David J Taylor
david-taylor at blueyonder.co.uk.invalid
Fri Jun 18 18:23:22 UTC 2010
"Ryan Malayter" <malayter at gmail.com> wrote in message
news:AANLkTimOvweJjo8TwxA2XWgep4dhEbGAlGQv1vnHiQe7 at mail.gmail.com...
> The fact that niether the reference implementation nor w32time have
> direct support for SNMP, a *far* more widely used and documented
> managment standard, would seem to be another way to look at "who has a
> problem." Windows Group Policy and WMI comprise a far more widley
> "accepted management standard" than NTP mode 6/7 packets in any case.
I would welcome SNMP in NTP and have said so on a number of occasions.
> I agree, when you *need* ntpd features for specific applications. I
> recommend w32time when you don't, as that is the simlpest from an
> operational perspective. Keep it simple.
In the cases I have seen recently, W32time simply isn't good enough. One
set of users are looking for sub-second accuracy (a lot with Windows XP),
and another set for about millisecond accuracy using Windows. The former
can be reference NTP over the Internet, and the latter reference NTP with
a local GPS source.
>> RFC1305 refers to NTP v3, by the way, but I think most are now on NTP
> There is still no published RFC for NTPv4. Alternative implementers
> cannot be expected to read the source code of the reference
> implementation and track it for compatibiltiy and work-alike behavior.
Completely agreed. "You can read the source code" is no substitute for a
proper specification and test profile.
More information about the questions