Re: [dnsext] we need help to make names the same, was draft-yao-dnsext-identical-resolution-02 comment

[Danny Mayer <[e-mail]>]

On 2/16/2011 4:29 PM, Mark Andrews wrote:
> 
> In message <[e-mail]>, Tony Fi
> nch writes:
>> On Wed, 16 Feb 2011, John Levine wrote:
>>>
>>> It would not be absurd to argue that the most reasonable way to solve
>>> the provisioning issues is for the SMTP and HTTP servers to ask the
>>> DNS what the canonical name for an otherwise unknown name is, so those
>>> servers are just provisioned with the canonical name and an "allow
>>> variants" flag.
>>
>> It used to be the case that SMTP servers would rewrite domains in
>> addresses by replacing a CNAME owner with its target. See RFC 1123 section
>> 5.2.2. This requirement no longer exists but there is still code out there
>> that supports it. I think it would be quite reasonable to add a feature
>> for optional cname-based canonicalization to an MTA. (You can probably do
>> it now using Exim's configuration language, though it'll probably be a bit
>> ugly.)
>>
>> There is also some HTTP server code out there that hooks into the DNS for
>> server name canonicalization - see Apache's UseCanonicalName DNS option,
>> which is my fault. It uses reverse DNS lookups (it was designed for
>> IP-based virtual hosting) but I don't think it would be hard to do
>> something similar based on CNAME records.
>>
>> Note that server features like this are nice to have but not absolutely
>> necessary.
> 
> HTTP abuses CNAME.  If HTTP clients where following the design
> principles behind CNAME then the HTTP request would be re-written
> when a CNAME was seen.  Instead they ignored the CNAME and as a
> result effectively treated it like a single MX record which is wrong
> and has caused problems all along.
> 
> When we actually want to use CNAMEs for what they are designed to
> be used for we find HTTP has hijacked them.
> 
> There still isn't a formal RFC for SRV with HTTP.

Even if there were I'm not convinced that it would be useful since there
is no way on the RHS to specify the path. It can give you a name, a
port, a weight and a priority but no path. There was a proposal for a
URL RR but I cannot find it right now and I don't think the wg is
considering it, at least it's not on the document list.

Danny
_______________________________________________
dnsext mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/dnsext