Re: [netmod] backward compatibility requirements in draft-verdt-netmod-yang-versioning-reqs-00

[Andy Bierman <[e-mail]>]

On Sat, Jul 21, 2018 at 3:12 AM, Juergen Schoenwaelder <
[e-mail]> wrote:

> On Sat, Jul 21, 2018 at 01:48:37AM -0700, Andy Bierman wrote:
> >
> > But you can tell the 2 subtrees apart this way.
> > If I change /foo from a container to a list, then how do you support both
> > implementations
> > of container /foo and list /foo at the same time?
> >
>
> Well, all of this is the consequence of moving from the current naming
> system (module, path) to (module,path,version). Once we allow
> non-backwards compatible changes, then we may have to find ways to
> support different versions of a module (i.e, during session
> establishment the client selects a version context to work with).
>
> To be clear about my involvement in the versioning design team: I am
> personally not convinced that a different versioning scheme is going
> to be simpler; certain things that are simple and robust today will
> become more complex and fragile. I decided to get involved in order to
> point out that moving to a (module,path,version) naming scheme has
> many implications since everywhere where we currently use
> (module,path) we need to think about now required version context is
> coming from. This goes far beyond YANG imports, this impacts likely
> protocols, the proposed instance document storage format, NACM rules
> may need to be interpreted in a version context etc.
>
>

Supporting multiple concurrent conflicting configuration management models
 is too complicated
and I would expect those interested in tractable solutions will look
elsewhere.


/js
>

Andy


>
> --
> Juergen Schoenwaelder           Jacobs University Bremen gGmbH
> Phone: [phone]         Campus Ring 1 | 28759 Bremen | Germany
> Fax:   [phone]         <https://www.jacobs-university.de/>
>