[Yeti DNS Discuss] Scoring the Yeti DNS Root Server System

Davey Song(宋林健) ljsong at biigroup.cn
Tue Dec 20 12:19:27 UTC 2016

Hi colleagues,


We repeat Geoff’s test against Yeti root server and make scores for them.
Comments are welcome.



There are findings I would like to list in this mail to attract some


* M root fragments only TCP segments. It is odd that large UDP responses of
M are not fragmented. But the yeti root server run by WIDE fragments only
UDP instead.

* We introduce a new star on good compression:If the server compresses all
labels in the packet, we give the server a star. Because we notice a
Microsoft DNS server does not do name compression for the mname and rname
fields in the SOA record which introduce additional 12 octets.

* #14 server run by SWITCH get ★★★★★ as a only 5-point server in Yeti
DNS project. Congratulations!

* As Geoff's second post on this topic IPv6 DNS fragmentation issue is still

* Since Yeti produces very large DNS response(1600+) which is fragmented by
every root server. To Truncate or to fragment is a problem. Or we can use
new proposals for very large DNS response suggested in our article.


Best regards,


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.yeti-dns.org/pipermail/discuss/attachments/20161220/0098490b/attachment.html>

More information about the discuss mailing list