Actually, I suggested that the password expire at its max age (regardles
of whether any warnings have been sent).
Jim
Andrew Sciberras <andrew.sciberras@eB2Bcom.com> 10/25/04 3:36:19 PM
G'Day
I'm generally satisfied with this.
If any directories exist today that use the following model:
* pwdMaxAge - Absolute maximum age of the password
* pwdExpireWarning - Time period before the max age in which warnings
will be delivered,
Then changing the semantics of these attributes would lead to
unexpected
behavior if an organization upgraded their directory server to the new
functionality.
Eg. A directory that wants to warn people 6 months before their
password
is due to expire.
pwdMaxAge = 31536000 (1 year)
pwdExpireWarning = 15768000 (6 Months)
Updating the directory server to one that supports Jim's suggestion
below will result in the password reaching its max age, then remaining
valid for another 6 months.
I'm not too sure if this is likely to be a serious problem for
implementations, but some text in the security considerations of the
draft indicating this might be appropriate.
Cheers
_________________________________________
Andrew Sciberras
eB2Bcom - Software Engineer
Jim Sermersheim wrote:
I believe the intent (however wrongly formulated) was to allow the
user
to receive a warning no matter what. Even if the password's max age
has
passed, the user would be allowed pwdExpireWarning seconds to change
the
pwd. The definition of pwdExpireWarning talks about this in a not
very
precise way (The number of seconds before the password will expire
after
the user is first warned of its upcoming expiration.)
Some history to help make sense of things:
The password policy I-D was created as a blend of the (then)
Netscape
and Novell directory password policies.
I believe the original implementors of pwdExpireWarning (Netscape)
used
this to both warn of expiration, and also allow some kind of grace
login
period.
Novell's implementation didn't include the notion of a warning
period.
Only a number of grace logins.
So now we have two ways of achieving 'grace login'.
A better way of specifying the pwdExpireWarning and pwdMaxAge
concepts
would have been to use one attribute to specify an age at which an
expiration warning is sent, and another attribute specified how long
these warnings will continue before the password finally expires.
I dislike having two similar but different grace mechanisms, so I
propose that we remove pwdExpireWarned, and expire the password when
it
reaches pwdMaxAge (regardless of whether any warnings have been
sent).
I'll update the I-D to reflect this without debate (because the
deadline is so near), and we can go from there.
Jim
Andrew Sciberras < andrew.sciberras@eB2Bcom.com > 9/14/04 7:48:25
PM
Hi Niel,
Neil Dunbar wrote:
<SNIP>
The pwdMaxAge should be the absolute maximum time that the password
can
be used by anyone as a credential. The pwdExpirationWarning time, I
think, should be the earliest opportunity that the directory server
can
warn the user that his/her password is approaching expiry. If the
user
comes into the expiry period late in the game - tough. You can
always
use the grace logins feature to allow the user with the dud password
to
change it after it has ceased to be a meaningful credential for
general
directory operations
</SNIP>
If someone was to implement the draft in its current form, their
first
warning time would indicate the time difference between the current
time
and the time that the password is due to expire. Subsequent logins
would
result in a warning time that will go beyond the specified pwdMaxAge
allowing the user to receive their full warning period.
Our implementation, which was based around the -05 version of the
draft
handled this inconsistency by returning an initial warning message of
pwdExpireWarning.
I've now noticed, in version -07 of the draft, that the following new
line exists within the description of pwdExpireWarning:
If not 0, the value must be smaller than the value of the pwdMaxAge
attribute.
This seriously implies that the author's intention is to ensure that
the
warning time does not exceed the maximum age of the password.
I'm not extremely passionate about whether a user should receive
their
full warning period. Some consensus on this issue, and the author's
opinion (Jim?) would be good though.
Andrew Sciberras
eB2Bcom - Software Engineer
------------------------------------------------------------------------
_______________________________________________
Ldapext mailing list
Ldapext@ietf.org
https://www1.ietf.org/mailman/listinfo/ldapext