An easy way to keep your users' passwords secure.
If you store user passwords in the clear, then an attacker who steals a copy of your database has a giant list of emails and passwords. Some of your users will only have one password -- for their email account, for their banking account, for your application. A simple hack could escalate into massive identity theft.
It's your responsibility as a web developer to make your web application secure -- blaming your users for not being security experts is not a professional response to risk.
bcrypt()
allows you to easily harden your application against these kinds of attacks.
Note: JRuby versions of the bcrypt gem <= 2.1.3
had a security
vulnerability that
was fixed in >= 2.1.4
. If you used a vulnerable version to hash
passwords with international characters in them, you will need to
re-hash those passwords. This vulnerability only affected the JRuby gem.
gem install bcrypt
The bcrypt gem is available on the following ruby platforms:
- JRuby
- RubyInstaller 1.8, 1.9, 2.0, 2.1, and 2.2 builds on win32
- Any 1.8, 1.9, 2.0, 2.1, 2.2, or 2.3 Ruby on a BSD/OS X/Linux system with a compiler
Note: Rails versions >= 3 ship with ActiveModel::SecurePassword
which uses bcrypt-ruby.
has_secure_password
docs
implements a similar authentication strategy to the code below.
require 'bcrypt'
class User < ActiveRecord::Base
# users.password_hash in the database is a :string
include BCrypt
def password
@password ||= Password.new(password_hash)
end
def password=(new_password)
@password = Password.create(new_password)
self.password_hash = @password
end
end
def create
@user = User.new(params[:user])
@user.password = params[:password]
@user.save!
end
def login
@user = User.find_by_email(params[:email])
if @user.password == params[:password]
give_token
else
redirect_to home_url
end
end
require 'bcrypt'
my_password = BCrypt::Password.create("my password")
#=> "$2a$10$vI8aWBnW3fID.ZQ4/zo1G.q1lRps.9cGLcZEiGDMVr5yUP1KUOYTa"
my_password.version #=> "2a"
my_password.cost #=> 10
my_password == "my password" #=> true
my_password == "not my password" #=> false
my_password = BCrypt::Password.new("$2a$10$vI8aWBnW3fID.ZQ4/zo1G.q1lRps.9cGLcZEiGDMVr5yUP1KUOYTa")
my_password == "my password" #=> true
my_password == "not my password" #=> false
Check the rdocs for more details -- BCrypt, BCrypt::Password.
bcrypt()
is a hashing algorithm designed by Niels Provos and David Mazières of the OpenBSD Project.
Hash algorithms take a chunk of data (e.g., your user's password) and create a "digital fingerprint," or hash, of it. Because this process is not reversible, there's no way to go from the hash back to the password.
In other words:
hash(p) #=> <unique gibberish>
You can store the hash and check it against a hash made of a potentially valid password:
<unique gibberish> =? hash(just_entered_password)
But even this has weaknesses -- attackers can just run lists of possible passwords through the same algorithm, store the results in a big database, and then look up the passwords by their hash:
PrecomputedPassword.find_by_hash(<unique gibberish>).password #=> "secret1"
The solution to this is to add a small chunk of random data -- called a salt -- to the password before it's hashed:
hash(salt + p) #=> <really unique gibberish>
The salt is then stored along with the hash in the database, and used to check potentially valid passwords:
<really unique gibberish> =? hash(salt + just_entered_password)
bcrypt-ruby automatically handles the storage and generation of these salts for you.
Adding a salt means that an attacker has to have a gigantic database for each unique salt -- for a salt made of 4 letters, that's 456,976 different databases. Pretty much no one has that much storage space, so attackers try a different, slower method -- throw a list of potential passwords at each individual password:
hash(salt + "aadvark") =? <really unique gibberish>
hash(salt + "abacus") =? <really unique gibberish>
etc.
This is much slower than the big database approach, but most hash algorithms are pretty quick -- and therein lies the
problem. Hash algorithms aren't usually designed to be slow, they're designed to turn gigabytes of data into secure
fingerprints as quickly as possible. bcrypt()
, though, is designed to be computationally expensive:
Ten thousand iterations:
user system total real
md5 0.070000 0.000000 0.070000 ( 0.070415)
bcrypt 22.230000 0.080000 22.310000 ( 22.493822)
If an attacker was using Ruby to check each password, they could check ~140,000 passwords a second with MD5 but only
~450 passwords a second with bcrypt()
.
In addition, bcrypt()
allows you to increase the amount of work required to hash a password as computers get faster. Old
passwords will still work fine, but new passwords can keep up with the times.
The default cost factor used by bcrypt-ruby is 10, which is fine for session-based authentication. If you are using a stateless authentication architecture (e.g., HTTP Basic Auth), you will want to lower the cost factor to reduce your server load and keep your request times down. This will lower the security provided you, but there are few alternatives.
To change the default cost factor used by bcrypt-ruby, use BCrypt::Engine.cost = new_value
:
BCrypt::Password.create('secret').cost
#=> 10, the default provided by bcrypt-ruby
# set a new default cost
BCrypt::Engine.cost = 8
BCrypt::Password.create('secret').cost
#=> 8
The default cost can be overridden as needed by passing an options hash with a different cost:
BCrypt::Password.create('secret', :cost => 6).cost #=> 6
bcrypt()
is currently used as the default password storage hash in OpenBSD, widely regarded as the most secure operating
system available.
For a more technical explanation of the algorithm and its design criteria, please read Niels Provos and David Mazières' Usenix99 paper: http://www.usenix.org/events/usenix99/provos.html
If you'd like more down-to-earth advice regarding cryptography, I suggest reading Practical Cryptography by Niels Ferguson and Bruce Schneier: http://www.schneier.com/book-practical.html
- Author :: Coda Hale coda.hale@gmail.com
- Website :: http://blog.codahale.com