The MPLS WG Archive

Cell Relay Retreat>MPLS WG Archive>month:2000-Jul> msg00055



[Date Prev][Date Next][Thread Prev][Thread Next]  
  [Date Index][Thread Index][Author Index][Subject Index]

draft-ietf-mpls-te-feed-00.txt

  • From: Alex Zinin <azinin@cisco.com>
  • Date: Sun, 9 Jul 2000 22:58:48 -0700
  • Organization: Cisco Systems


  Peter, et al.

  In your draft you recommend to change the contents
  of IGP's LSDB whenever the feedback from the signaling
  component is received. I'm afraid it is not quite good.

  1. If the checksum of the LSA/LSP is not recalculated
     after the contents were changed, the IGP may report
     a fatal error if the checksum is verified before
     the next version of the LSA/LSP arrives.

  2. Recalculating the checksum is also wrong, because:

      a) only the originating router must do this

      b) if this is done by other routers it may affect
         LSA/LSP version comparison algo.

  Because of the above and because it is a 'bad thing' to
  modify link-state PDUs originated by others, I think it
  is safer and makes some more sense to say that implementations
  must account for a method to link the information provided
  by the signaling component to the corresponding elements
  of the TE database, instead of overwriting it.

  Cheers,
  
-- 
Alex Zinin