Re: SPF isn't going to change, was Deprecating SPF

>I'm not sure I fully understand the latency argument here.
>Wouldn't you just fire off the SPF-99 and TXT-16 queries
>asynchronously? Waiting for two answers shouldn't be much longer
>than waiting for one answer when done in parallel.

If your DNS library permits, sure, you can do them in parallel.  But
experience has shown that due to cruddy middleboxes or the like, type
99 queries get lost and time out more than you might think.  If the
type 16 query doesn't give you a usable SPF record, you have to wait
for the type 99 query.  People who run large mail servers tell me the
slowdown is noticable.


>And long term, wouldn't there be benefits from not having to sift
>through any other TXT records you receive for the ones that look
>like SPF records?

In a word, no.  The SPF rules for handling type 16 and type 99 records
were identical: there has to be exactly one record that starts
"v=spf1".  You have to look at all the records to be sure that there
aren't multiple v=spf1 records (which invalidates all of them) and
once you're doing that, ignoring the ones that don't have the right
prefix is about three lines of code in a multi thousand line library.

R's,
John
_______________________________________________
dnsext mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsext