The MPLS WG Archive

[Curtis Villamizar <curtis@fictitious.org>]

In message <F69EB380D594D611BBEA00065B3F14B4A35254@MAIL>, Steve Yao writes:
> --=_IS_MIME_Boundary
> Content-Type: text/plain;
> 	charset="iso-8859-1"
> 
> >
> >Which comes back as as a RESV with an RRO with the loose hops replaced
> >by the actual hops.  The ERO can then be changed to pin the path, but
> >if not, the RRO should get updated if the path changes.
> >
> >Curtis
> >
> 
> The RRO received at the HE only contains the IP addresses that map to half
> of the LSAs or LSPs that the LSP traversed. 
> 
> Steve


I'm not sure how you reach that conclusion.  There is both addr and
neighbor in the IGP TE extensions for numbered interfaces.  The RRO
hops match the neighbor side.  That gives the exact IGP LSAs or LSPDU
TLVs and gives the near side (the one likely to arrive first).

When a link goes down, one LSA or LSPDU fragment is originated by two
LSR, one on each side of the link.  When either LSA/LSPDU arrives, the
TE-LSDB unidirectional adjacency (or link) in both directions is
considered down.  The two unidirectional data structures each have a
set of MPLS LSP data structures linked to them (actually other small
structures that point to the LSP, but that's some of the data
structures 101 material).

If you mean that the LSA/LSPDU for the type-2/pseudonode of a
broadcast interface is not explicitly present, then quite frankly,
knowing the node of the previous hop and the next interface this is
not an incredibly hard problem to solve either.

If that's not what you meant, please let me know.

Curtis