Re: [Ntp] Antw: [EXT] Re: NTPv5: big picture

[Magnus Danielson <[e-mail]>]

Hi,

On 2021-01-04 15:35, Kurt Roeckx wrote:
> On Mon, Jan 04, 2021 at 03:16:39PM +0100, Ulrich Windl wrote:
>> Should dynamic addition, reconfiguration and deletion of time sources be a
>> SHOULD for v5 at least?
> I think the draft should not document a protocol for how to do
> that. That is an implementation detail. It should probably not
> even go over the UDP port.

I think there can be good reasons to actually put it into a standard, as
with gazillion of boxes and implementations, making less issues of how
to manage this hurd of equipment with a wild mix of open source and
various vendors interpretations of things, it will be a pain never the less.

Now, this does not need to be the core standard, but a separate
management standard. Then, those needing it, can require conformance
with that separate standard and then may the market forces decide if it
needs to be implemented or not. In practice, it is probably needed.

There is a limit to how much one can deferr to "implementation issues".
Management and monitoring issues is a huge issue in practice. While we
can keep such things out of the core specification, it probably needs to
be in the suite of standards offering a set of solutions for various
uses. The alternative is that some will risk divert away from using NTP.

Cheers,
Magnus