The important points were
>Btw -- large password files using anything like this scheme are obsolescent.
>You can't use a hashed password for challenge/response, 
....
>The fundamental problem is that users pick bad passwords and passphrases ...

Yup.  I like S/Key better than the annoying SecureID card I use to
log in to work, or public-key challenge/responses where there's
an intelligent client that can use them.

>> >> b.  Use a unique per-passphrase salt of at least 32 bits.

If you're going to bother with a salt, might as well make it 64 or 128 bits;
increasing storage by 2**32 is fine, but some combination of Moore's law,
holographic storage, tape robots or whatever may catch up with you, 
but if you're doing an iterated-SHA1 or equivalent, you can allow 
long passphrases and still use enough salt to make things unstorable,
forcing the cracker to iterate calculations every time.

>You're arguing with 20-20 hindsight.  16 gig of disk space wasn't even
>comprehensible then.  16 *meg* of disk space sufficed to bring up UNIX.

On the other hand, 16 gig of tape was comprehensible then, if large,
and tape sorting was still common technology - much more annoying with
160MB tapes than Exabytes or whatever the current big tapes are,
but you could sort things into some convenient retrieval order.

>How would I like to do it, given a blank slate?  Most likely, I'd use
SHA-1 on
>the user's password, probably concatenated with a salt, to produce a DSA
>private key; the server would store the corresponding public key.  (It's
>harder to pull a trick like that using RSA keys.) 

A while back I did a login protocol based on Diffie-Hellman;
it turned out to be relatively easy (though unfortunately someone from
Siemens had also discovered it and patented it in Germany and then the US :-)
But almost any public-key system can give you a good mechanism for a
challenge/response and set up a shared secret for encrypting or AHing
a login session so it doesn't get hijacked.
                                Thanks! 
                                        Bill
Bill Stewart, [EMAIL PROTECTED]
PGP Fingerprint D454 E202 CBC8 40BF  3C85 B884 0ABE 4639

Reply via email to