phc-discussions - Re: [PHC] "Why I Don't Recommend Scrypt"

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: Thu, 13 Mar 2014 17:43:43 +0100
From: CodesInChaos <codesinchaos@...il.com>
To: discussions@...sword-hashing.net
Subject: Re: [PHC] "Why I Don't Recommend Scrypt"

1) When you store passwords to authenticate users on login, you mainly
care about lower class attackers who use a GPU.
   When deriving encryption keys you also care about high class
attackers who can afford the development cost of ASIC.

   bcrypt is GPU unfriendly, scrypt only with quite large memory use.
Scrypt allows time memory tools which are useful for speeding it up on
GPUs.

2) On login servers you have much tighter time constraints. Perhaps
only 10ms per password. This limits the memory use in scrypt a lot.
    For disk encryption you can afford spending a second to derive the key.

    Since scrypt's advantage over bcrypt grows with the amount of
memory you use, it doesn't gain much when you only have 10ms to
compute it, and thus don't have
    time to access hundreds of megabytes.

=> For GPU based attackers and defenders with short time constraints
the advantage of scrypt over bcrypt rather small or it's possibly even
worse. Servers can't afford
     hashing for a second and using a GB of RAM in the process, so
scrypt can't shine.