Hi Juergen, New draft is posted.
https://tools.ietf.org/html/draft-asechoud-netmod-diffserv-model-02 Let us know of any further comments. Regards, Aseem On 5/19/15, 3:49 PM, "Aseem Choudhary (asechoud)" <[email protected]> wrote: > > >On 5/19/15, 1:39 PM, "Juergen Schoenwaelder" ><[email protected]> wrote: > >>On Tue, May 19, 2015 at 07:57:26PM +0000, Aseem Choudhary (asechoud) >>wrote: >>> Hi Juergen, >>> >>> Please find the replies inline. >>> >>> -thanks, >>> Aseem >>> >>> On 5/18/15, 10:20 AM, "Juergen Schoenwaelder" >>> <[email protected]> wrote: >>> >>> >Hi, >>> > >>> >what is the purpose of the feature policy-template-support in >>> >draft-asechoud-netmod-diffserv-model-01.txt? Can I use the model >>> >without this feature? >>> > >>> >I also note that no implementable object is left in >>> >ietf-diffserv-policy if I do not support the feature >>> >policy-template-support. Is this useful? >>> >>> [AC] ietf-diffserv-policy module defines policy as a template. The >>>other >>> option is to configure & apply it directly under target as defined in >>> ietf-diffserv-target module. >> >>OK, but why is there a feature since this 'policy' is an extension >>defined in a separate YANG model? > >[AC] OK, I see the point you are making, but there are some changes >regarding it in next version which will alleviate your concern. >> >>> >Is it a bug or a feature to support inline classifier definitions? Is >>> >it correct that actions are always 'inline' so to say? Is this a bug >>> >or a feature? >>> >>> [AC] Classifier can be defined inline or can be defined as a separate >>> template and referred in the policy definition. >> >>Yes, but is this a bug or a feature? From an implementation >>perspective, this sounds like extra cost. So what is the benefit of >>having both options? > >[AC] There are pros and cons of defining either way: When defined as a >separate template, it becomes reusable objects. Any change in the object >updates the referencing objects too. > While when defined inline, it is local change only. Any modification >has to be done on every instance. > There are use-cases where users may use template and others where they >define inline. There are vendors which support either/or/both. Also, >Inline is defined as a feature. > >> >>> All the actions are defined inline. This is the most common use-case. >> >>Hm. Why are classifiers and actions treated differently? Perhaps some >>discussion of the design decisions somewhere in the document would be >>useful. > >[AC] Actions are typically not reusable objects. Like metering rate or >RED thresholds are specific values vs classification of voice, video or >data. > I will take this point of adding to the document. >> >>/js >> >>-- >>Juergen Schoenwaelder Jacobs University Bremen gGmbH >>Phone: +49 421 200 3587 Campus Ring 1 | 28759 Bremen | Germany >>Fax: +49 421 200 3103 <http://www.jacobs-university.de/> > >_______________________________________________ >netmod mailing list >[email protected] >https://www.ietf.org/mailman/listinfo/netmod _______________________________________________ netmod mailing list [email protected] https://www.ietf.org/mailman/listinfo/netmod
