Coming back to this unanswered question: from my point of view the color matching seems a bit harsh with regard to operational reality. I certainly won't get any nightmares for non exhaustive additional sections in response to extremely large queries.
This said, while looking at this I noticed something that leaves an uneasy feeling: for an extremely large query, I get 2 types of responses from the root servers. One set of root servers fills the additional section with 2 names, and their v4 and v6 addresses. But it's always the same two servers, indepently of the server asked. The other set answers with a bit more servers, but only v4 adresses, and here again, always the same list. I was expecting (perhaps naively) that the additional section would be filled randomly until no space is left. The current setup seems to restrict some of the inherent redundancy of the DNS. Any thoughts? Gilles On 22/5/12 17:00 , Edward Lewis wrote: > I looked back at this and noticed I should have said "the right matrix" > is more important. > > Still, I'm a little surprised there's been no response from a comparison > of a DNS Operations WG document to real-world measurements. > > Are we okay with a document that issues criteria that only 70% of TLDs > fall into "green?" Are the grades overly harsh? Should the TLDs be > urged to work on getting to green? > > At 13:55 -0400 5/10/12, Edward Lewis wrote: >>> A URL for this Internet-Draft is: >>> http://www.ietf.org/internet-drafts/draft-ietf-dnsop-respsize-14.txt >> >> For kicks I ran the included perl code against the 313 delegations >> from the root zone (313 does not include "root" and does include the >> 11 test TLDs). >> >> ...I'll leave it to the draft to explain the headings and rankings >> below. I did this just to measure the draft's assessments against the >> TLD settings. Not that the TLDs are the only audience of this draft, >> but it's a convenient data set. >> >> Measured in %'ages of 313: >> >> Max length domain name || Average length domain name >> Color A-only A+AAAA A-preferred || A-only A+AAAA A-preferred >> || >> Green 59% 19% 19% || 99% 69% 69% >> Yellow 35% 41% 18% || 1% 31% 28% >> Orange 4% 21% 9% || 1% >> Red 2% 11% 53% || 2% >> >> The left matrix is more important (operationally). That's about the >> only commentary I'll add, just to head off the concern of "red"s in >> the right matrix. >> >> To Paul and Akira, here's more or less what I did with the results of >> the pl: >> >> looping through the names in the root zone: >> server_set=`dig @127.0.0.1 +short $name ns` >> size_estimate=`perl respsize.pl $server_set` >> results=`echo $size_estimate | sed "getting '()'s#$name \2 \3 \4 \6 >> \7 \8#"` >> echo "$results" >> and then used a spreadsheet to do the percentages. (Shown in case I >> messed up something.) >> -- >> -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- >> >> Edward Lewis >> NeuStar You can leave a voice message at >> +1-571-434-5468 >> >> 2012...time to reuse those 1984 calendars! >> _______________________________________________ >> DNSOP mailing list >> DNSOP@ietf.org >> https://www.ietf.org/mailman/listinfo/dnsop > -- Fondation RESTENA 6, rue Coudenhove-Kalergi L-1359 Luxembourg tel: (+352) 424409 fax: (+352) 422473 _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
