[ietf-dkim] Are lookalike domains like parent domains?

[Steve Atkins]

On May 1, 2008, at 10:37 AM, Jim Fenton wrote:

> Steve Atkins wrote:
>> On Apr 30, 2008, at 11:40 AM, Arvel Hathcock wrote:
>>
>>
>>> So, from the receiver side, I don't have a problem with the
>>> "treewalking" - whether it stays or goes.  I don't see it's   
>>> existence as
>>> the huge problem that others do but if it disappeared tomorrow this
>>> would not destroy the essence of ADSP (it would just increase it's
>>> deployment complexity).
>>>
>>
>> It would increase deployment complexity for senders.
>>
>> It would decrease deployment complexity for receivers.
>>
>
> That's true, but you also need to consider the nature and degree of  
> deployment complexity required.
>
> Senders who want full coverage would need to either add new ADSP  
> records for all their hosts or deploy new (and, as far as I know,  
> currently nonexistent) DNS tools to automatically publish ADSP  
> records for all of their hosts.

Not at all. It's a purely internal matter. It has no impact outside  
the site of someone who has chosen to use ADSP.

It's less important than it being purely a voluntary effort for ADSP  
users, but there's also no need for them to deploy or develop new  
tools to do so. If they can publish an A record, they can publish a  
TXT record with little extra effort.

>
>
> Otherwise, eceivers would need to deploy an ADSP implementation that  
> queries the parent domain, as currently described in SSP-03.  Other  
> than that, they don't need to do anything special.

Which would add a network session for every inbound email, whether  
that email was sent by an ADSP user or not. Increasing the latency of  
inbound mail processing is a significant cost.

Cheers,
   Steve