phc-discussions - Re: [PHC] Native server relief support for password hashing in browsers

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Sat, 18 Jan 2014 12:14:34 -0800
From: Andy Lutomirski <luto@...capital.net>
To: discussions <discussions@...sword-hashing.net>
Subject: Re: [PHC] Native server relief support for password hashing in browsers

On Sat, Jan 18, 2014 at 12:03 PM, Bill Cox <waywardgeek@...il.com> wrote:
> On Sat, Jan 18, 2014 at 2:44 PM, Tony Arcieri <bascule@...il.com> wrote:
>> I suggested to Brendan Eich (via Twitter) that browsers directly implement
>> server relief for password hashing functions:
>>
>> https://twitter.com/BrendanEich/status/424282367335731201
>>
>> I think it'd be pretty cool if that happened!
>
> I agree.  One thing I haven't figured out is how to do Blakerypt style
> session key protection with server relief.  A Blakerypt session key is
> a secret associated with the password which is decrypted with a secret
> master key on the server.  So long as it remains a secret, off-line
> brute-force attacks become impractical.  It's an awesome extra level
> of protection for servers that implement it.  The problem is how can
> we keep the session key secret if we broadcast it to clients?  I
> suspect this is solvable...
>

What protection does this offer above the following simple scheme:

Choose a random key k.  Store it in an HSM.

Server stores, for each user, (username, salt (etc), E_k(password
hash)).  HSM has a primitive that takes E_k(password hash) and the
hash and tells you whether they match.

For even more fun, an HSM instead implement the server end of SRP or
another augmented PAKE given E_k(the verifier).  That verifier could
be based on an expensive hashed version of the password.

Of course, once you have HSMs in the picture, one wonders whether
expensive hashing is needed at all.

--Andy