[atlas] DNS RTT over TCP: twice as long than UDP?

On Fri, Jul 05, 2019 at 12:33:25PM +0000,
 Giovane Moura <giovane.moura at sidn.nl> wrote 
 a message of 26 lines which said:

> My hypothesis is that the field *rtt* on DNS queries on Atlas[1],
> for TCP, is, in fact, measuring 2 RTTs: the RTT of the TCP
> handshake, and the RTT of query/response itself.

I assume (I didn't check with tcpdump) that Atlas starts the clock
when the SYN packet leaves, and dig when the DNS request leaves. That
would explain your observation.

Both make sense, and I disagree when you say that the second method is
the only right one.

In practice, which method is the most relevant depends on whether you
use persistent TCP connections or not.