Technical strategy and performance

Eric S. Raymond esr at
Thu Jun 30 04:31:52 UTC 2016

Hal Murray <hmurray at>:
> The big picture question that comes to mind is why did we start by forking 
> ntp classic? Why not start from scratch?  Did anybody consider chrony?  What 
> other options are/were there?

We forked from Classic for the same reasons we didn't start from
scratch and didn't use chrony - that is, the demand from the large
users driving our funding.  Originally, that was (implicitly)
government users for which the initial NSF term grant was buying a
more secure NTP. Later the customer base expanded and shifted towards
the large corporate users backing LF.

The thing about these users is they're intensely conservative.  They
won't buy a time service implementation that doesn't reassure them by
looking and smelling like what they're used to.  In many cases, especially in
governmant, they *can't* -- they have lengthy certification requirements
for new infrastructure components.

Thus, while it is in theory possible that something like timed or chrony
could be a better solution, it's not a practical one.  From Day One what
the people driving our funding have wanted is Mills NTP without the suck.

That's what I've been hired to do, and that's what I intend to deliver.

So a practical way for you to grasp what is driving my design
decisions and even the style of the documentation is: to ask what
features and guarantees you would want from NTPsec, and what would you
find most reassuring, if you were any of: Cisco. The National Science
Founation.  IBM.  NASA.  Samsung Electronics.  A random three-letter
spook shop.  Amazon.  The NOAA.  Google.

These are one sort of users I'm thinking about constantly - the ones
with big modern systems. The other kind is hackers and makers with
*small* modern systems - Pis and other hackerboards, mobile devices,
embedded deplyments.

Serving time hobbyists trying to squeeze the last bit of service life
out of 10 or 20-year hardware is sentimentally nice, and I'll do it
when it doesn't conflict with our other objectives, but it definitely
comes in third behind making the previous two groups happy.

I speak under the possibility of correction by Mark, whose job it is to
set project policy.  But I believe our priorities are very similar.

> Where would I look to find a crisp statement of the goals of the project?

On the project website.
		<a href="">Eric S. Raymond</a>

More information about the devel mailing list