Re: [gpsd-dev] Fwd: Re: Clarifications needed for the time-service HOWTO

[Greg Troxel]

"Gary E. Miller" <address@hidden> writes:

> Several RFCs say to prefer IPv6 over IPv4.  Almost all IPv6 aware
> programs like browsers Part of the rational is to alert people to their
> sucky IPv6 and get it fixed. :-)

Agreed.  My v6 is actually pretty much ok, but it has 30 ms more
latency.  A lot of my mail and some web goes over it.

> Many programs, like ssh and rsync, have -4 and -6 options to prefer one
> over the other.  Nice to have after you have identified your problem and
> your upstream will not fix it.

Sure.  ntpq does too.   I meant something more, because "geolocation" is
really "network topology location", and one's v6 location and v4
location are different.   But given that ntpd does dns queries over some
AF to a local resolver and the resolver might be dual-stack and use
either, the notion that geolocation hints obtained from a server using
one AF are valid for the other AF is both suspect and hard to work
around.

pgpdEwQs0EIjY.pgp
Description: PGP signature