To:
"Larson, Matt" <mlarson@verisign.com>
Cc:
dnsop@cafax.se
From:
Peter Koch <pk@TechFak.Uni-Bielefeld.DE>
Date:
Thu, 05 Apr 2001 19:28:15 +0200
In-reply-to:
Your message of "Mon, 02 Apr 2001 16:27:01 EDT." <DF737E620579D411A8E400D0B77E671DA20111@regdom-ex01.prod.netsol.com>
Sender:
owner-dnsop@cafax.se
Subject:
Re: RR distribution by type?
> Has anybody ever done or seen a (recent) survey of how various RR types are > distributed "in the wild" in real-world zones? Ideally, I'd be interested Here's an excerpt of what was discovered during the latest RIPE DNS hostcount <http://www.ripe.net/ripencc/pub-services/stats/hostcount/> for the TLD DE <http://www.TechFak.Uni-Bielefeld.DE/~pk/dns/hostcount/latest.html>. 38506954 total 13173185 A 11301138 NS 7970557 MX 4916802 SOA 829975 CNAME 202454 HINFO 90459 TXT 9544 PTR 4101 SRV 2465 WKS 2456 RP 1105 AAAA 1037 MB 943 MR 621 LOC 50 A6 23 KEY 18 MINFO 17 AFSDB 2 NSAP 2 MG The figures for A, NS and SOA are off by a factor of two (see ref'd document for an explanation). Some observations: o HINFO is rather popular o PTR RRs are due to RFC1101 and classless delegation (RFC2317) pointing into the "forward" zone o SRV is still rather low, but the survey only covers some 80 % (== 2.45 million) of the zones and only sees the externally visible part of the namespace o Given the distribution of AAAA and A6 - does this call for combined operations? Finally, the question whether a certain RR type is "used" remains unanswered for most of them, because you'd have to take into account both zone population and query patterns. -Peter