Re: [dnssd] SRP: support of SRP message with zero service?

Also bear in mind that SRP isn’t just for the constrained device use case.
There may be cases where it’s useful to register a well known name and
unnecessary to advertise a service. E.g. so you can ssh to your laptop by
name.

Op wo 7 dec. 2022 om 07:52 schreef Esko Dijk <esko.dijk@iotconsultancy.nl>

> Yes we have, the link to the start of the thread is:
> https://mailarchive.ietf.org/arch/msg/dnssd/01voO6MkMBBvxeInncR6e-3u1bM/
>
> (somewhat long discussion)
>
>
>
> One use case mentioned also is an SRP client adding or removing an IP
> address in its registration, while keeping its services unchanged.
>
>
>
> Esko
>
>
>
> *From:* Kangping Dong <wgtdkp@google.com>
> *Sent:* Wednesday, December 7, 2022 12:48
> *To:* Esko Dijk <esko.dijk@iotconsultancy.nl>
> *Cc:* Ted Lemon <mellon@fugue.com>; Abtin Keshavarzian <abtink@google.com>;
> DNSSD <dnssd@ietf.org>; Jonathan Hui <jonhui@google.com>
> *Subject:* Re: SRP: support of SRP message with zero service?
>
>
>
> > Why is it useful to have a hostname and address generally?  :)
>
>
>
> If the client knows only the hostname and address, it doesn't know what
> protocol should be used to communicate with the device, so it doesn't seem
> to be useful to
>
> know only the address. One use case in my mind is that a device may
> register a well-known hostname, but didn't find a real case in local
> constraint networks...
>
>
>
> > I recall that we discussed that a device could temporarily not have a
> particular service; and some expectation that in due time a service will
> become available again.
>
> > During this time the device wants to keep its name registered, to show
> that it is still there and functional, and defend its name against
> potential other devices taking that name.
>
> *> *So one use case is deleting its service(s) but keeping the name
> registered.
>
>
>
> Will this encourage devices to always register their hostname even when
> they don't provide any service? Do we have a link to that discussion? @Esko
> Dijk <esko.dijk@iotconsultancy.nl>
>
>
>
> BRs,
>
> Kangping
>
>
>
>
>
>
>
> On Wed, Dec 7, 2022 at 7:32 PM Esko Dijk <esko.dijk@iotconsultancy.nl>
> wrote:
>
> I recall that we discussed that a device could temporarily not have a
> particular service; and some expectation that in due time a service will
> become available again.
>
> During this time the device wants to keep its name registered, to show
> that it is still there and functional, and defend its name against
> potential other devices taking that name.
>
>
>
> So one use case is deleting its service(s) but keeping the name registered.
>
>
>
> Regards
>
> Esko
>
>
>
> *From:* Ted Lemon <mellon@fugue.com>
> *Sent:* Wednesday, December 7, 2022 12:15
> *To:* Kangping Dong <wgtdkp@google.com>
> *Cc:* Abtin Keshavarzian <abtink@google.com>; DNSSD <dnssd@ietf.org>;
> Esko Dijk <esko.dijk@iotconsultancy.nl>; Jonathan Hui <jonhui@google.com>
> *Subject:* Re: SRP: support of SRP message with zero service?
>
>
>
> Why is it useful to have a hostname and address generally?  :)
>
>
>
> Op wo 7 dec. 2022 om 06:09 schreef Kangping Dong <wgtdkp@google.com>
>
> Hi,
>
>
>
> Regarding SRP registration, it's unclear if at least one "Service
> Description Instruction" is required in a SRP message when the host is not
> being deleted.
>
> If it's not required to register at least one service, what's the use case
> of registering only the host/address? Any examples? (It doesn't seem to be
> useful
>
> to advertise only the host name and addresses with the Advertising Proxy :)
>
>
>
> Thanks!
>
>
>
>
>
>
>
> BRs,
>
> Kangping
>
> _______________________________________________
> dnssd mailing list
> dnssd@ietf.org
> https://www.ietf.org/mailman/listinfo/dnssd
>