RE: [dss] Observations on TST

[Tim Moses <tim.moses@entrust.com>]

Colleagues - Another approach (that I quite like) is to include a time whose
semantics are that the authority asserts that the particular configuration
of bits existed on or before that time.  The authority would take its own
accuracy into account in determining what time to assert.  I.e. it would add
its own estimate of uncertainty to the time.  Then a client can simply rely
on the fact that (according to the authority) the particular configuration
of bits existed on or before the time it sees in the token.  What value is
there in knowing that the authority may be wrong and that they could have
existed earlier, by the amount of the uncertainty; the requestor might have
held onto the bits for an amount of time that far exceeds the authority's
uncertainty, before submitting them for timestamping.  All the best.  Tim.

-----Original Message-----
From: Trevor Perrin [mailto:trevp@trevp.net]
Sent: Friday, August 15, 2003 2:57 PM
To: Nick Pope; 'DSS'
Subject: RE: [dss] Observations on TST


At 09:32 AM 8/15/2003 +0100, Nick Pope wrote:
> > Trevor wrote:
> > One thing that worries me about the Accuracy field, as in RFC
> > 3161 and the
> > Entrust proposal, is that the client has to add and subtract the
Accuracy
> > field with the timestamp time to determine the upper and lower bounds.
> >
> > Isn't this non-trivial?  The client might have to know how many
> > days are in
> > each month, leap-years, leap-seconds, that sort of thing.
> >
> > Would it be easier for the server to explicitly return an upper and
lower
> > bound, instead of an Accuracy?
> >
>
>I am not expecting that the upper and lower bounds is the main interest.  I
>see accuracy as a quality indicator.  Changing to upper and lower bounds
>make it much more difficult to decide whether the time-stamp is accurate
>enough for my purposes.

That's true.  I wonder if there's anyone with experience using RFC 3161 who 
can tell us how useful Accuracy is, and what it's most commonly used for.

Trevor 


You may leave a Technical Committee at any time by visiting
http://www.oasis-open.org/apps/org/workgroup/dss/members/leave_workgroup.php