[dns-operations] Allowance for inaccurate clocks

Peter Koch pk at DENIC.DE
Thu Mar 17 02:34:56 UTC 2011


On Wed, Mar 16, 2011 at 06:50:04PM -0400, Olafur Gudmundsson wrote:

> I agree with Mark, here 1 day in the past is a good time for signature
> initiation time.

while I cannot see a specific reason to be too strict with the inception
time, the logic re: clock fuzz would hold in the opposite direction, i.e.
whereever you'd have remaining RRSIG lifetimes of n days, you'd have to account
for n+1 days instead.  Not sure I buy that.

-Peter



More information about the dns-operations mailing list