ietf-clear
[Top] [All Lists]

[ietf-clear] Re: Make CSV backwards compatible with legacy SPF records?

2004-12-09 07:11:51
In <41B81380(_dot_)7030600(_at_)elvey(_dot_)com> Matthew Elvey 
<matthew(_at_)elvey(_dot_)com> writes:

On 12/1/2004 11:57 PM, wayne sent forth electrons to convey:

For the purposes of HELO checking, the question is how many DNS
lookups does it take to resolve the SPF record for whatever HELO
domain you use, not elvey.com.  In most cases, this will be "v=spf1 a -all",

No it won't.  "v=spf1 a mx ptr -all" is most popular, based on stats
you provided*.

That is the most common SPF record at the domain name used in the
From: and/or the domains immediately under com/net/org.  (I've done
surveys of both.)  The HELO domain is supposed to be that of the SMTP
client, which is often (almost always?) different.

The SPF wizard at http://spf.pobox.com/wizard.html has, since last
year(?), tried to guess the HELO domains that will be used and told
you to add just the short "v=spf1 a -all" record at those locations.


To try and keep this post a little bit more on-topic for this list,
does anyone have any good data on what real-world HELO domains look
like?  Is it typically a subdomain off of the main domain name?  How
often does a single HELO domain get used by multiple IP addresses?  (I
confess I've never even bothered to double check, but I hear that
hotmail.com always uses just hotmail.com in the HELO domain and if so,
I suspect MS uses it on hundreds of different IP addresses.)  How
often does HELO domain intentionally resolve to a different IP address
than the client SMTP IP address?  How often is this accidental?

Understanding the answers to these questions would help create a
better CSV deployment guide, pointing out both common errors and
common situations that must be changed if the domain owner wants to
use CSV.  (And, of course, SPF/HELO checking.)


-wayne