[LEAPSECS] Google, Amazon, now Microsoft

[Warner Losh]

> On May 31, 2015, at 1:55 PM, Daniel R. Tobias <dan at tobias.name> wrote:
> 
> On 31 May 2015 at 19:33, Poul-Henning Kamp wrote:
> 
>> Most likely, at some random time after the leapsecond, your clock
>> steps a second.
> 
> ...which is basically how most computers deal with time
> synchronization, excepting the minority that actually attempt
> continuous high precision and accuracy; the computer's clock, not all
> that accurate a timepiece, drifts off from external standards within
> the period (hours, days, weeks) between synchronizations, and has to
> step a few seconds one way or the other to catch up (possibly
> smoothed out to prevent discontinuities that harm processes in
> progress); the leap second, if any, is lost in the noise.

This is precisely the attitude that prevents leap seconds from
being implemented properly. “It’s just a second” and “we’ve never
cared, so why should we start” are lame cop-outs.

And most systems I’ve used certainly are synchronized via NTP.
All macs, all my servers, etc all maintain a reasonably coherent
time scale. Failure to do this, btw, will lead many of the synchronization
algorithms and such to have false timeouts, or worse. Some
are resilient in the face of these weird steps. Others less so.

Then again, my data is skewed, since I never use Windows.

Warner

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 842 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <https://pairlist6.pair.net/pipermail/leapsecs/attachments/20150531/a4c2e111/attachment.pgp>