CNAME caching

[Kevin Darcy]

No, it really *has* to work this way, since the target of the CNAME 
could own multiple record types, not just A records. Also, under the 
robustness principle, caching resolvers should be able to do something 
sane with nested aliases, and this gets to be problematic if you 
conflate an alias record with its target...

                                                                         
                                       - Kevin

Jim McAtee wrote:

>So, what you're saying is that the resolving server would cache both records
>and then follow the chain through its own cache?  I din't realize that was how
>records were cached.  Just figured it would cache a single record with host
>name and the resolved IP address.
>
>Thanks.
>
>
>Roger Ward wrote:
>
>  
>
>>Both.
>>
>>-Roger
>>
>>-----Original Message-----
>>From: bind-users-bounce at isc.org [mailto:bind-users-bounce at isc.org] On Behalf
>>Of Jim McAtee
>>Sent: Thursday, October 02, 2003 2:38 PM
>>To: bind-users at isc.org
>>Subject: CNAME caching
>>
>>
>>When you have a CNAME pointing at an A record, what is usually used by a
>>resolving server for the TTL - that of the CNAME or that of the record to
>>which it resolves?
>>    
>>
>
>
>
>
>
>  
>