[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: comments on draft-ietf-ltans-pki-retention-00.txt



Haven't yet got the time to read the draft but some comments inline.

>
> <snip>
> > So the first thing will be to define the adequate
> > certificate/PKI/ policy  for your needs and then decide abou
> > the necessary cert.
> > management structure. Legally spoken: it is my responsibility
> > to store the necessary verification data, how I will do that,
> > should remain in my responsibility. Having said this, I do
> > not think this should be part of the LTANS standard.
>
> Do you think LTANS specs can play a role here, i.e., ERS?  If not, what
> specs apply?  The retention spec defines an optional binding between
> certs and evidence records that enables you to fulfill your
> responsibility of preserving the necessary verification data.

No, I don't thnik ERS should have anything to do with retetnion. IMO, this is
pure process managerial issue and should be binded to the archive objects build
of several parts (see the LTAP slides for Dallas), namely:

- archive data
- metainformation (including operation information, sec. attributes, etc.)
- ERS

  It's not
> clear to me that certs must be preserved as part of a signed data object
> especially since not doing this has significant benefit, e.g., storage
> savings and decoupling trust anchors from signed data.
> </snip>

Law requirments usually clearly specify the availability of 'complementary'
information needed for security attributes validation in time (and this issue
is addressed by by RFC3126 and its XML interpretation (XAdES), however from
different perspective than ERS). Such information may be obtained on request or
be part of the archive object. The problem is the availability and integrity.
Such information may cease to exist for whatever reason, including decomposing
CA. Henece the need to place it next to archive data. But you are right, this
is storage saving issue also. So one may simply use pointers for example, while
use ERS to demonstrate integrity.

>
> Not an answer but a comment to this question:
> I agree with Carl in the point that the retention of certificates etc.
> is important and that it would be good to have this at a central point.
> At the first glance it might even be better located in PKIX - on the
> other side  the charter of LTANS "Long term archiving and notary
> services" does in my eyes not exclude such a spec AND which is most
> important it is an important (at the moment not really specified part)
> of the puzzle for an application for long term storage of signed
> documents.

Another problem with retention period is that it may change. So placing this
information on a specific place does rise some issues. This is why I think it
is an archive data handling information.

Aleksej