Re: [dnssd] Setting device friendly names

[Ted Lemon <[e-mail]>]

Hm.   That's not what I'm saying.  The point Barbara made that I agree with
and restated was that the device should report the same name no matter who
asks or how.   IOW, we can't just change its name in the homenet naming
infrastructure: it needs to know and accept the new name.

In some cases, we don't want to trust the network infrastructure to set our
name--e.g., it used to bug the hell out of me when my Mac would take its
name from the PTR record and wind up with something nonsensical;
fortunately my Chromebook doesn't do this, nor does my Android phone.

For these cases, the only alternative we have at present is a UI on the
device, or on a web page provided by the device.   It might be nice to have
a trustworthy mechanism for doing this, but we don't have a security
architecture that would support that.

However, for other devices, being able to assign the name using the network
infrastructure is perfectly fine.   I think the main distinction is
"infrastructure/IoT devices" versus personal devices that roam with the
user.


On Thu, Jul 21, 2016 at 5:34 PM, Stuart Cheshire <[e-mail]> wrote:

> On 21 Jul 2016, at 06:22, STARK, BARBARA H <[e-mail]> wrote:
>
> > +1
> > Thx. You said that so much better than I did. :)
> > And even where network configuration is desirable, a one-size-fits all
> ubiquitous solution is not the right answer. We should not design an
> architecture with such a dependency.
> > Barbara
>
> Sorry, I’m getting confused by this discussion.
>
> The discussion in the meeting was that pretty much all network devices
> (apart from my SMA solar inverters) can have their name set today over the
> network, but not using a one-size-fits-all ubiquitous solution. It’s ad
> hoc. Many devices have their name set via web UI. Some devices have their
> name set via some proprietary app. Others have their name set using some
> non-web protocol like Apple’s HomeKit. Some have their name set via
> physical controls on the device itself. The assertion in the meeting was
> that this is no good and we need a one-size-fits-all ubiquitous solution.
> UPnP was offered as that one-size-fits-all ubiquitous solution. Now we’re
> saying that a one-size-fits-all ubiquitous solution is *not* the right
> answer.
>
> I’m confused. It sounds like we’re back to saying what we have already
> today is just fine (apart from my SMA solar inverters -- I need to find out
> how to file a bug report for that).
>
> Stuart Cheshire
>
>