Re: dhc-addr-notification: input required on the client behaviour

+1. Avoiding a small amount of extra multicast traffic in these unusual
situations (misconfiguration, rollback) doesn't seem worth complicating the
implementation and increasing the risk of implementation bugs.

If the network has really serious problems with these packets (e.g., DHCPv6
server catches fire when it receives them) then relying on implementations
not to send them doesn't seem like a good strategy anyway.

On Wed, Jan 10, 2024 at 1:42 AM Warren Kumari <[email protected]> wrote:

>
>
>
>
> On Tue, Jan 09, 2024 at 10:07 AM, Ted Lemon <[email protected]> wrote:
>
>> On Tue, Jan 9, 2024 at 4:15 AM Jen Linkova <[email protected]> wrote:
>>
>>> Would the group be OK if -08 contains Bernie's proposal?
>>>
>>> Not speaking for the group, obviously, but I think this is the right
>> choice. Anything else is an unnecessarily complicated protocol fix to a
>> configuration mistake.
>>
>
>
> Yup / +1.
> If the "failure mode" was more harmful then I'd have a different view, but
> seeing as it is a trivial amount of traffic (and in a weird corner case), I
> think that it's the right solution….
>
> W
>
>

_______________________________________________
dhcwg mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dhcwg