[Ntp] Re: NTP Extension Field additions, and IANA NTP table management process

[Miroslav Lichvar <[e-mail]>]

On Wed, Feb 19, 2025 at 02:12:17AM -0800, Harlan Stenn wrote:
> The format of EFs and the Type field has been as I have described it since
> the inception of extension fields in NTPv4.

Maybe that was the intention, but RFC5905 didn't describe it like
that. RFC5906 is informational.

> I don't understand.  Are you saying that 256 type may not be enough for
> NTPv4?

Yes. It's not just NTPv4. There will be NTPv5, v6, ...

> I'll point out that we're now talking about 18 or 19 (or if your number is
> correct, 23) out of 256 total being consumed in *25 years*.

That doesn't mean the rate cannot accelerate. The 13 proposed ones
could be allocated in a single year or only few years apart.

> Beyond that, it would be trivial to declare a new EF type, say an Extended
> Extension Field, which would be implemented in that packet's data payload,
> that would have room for at least 16*256 *more* EFs.

Or we could just fully use the existing 16-bit type, avoiding
complications in future.

> The working group is already talking about v5, and v5 can change the EF
> stuff in any way it wants.

The current plan is to keep all types in one table to keep it simple.
The EFs specified for NTPv4 will still work in NTPv5.

-- 
Miroslav Lichvar