind it. > > Thanks!! > > 2010/8/13 dongya zhang <[email protected]> > >> Try to read this link: >> >> http://www.juniper.net/techpubs/software/erx/erx50x/swconfig-routing-vol2/html/bgp-mpls-vpns-config5.html >> >> IMHO,Domain id and VPN route tag is used to prevent route loop. >> >> 2010/8/14 Jeferson Guardia <[email protected]> >> >>> Guys, >>> >>> I am studying MPLS vpn stuff and it came across some new stuff of mpls, >>> but >>> this one is not quite clear for me yet, the domain-tag. I had an >>> idea that somehow is used to prevent redistributed routes that originated >>> from the MPLS backbone from returning via some other location on the MPLS >>> backbone. >>> >>> Am I right? anything else to add or any other concept I should be aware >>> of? >>> >>> Best regards, >>> _______________________________________________ >>> cisco-nsp mailing list [email protected] >>> https://puck.nether.net/mailman/listinfo/cisco-nsp >>> archive at http://puck.nether.net/pipermail/cisco-nsp/ >>> >> >> > _______________________________________________ > cisco-nsp mailing list [email protected] > https://puck.nether.net/mailman/listinfo/cisco-nsp > archive at http://puck.nether.net/pipermail/cisco-nsp/ > _______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
