Re: [netmod] draft-bjorklund-netmod-structural-mount: Namespace issues

[Ladislav Lhotka <[e-mail]>]

Martin Bjorklund <[e-mail]> writes:

> Ladislav Lhotka <[e-mail]> wrote:
>> Hi Anton,
>> 
>> Anton Tkáčik <[e-mail]> writes:
>> 
>> > Hi,
>> >
>> > if I understand correctly netmod-structural-mount inlines "mounted" data directly to container / list which is used,
>> >
>> > which brings up following issues:
>> >
>> >
>> > 1. It is possible to have identifier conflict between data from model
>> > defining mount and mounted data (if mounted schema
>> 
>> I think we need to eliminate all kinds of recursive mount, so this
>> should not happen. 
>> 
>> >
>> > contains same model).
>> >
>> > e.g.
>> >
>> >       module mount-under-mount {
>> >
>> >              container mounts {
>> >
>> >                    list mount {
>> >
>> >                        key name;
>> >
>> >                        leaf name;
>> >
>> >                               container mounts {
>> >
>> >                                     // List of discovered remote mounts
>> >
>> >                               }
>> >
>> >                               mnt:mount-point data;
>> >
>> >                    }
>> >
>> >              }?
>> >
>> >       }
>> >
>> >
>> > 2. Expanding data directly under container / list may be problematic
>> > for clients which do not support netmod-structural-mount.
>> 
>> I don't think it can work if the client doesn't support the mount mechanism.
>> 
>> >
>> >
>> > I believe both can be solved elegantly by requiring mount-point
>> > extension to be used inside anydata element, which signifies
>> 
>> This has been already discussed in the mailing list. IMO, the biggest
>> problem of anydata (as it is currently defined, at least) is that it
>> permits really anything, i.e. not only stuff contributed by the mounted
>> modules.
>
> The fact that anydata can represent "really anything" does not mean
> that every server MUST allow the client to create "really anything"
> for all anydata config nodes.  It will depend on the semantics of each
> particular anydata node.

Schema validation is effectively disabled for anydata nodes. This may be
a problem especially for implementations that perform validation as a
separate step, perhaps automatically from the data model.

And with schema validation out of the way, it is easier to exploit
potential bugs in the server.

>
> I think that using mount-point as a substatement to anydata is in fact
> the only really safe option to mount.  Both the proposed solutions in

Well, if safe means (partial) compatibility of old clients - and I am
not even convinced about this. Otherwise it is IMO less safe because the
schema is loosened.

> the current structural mount and ysdl drafts are unsafe for clients
> that don't understand the mount.

True.

Lada

>
>
> /martin

-- 
Ladislav Lhotka, CZ.NIC Labs
PGP Key ID: E74E8C0C