[Lsr] Considerations for Extended TE Metrics (draft-ietf-isis-te-app)

[Alvaro Retana <[e-mail]>]

Dear authors:


When reviewing the updates to draft-ietf-ospf-te-link-attr-reuse-11, I
noticed the following difference with respect to
draft-ietf-isis-te-app-12:


[] draft-ietf-isis-te-app-12:

437	4.2.3.  Considerations for Extended TE Metrics

439	   [RFC8570] defines a number of dynamic performance metrics associated
440	   with a link.  It is conceivable that such metrics could be measured
441	   specific to traffic associated with a specific application.
442	   Therefore this document includes support for advertising these link
443	   attributes specific to a given application.  However, in practice it
444	   may well be more practical to have these metrics reflect the
445	   performance of all traffic on the link regardless of application.  In
446	   such cases, advertisements for these attributes will be associated
447	   with all of the applications utilizing that link.


[] draft-ietf-ospf-te-link-attr-reuse-11:

449	8.  Considerations for Extended TE Metrics

451	   [RFC7471] defines a number of dynamic performance metrics associated
452	   with a link.  It is conceivable that such metrics could be measured
453	   specific to traffic associated with a specific application.
454	   Therefore this document includes support for advertising these link
455	   attributes specific to a given application.  However, in practice it
456	   may well be more practical to have these metrics reflect the
457	   performance of all traffic on the link regardless of application.  In
458	   such cases, advertisements for these attributes can be associated
459	   with all of the applications utilizing that link, for example, by
460	   listing all applications in the Application Bit-Mask.



The difference is in the last sentence: the OSPF draft points at how
the user can associate the attributes with all the
applications...while the ISIS draft just says that the "attributes
will be associated with all of the applications".

I'm assuming that the ISIS operation is similar: the new sub-TLV would
have to include all the appropriate bits in the Application Bit-Mask.
Is that a correct assumption, or will ISIS behave differently?

Please clarify the text in the ISIS draft.


As I mentioned in the thread about draft-ietf-ospf-te-link-attr-reuse,
I am starting the IETF Last Call for both documents.  We can work on
this clarification during that time.


Thanks!

Alvaro.