Deciding what modes to keep.

[Daniel Franke]

On 9/29/16, Gary E. Miller <gem at rellim.com> wrote:
> I use that [peer mode] all the time.  Why would someone not?  it allows some
> optimizations in the protocol and in the relationship between servers.

Compared to simply making two hosts clients of each other, the only
optimization you get is a 50% reduction in network traffic (who
cares?). In theory if not also in practice, you sacrifice some
precision, because the much longer in-flight time of each exchange
gives more time for your oscillator to drift, resulting in wider error
bars on the value of (t4-t1). It forces you to accept packets with
invalid origin timestamps because otherwise a single dropped packet
will DoS the association forever; this makes it completely insecure
without authentication (anybody on the internet can manipulate your
clock), and even with authentication certain replay attacks remain
possible.

With all that said, I'm not actually advocating for removing it,
because I want to keep it around to use as a testbed for the
DTLS-encapsulated-NTPv4 portion of my NTS proposal.