[dns-operations] new public DNS service: 9.9.9.9

[Jared Mauch]

> On Nov 23, 2017, at 11:29 AM, Stephane Bortzmeyer <bortzmeyer at nic.fr> wrote:
> 
> On Thu, Nov 23, 2017 at 03:46:02PM +0200,
> Taras Heichenko <tasic at hostmaster.ua> wrote 
> a message of 25 lines which said:
> 
>> If you know folks from quad9 could you send to them this link
>> https://www.ripe.net/ripe/mail/archives/ripe-atlas/2017-November/003495.html
>> which point to letter that describe weird effect of quad9 work?
> 
> There is nothing wrong here and Quad9 has nothing to fix. Using ICMP
> to measure the RTT of a DNS server is brittle and yield unexpected
> results. If you want to know the RTT of Quad9, send DNS requests.

Yup.

It’s a proxy for what works well (or might work well), but for what doesn’t work well - it often require more study.

Those probes could be bad (for example).  For the ICMP case I just created a one-off with 500 random probes.

If you want Ripe ATLAS credits to do your own study, send me a email with what your RIPE username is so I can transfer you some credits.

- jared