The MPLS WG Archive

["Brennen, Jack" <John.Brennen@marconi.com>]

Your reading of the draft appears to be correct -- the draft
seems to allow that although it is not really a useful behavior.

It seems that any of the following EROs could be sent from
B1 to B2 and still maintain correct end-to-end operation:

              A--B1--B2--B3--C

   {B,C}
   {B2,B,C}
   {B2,B3,B,C}
   {B2,B3,C}
   {Bx,C}        -- where Bx is any subset of node B containing
                    both B2 and B3.

This is not a complete list -- the draft leaves a great deal of
room for variations on these themes.

All of these EROs seem to be permitted by the draft, primarily
by the third paragraph of 4.3.4.2:

   Otherwise, if the node is a member of the abstract node for the first
   subobject, a series of subobjects MAY be inserted before the first
   subobject or MAY replace the first subobject.  Each subobject in this
   series MUST denote an abstract node that is a subset of the current
   abstract node.

My personal preference as an implementor would be to forward the
incoming ERO unmodified -- send the ERO {B,C} to node B2.

But certainly nothing useful can come from sending the
ERO {B2,C} to node B2.

> -----Original Message-----
> From: Michael H. O'Meara [mailto:mmeara@mars.iol.unh.edu]
> Sent: Thursday, January 04, 2001 4:38 PM
> To: Eric Gray
> Cc: John Sparr; Bill Sanford; MPLS Mailing List
> Subject: Re: ERO and strict/loose subobject
> 
> 
> Hello all,
>      While there's a thread on the subject, I was hoping that someone
> might help clarify some of my confusion in regards to the 
> processing of an
> ERO. In particular, I'm having some difficulty understanding 
> the selection
> of a next hop for a node in the interior of an abstract node. 
> According to
> draft-ietf-mpls-rsvp-lsp-tunnel-07.txt, section 4.3.4.1:
> 
> 5) Interior of the Abstract Node Case: Otherwise, the node selects a
>    next hop within the abstract node of the first subobject (which the
>    node belongs to) that is along the path to the abstract node of the
>    second subobject (which is the next abstract node).  
> ...
> 6) Finally, the node replaces the first subobject with any subobject
>    that denotes an abstract node containing the next hop.  This is
>    necessary so that when the explicit route is received by the next
>    hop, it will be accepted.
> 
>     What, exactly, is meant by "any subobject that denotes an abstract
> node containing the next hop?" For example, given the 
> following topology:
> 
>     A--B1--B2--B3--C
> 
>     Say that A sends an ERO={B,C}, where B describes an 
> abstract node that
> contains B1, B2, and B3, and B and C are strict hops. According to the
> steps for making the selection of a next hop, B1 could send 
> an ERO={B2,C}
> to B2.
> Since C is listed as a strict hop, B2 would return a 
> "Bad strict
> node error." I may be missing something, but it would seem 
> that nothing in
> the steps for the selection of a next hop, nor in section 
> 4.3.4.2 on the
> addition of subobjects to the ERO, would preclude this from 
> happening. I
> would certainly appreciate any help in clarifying this matter.
> 			Thanks in advance,
> 				Mike    
> 
> ***************************
> Michael O'Meara
> Technician
> MPLS Consortium	
> InterOperability Lab
> University of New Hampshire
> ***************************
> 
> 
>