Re: [Ntp] Antw: [EXT] Re: Splitting the Roughtime draft?

[Magnus Danielson <[e-mail]>]

Hi,

On 2021-02-01 22:40, Warner Losh wrote:
>
>
> On Mon, Feb 1, 2021 at 12:29 AM Ulrich Windl
> <[e-mail]
> <mailto:[e-mail]>> wrote:
>
>     >>> Hal Murray <[e-mail]
>     <mailto:[e-mail]>> schrieb am 31.01.2021 um 10:06 in
>     Nachricht
>     <[e-mail]
>     <mailto:[e-mail]>>:
>
>     > [e-mail] <mailto:[e-mail]> said:
>     >> I think we need to be very clear about the fact that all trust in
>     Roughtime
>     >> is rooted in the long‑term keys and that they are expected to
>     be valid for
>     a
>     >> very long time indeed.
>     >
>     > How long is "very long"?
>
>     The "industry standard" seems to be 10 years for that, while
>     "long" nowadays
>     is probably only two years...
>
>
> Consumer grade stuff is like 1-2 years. But deployed, embedded gear
> still needs 5-10 years depending on the segment it is in. You don't
> want to climb a lot of telephone poles to redeploy every couple of
> years, for example...

Actual Consumer grade stuff is longer, much longer than the vendors of
it think. For many operator purposes systems can survive 15-20 years,
easily.

It's hard to set such life-spans. Some systems have much longer
life-spans, but then upgrades occurs. What we should focus on is to make
it long enough so that a upgrade over the net can occur, even if that
occurs over a low enough rate. Re-keying over the air is the way to go
long-term. Then you have the problem with long storage times, and ways
to upgrade them with keys, and override if they gone over the time.

Requirements is starting to emerge on this, and I have been pushing for
it, but it takes time before you can rely on it in such system designs.
So we will have to push both sides.

Cheers,
Magnus