Difference between revisions of "Password Storage Cheat Sheet"

From OWASP
Jump to: navigation, search
m (Use a modern hash algorithm)
(Grouped the information regarding work factor to avoid being redundant, while still adding more information. Added an example of the number of iterations used by a large company.)
(46 intermediate revisions by 7 users not shown)
Line 1: Line 1:
= DRAFT CHEAT SHEET - WORK IN PROGRESS =
 
 
 
= Introduction =
 
= Introduction =
  
This article is focused on providing guidance to storing a password in order to help prevent password theft. Too often passwords are stored as clear text. Thus the password can be read directly by the database’s administrator, super users or via data theft by SQL Injection. Database backup media is also vulnerable to password theft via password storage. It is recommended that you avoid storing the clear text password or an encrypted version of the password.
+
Media covers the theft of large collections of passwords on an almost daily basis. Media coverage of password theft discloses the password storage scheme, the weakness of that scheme, and often discloses a large population of compromised credentials that can affect multiple web sites or other applications. This article provides guidance on properly storing passwords, secret question responses, and similar credential information. Proper storage helps prevent theft, compromise, and malicious use of credentials.
 +
Information systems store passwords and other credentials in a variety of protected forms. Common vulnerabilities allow the theft of protected passwords through attack vectors such as SQL Injection. Protected passwords can also be stolen from artifacts such as logs, dumps, and backups.
 +
 
 +
Specific guidance herein protects against stored credential theft but the bulk of guidance aims to prevent credential compromise. That is, this guidance helps designs resist revealing users’ credentials or allowing system access in the event threats steal protected credential information. For more information and a thorough treatment of this topic, refer to the Secure Password Storage Threat Model here [http://goo.gl/Spvzs http://goo.gl/Spvzs].
 +
 
 +
= Guidance =
 +
 
 +
==  Do not limit the character set or length of credentials ==
 +
 
 +
Some organizations restrict the 1) types of special characters and 2) length of credentials accepted by systems because of their inability to prevent SQL Injection, Cross-site scripting, command-injection and other forms of injection attacks. These restrictions, while well-intentioned, facilitate certain simple attacks such as brute force.
 +
 
 +
Do not apply length, character set, or encoding restrictions on the entry or storage of credentials. Continue applying encoding, escaping, masking, outright omission, and other best practices to eliminate injection risks.
 +
 
 +
== Use a cryptographically strong credential-specific salt ==
 +
 
 +
A salt is fixed-length cryptographically-strong random value. Append credential data to the salt and use this as input to a protective function. Store the protected form appended to the salt as follows:
 +
 
 +
<code>[protected form] = [salt] + protect([protection func], [salt] + [credential]);</code>
 +
 
 +
Follow these practices to properly implement credential-specific salts:
 +
 
 +
* Generate a unique salt upon creation of each stored credential (not just per user or system wide);
 +
* Use cryptographically-strong random [*3] data;
 +
* As storage permits, use a 32bit or 64b salt (actual size dependent on protection function);
 +
* Scheme security does not depend on hiding, splitting, or otherwise obscuring the salt.
 +
 
 +
Salts serve two purposes: 1) prevent the protected form from revealing two identical credentials and 2) augment entropy fed to protecting function without relying on credential complexity. The second aims to make pre-computed lookup attacks [*2] on an individual credential and time-based attacks on a population intractable.
 +
 
 +
== Impose infeasible verification on attacker ==
 +
 
 +
The function used to protect stored credentials should balance attacker and defender verification. The defender needs an acceptable response time for verification of users’ credentials during peak use. However, the time required to map <code><credential> → <protected form></code>  must remain beyond threats’ hardware (GPU, FPGA) and technique (dictionary-based, brute force, etc) capabilities.
 +
 
 +
Two approaches facilitate this, each imperfectly.
 +
 
 +
=== Leverage an adaptive one-way function ===
 +
 
 +
Adaptive one-way functions compute a one-way (irreversible) transform. Each function allows configuration of ‘work factor’. Underlying mechanisms used to achieve irreversibility and govern work factors (such as time, space, and parallelism) vary between functions and remain unimportant to this discussion.
 +
 
 +
Select:
 +
 
 +
* PBKDF2 [*4] when FIPS certification or enterprise support on many platforms is required;
 +
* Scrypt [*5] where resisting any/all hardware accelerated attacks is necessary but support isn’t.
 +
 
 +
Example protect() pseudo-code follows:
 +
 
 +
<code>return [salt] + pbkdf2([salt], [credential], c=10000); </code>
 +
 
 +
Designers select one-way adaptive functions to implement protect() because these functions can be configured to cost (linearly or exponentially) more than a hash function to execute. Defenders adjust work factor to keep pace with threats’ increasing hardware capabilities. Those implementing adaptive one-way functions must tune work factors so as to impede attackers while providing acceptable user experience and scale.
 +
 
 +
Additionally, adaptive one-way functions do not effectively prevent reversal of common dictionary-based credentials (users with password ‘password’) regardless of user population size or salt usage.
 +
 
 +
==== Work Factor ====
 +
 
 +
Since resources are normally considered limited, a common rule of thumb for tuning the work factor (or cost) is to make protect() run as slow as possible without affecting the users' experience and without increasing the need for extra hardware over budget. So, if the registration and authentication's cases accept protect() taking up to 1 second, you can tune the cost so that it takes 1 second to run on your hardware. This way, it shouldn't be so slow that your users become affected, but it should also affect the attackers' attempt as much as possible.
 +
 
