The MPLS WG Archive

[Curtis Villamizar <curtis@fictitious.org>]

In message <3E70EBE8.6E9673AF@lucent.com>, Yangguang Xu writes:
> 
> You were talking about implementation with linked list to each port to speed 
> up
> traversing. His point is, I think, the failure can happen any port/link in th
> e
> network, construct data structure according to the HE local ports may not
> helpful in some cases. 


That is not what I suggested and yes it would not be of much use.

For each link in the path, not just the first hop (port), one entry is
created.  How to create this sort of data structure is grad school
data structures 101 material.

Besides, if you only had a pointer from the immediate hop then we
wouldn't be discussing flooding because the failure is on a directly
connected link and doesn't need to be flooded.

Curtis


> Curtis Villamizar wrote:
> > 
> > In message <F69EB380D594D611BBEA00065B3F14B4A35250@MAIL>, Steve Yao writes:
> > >
> > >
> > > >Then the path-tear and/or the directed notify would still notify the
> > > >ingress.
> > > >
> > > >What circumstances are you thinking of.
> > > >
> > > >Curtis
> > > >
> > >
> > > I guess it is suffice to say that the linked list mechanism does not cove
> r
> > > all the cases.
> > >
> > > Steve
> > 
> > It wasn't intended to cover all cases and does not depricate use of
> > path-tear.  You still didn't answer my question.
> > 
> > Curtis
>