Re: [NGO] comments on CANMOD BoF

Phil Shafer wrote:
> Andy Bierman writes:
>> The NETMOD WG (if there ever is one) should deal
>> with the entire problem of standardized CM, which includes
>> secure operation in a multi-user environment.
> 
> With this approach, we'd still be working on the NETCONF draft.  We
> need to find what we agree on, build concensus on that, publish,
> gain experience, learn, and evolve.
> 
> If the plan is "all or nothing", we'll get nothing.
> 

The NETCONF access control model is "all or nothing",
not the NETCONF feature development plan.  That is in its
3rd phase, and new stuff like partial-locking and yet another
optional transport (which is only needed to avoid the mandatory
transport) are given higher priority than security.

You have to design the 2nd floor of the house, even though
you start out by building the first floor.  If you don't,
at best the project will cost 10X more expensive than it should,
and at worse, the house will collapse when you add the weight
the load-bearing walls on the first floor were never designed to handle.

A standard access control model is clearly needed.
Every NETCONF implementation has its own proprietary ACM.
The requirements for standardized access control could impact
the NETMOD architecture and the DML.  Ignoring access control
and bolting it on later might be a huge mistake.

So by 'all', I mean a coherent and well-planned execution strategy
to reach a complete standardized CM solution for NETCONF,
as opposed to an ad-hoc free-for-all that continues to produce
zero writable standard objects for NETCONF.

> Thanks,
>  Phil
> 
> 
> 

Andy

_______________________________________________
NGO mailing list
NGO@ietf.org
https://www.ietf.org/mailman/listinfo/ngo