The MPLS WG Archive

["chetan kumar s" <chetan.kumar@wipro.com>]

On Fri, 18 Oct 2002, David Charlap wrote:

> Hong Liao wrote:
> >

Hi,
	This question has arised due to the fact that RFC has both errors
mentioned in the section 4.7.4. But however many ppl follow the most
obvious error "Routing Problem". I do not understand why authors had
mentioned both the error in this section.. Well is there any
implementation that use "policy control failure"

Thanks
Chetan S

> > I have two questions for section 4.7.4.
> >
> > 1) if the three tests mentioned in this section fail, should router send a
> > patherr msg with the error code 'routing problem' or error code 'policy
> > control failure'?
> >  I think it should be the error code 'routing problem'.
>
> That's what section 4.7.4 says:
>
> 	For a link to be acceptable, all three tests MUST pass.  If
> 	the test fails, the node SHOULD send a PathErr message with
> 	an error code of "Routing Problem" and an error value of
> 	"no route available toward destination".
>
> > 2)  In the spec and this section, it was not clear that what kind of value
> > should be put in the 'link-attr', 'exclude-any',
> > 'include-any','include-any' in order to trigger the router to generate the
> > patherr msg above.  I would like to know how to set the value in this
> > field, and give me an example on how to create this scenario with what kind
> > of value in these fields?
>
> Every link has attributes configured.  Other routers learn them through
> routing.
>
> If an ingress node cares about the attributes, it will generate an ERO
> that sends the LSP over links that match the required criteria.  It will
> also set the resource affinities in the session attributes so that
> transit nodes can validate the locally-configured attributes against
> those requirements.
>
> When choosing an egress interface, a router must choose one such that
> none of the "exclude-any" bits correspond with the link's configured
> bits, at least one of the "include-any" bits corresponds with the link's
> configured bits, and all of the "include-all" bits correspond with the
> link's configured bits.
>
> Validating a link (fetched from an ERO or from routing) against an
> interface's attribute bits is easy.  Using these bits as part of the
> criterial for choosing a link may be more complicated, depending on what
> kind of routing features your box supports.
>
> -- David
>

-- 
------------
Thinking is progress, non thinking is destruction of individual, organization
and nation

-Abdul Kalam