home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky sci.crypt:2723 comp.security.misc:786
- Newsgroups: sci.crypt,comp.security.misc
- Path: sparky!uunet!cs.utexas.edu!qt.cs.utexas.edu!yale.edu!nigel.msen.com!hela.iti.org!widener!iggy.GW.Vitalink.COM!nocsun.NOC.Vitalink.COM!indetech!cirrus!dhesi
- From: dhesi@cirrus.com (Rahul Dhesi)
- Subject: Re: Crypt should be based on MD5 (was: the Crypt 16 discussion)
- Message-ID: <1992Jul23.013037.24195@cirrus.com>
- Sender: news@cirrus.com
- Organization: Cirrus Logic Inc.
- References: <2a510a22@babyoil.ftp.com> <709960260@romeo.cs.duke.edu> <62451@cup.portal.com> <16990@ulysses.att.com>
- Date: Thu, 23 Jul 1992 01:30:37 GMT
- Lines: 50
-
- Some general comments about password expiration. Assumption: the
- intruder is trying to guess a password by already having access to the
- encrypted form.
-
- Suppose T1 is the time it takes for the intruder to determine a password
- by some sort of search (guessing, exchaustive, whatever). Suppose T2 is
- the time it takes for an intruder to actually *begin using* a password
- once he has guessed it. So if the intruder acquires an encrypted
- password at time t0, he can begin using it at time t0 + T1 + T2.
-
- It makes sense to set the password expiration period to be less than T1 +
- T2. Then, by the time the intruder is ready to use the password, it has
- been changed.
-
- So what are useful values for T1 and T2?
-
- My guess is that T2 can be as little as 15 seconds, the time it takes to
- log in.
-
- For well-chosen Unix passwords T1 is at least decades, if not several
- centuries. We need to expire a password in (some decades + 15 seconds).
- It should be reasonable to expire a password in about a year.
-
- For *poorly* chosen passwords, T1 might be as short as a day. In that
- case, passwords should expire every few hours.
-
- CONCLUSION 1: If passwords are well-chosen, password expiration adds
- nothing to security. If passwords are poorly-chosen, password expiration
- must occur too rapidly to be of much use.
-
- However, consider the intruder's point of view. If the intruder can
- immediately get access to the encrypted form a password when it is
- changed, then password expiration becomes irrelevant. Mathematically
- speaking, if the intruder is always trying the *current* encrypted
- password, his chances of guessing it neither increase nor decrease if the
- encrypted password changes every 6 hours (or however often).
-
- CONCLUSION 2: If the intruder has immediate access to changed encrypted
- passwords, then password expiration has no beneficial effect at all.
-
- FINAL CONCLUSION: Password expiration sounds good, may impress
- customers, and certainly makes a sysadmin feel powerful and in control
- (the VAX/VMS syndrome). But it's far, far better to use a "passwd"
- program that doesn't allow poor passwords in the first place. And if you
- do so, shadow password files aren't really needed, though they are still
- a good thing.
- --
- Rahul Dhesi <dhesi@cirrus.com>
- also: dhesi@rahul.net
- "He's metabolically challenged, Jim."
-