Is there any flag in an NTP response to indicate Google's leap second smearing is in effect?

Skip to first unread message

Kerry Shetline

Nov 20, 2019, 2:33:02 PM11/20/19
to public-ntp-discuss
Without having to maintain a list of NTP servers which use leap second smearing, and those which don't, is there a way to tell, just by looking at the response from an NTP server, whether leap-second smearing will be applied or not?

I think I can be sure about this for any time server, but since Google proposes their smearing as a standard, I imagine many other time servers might follow suit.

I know from what I've read that the LI field will always be zero when smearing is used. I suppose that if I simultaneously consulted a different NTP server that I could be certain did NOT using smearing, then I'd know smearing would happen when the LI values differed. A more straight-forward test would be desirable, however.

Ideally this would be a flag that would be passed on by any NTP server which is itself synced to a Google NTP server, or other leap-second smearing server.

Bryan Tinsley

Dec 3, 2019, 9:12:16 AM12/3/19
to Kerry Shetline, public-ntp-discuss
As you probably found in your reading, there is a recently expired draft RFC here: (slightly more consumable version:  We currently return int32 encoded GOOG in S1 refid.  Even if we implemented the above, you would only be able to affirm that you are (or are not) receiving smeared time when you received a response from an NTP server that provided the 254.b1.b2.b3 encoded correction.  

We previously decided to not go down this route because of lack of client support and speculative benefits.  We're also highly reluctant to add more code paths that are only exercised during a leap smear.  

Hope this helps!

You received this message because you are subscribed to the Google Groups "public-ntp-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
To view this discussion on the web visit
Reply all
Reply to author
0 new messages