Re: [Ntp] NTPv5 draft

[Miroslav Lichvar <[e-mail]>]

On Tue, Dec 01, 2020 at 10:20:48AM +0100, James wrote:
> Given that largest number of deployments of NTPv4 operate on public
> internet, if our intention is to have a protocol to supersede NTPv4 and
> solve the issues of those existing use cases, the protocol must provide
> mitigation to the existing threats of both server malfeasance and middlebox
> tampering and prevent downgrading that would enable either.

Which security-related issues of existing use cases do you expect
NTPv5 to solve?

> NTS is not the
> only option and roughtime has shown it is possible to provide authentication
> primitives as a core part of the protocol without the larger overheads that
> a bolt-on requires.

Symmetric keys are not practical on public internet. If roughtime has
some significant advantages over NTS that would be useful in NTP,
there is nothing stopping us from specifying a roughtime extension for
NTP. I'm not sure why it would need to be included in the core protocol.

-- 
Miroslav Lichvar