[YANG] Re: Underestimating the problem ([was something else])

Phil Shafer wrote:
> Andy Bierman writes:
>> YANG does not address the real problems preventing standard
>> data models for configuration either.
> 
> Yup, it's just another step.  My concern is that we don't
> spend seven years on this step.
> 

Some of us will vote with our feet way before then ;-)

>> But if not, then maybe registry-based-indirection mechanisms
>> would work instead.
> 
> The translation framework needs to give flexibility not only
> to the expressiveness of the device configuration, but also
> to the ability of the provider to tailor the translation to
> their needs/networks/products/customers/etc.
> 

I can remember (from my former life as an IOS SW developer),
several bold attempts to provide this sort of translation in a closed,
self-contained application -- and they all failed spectacularly.
It is way too hard for 1 apps-dev team (or person!) to keep the
database up to date or even tractable.

But, if WEB apps existed that provided an easy-to-use Wikipedia
kind of community-input approach, combined with structured
use of NETCONF capabilities, YANG data models, and some new (very TBD)
indirection mechanisms in YANG, then it might be possible
to get 80% of a router configured auto-magically, and drastically
simplify the other 20%.

> Thanks,
>  Phil
> 
> 

Andy

_______________________________________________
YANG mailing list
YANG@ietf.org
https://www1.ietf.org/mailman/listinfo/yang