 +
While there is a minimum number of iterations recommended to ensure data safety, this value changes every year as technology improves. An example of the iteration count chosen by a well known company is the 10,000 iterations Apple uses for its iTunes passwords (using PBKDF2)[http://images.apple.com/ipad/business/docs/iOS_Security_May12.pdf](PDF file). However, it is critical to understand that a single work factor does not fit all designs. Experimentation is important.[*6]
 +
 
 +
=== Leverage Keyed functions ===
 +
 
 +
Keyed functions, such as HMACs, compute a one-way (irreversible) transform using a private key and given input. For example, HMACs inherit properties of hash functions including their speed, allowing for near instant verification. Key size imposes infeasible size- and/or space- requirements on compromise--even for common credentials (aka password = ‘password’).
 +
Designers protecting stored credentials with keyed functions:
 +
 
 +
* Use a single “site-wide” key;
 +
* Protect this key as any private key using best practices;
 +
* Store the key outside the credential store (aka: not in the database);
 +
* Generate the key using cryptographically-strong pseudo-random data;
 +
* Do not worry about output block size (i.e. SHA-256 vs. SHA-512).
 +
 
 +
Example protect() pseudo-code follows:
 +
 
 +
<code>return [salt] + HMAC-SHA-256([key], [salt] + [credential]);  </code>
 +
 
 +
Upholding security improvement over (solely) salted schemes relies on proper key management.
 +
 
 +
== Design protection/verification for compromise ==
  
== Password Storage Rules ==
+
The frequency and ease with which threats steal protected credentials demands “design for failure”. Having detected theft, a credential storage scheme must support continued operation by marking credential data compromised and engaging alternative credential validation workflows as follows:
  
Passwords are secrets. There is no reason to decrypt them under any circumstances. It is crucial that passwords are stored in a way that they can be *verified* but not *reversed* in any way, even by insiders. To accomplish this, store the salted hashed value of the password. Preferably use a different random salt for each password hash instead of a constant long salt.
+
1) Protect the user’s account
 +
  a. Invalidate authN ‘shortcuts’ disallowing login without 2nd factors or secret questions
 +
  b. Disallow changes to account (secret questions, out of band exchange channel setup/selection, etc.)
  
=== Use a modern hash algorithm ===
+
2) Load and use new protection scheme
# SHA
+
  a. Load a new (stronger) protect(credential) function
# bcrypt
+
  b. Include version information stored with form
 +
  c. Set ‘tainted’/‘compromised’ bit until user resets credentials
 +
  d. Rotate any keys and/or adjust protection function parameters (iter count)
 +
  e. Increment scheme version number
  
=== Use a long cryptographically random salt ===
+
3) When user logs in:
 +
  a. Validate credentials based on stored version (old or new); if old demand 2nd factor or secret answers
 +
  b. Prompt user for credential change, apologize, & conduct OOB confirmation
 +
  c. Convert stored credentials to new scheme as user successfully log in
  
If each password is simply hashed, identical passwords will have the same hash. There are two drawbacks to choosing to only storing the password’s hash:
+
Supporting workflow outlined above requires tight integration with Authentication frameworks and workflows.
# Due to the birthday paradox (http://en.wikipedia.org/wiki/Birthday_paradox), the attacker can find a password very quickly especially if the number of passwords the database is large.
+
# An attacker can use a list of precomputed hashed (http://en.wikipedia.org/wiki/Rainbow_table) to break passwords in seconds.
+
In order to solve these problems, a salt must be concatenated in front of the password before the digest operation.
+
  
A salt is a cryptographically random number of a fixed length. This salt must be different for each stored entry.  Since rainbow tables are already passing 24 characters, a salt of 24 bytes or longer is the recommended minimum length.
+
= References=
  
=== Iterate the hash ===
+
* [1] Morris, R. Thompson, K., Password Security: A Case History, 04/03/1978, p4: http://cm.bell-labs.com/cm/cs/who/dmr/passwd.ps
 +
* [2] Space-based (Lookup) attacks: Space-time Tradeoff: Hellman, M., Crypanalytic Time-Memory Trade-Off, Transactions of Information Theory, Vol. IT-26, No. 4, July, 1980 http://www-ee.stanford.edu/~hellman/publications/36.pdf Rainbow Tables -http://ophcrack.sourceforge.net/tables.php
 +
* [3] For example: http://docs.oracle.com/javase/6/docs/api/java/security/SecureRandom.html
 +
* [4] Kalski, B., PKCS #5: Password-Based Cryptography Specification Version 2.0, IETF RFC 2898, September, 2000, p9 http://www.ietf.org/rfc/rfc2898.txt
 +
* [5] Percival, C., Stronger Key Derivation Via Sequential Memory-Hard Functions, BSDCan ‘09, May, 2009 http://www.tarsnap.com/scrypt/scrypt.pdf
 +
* [6] For instance, one might set work factors targeting the following run times: (1) Password-generated session key - fraction of a second; (2) User credential - ~0.5 seconds; (3) Password-generated site (or other long-lived) key - potentially a second or more.
  
To slow down the computation it is recommended to iterate the hash operation many times. While hashing the password many times does slow down hashing for both attackers and typical users, typical users don't really notice it being that hashing is such a small percentage of their total time interacting with the system. On the other hand, an attacker trying to crack passwords spends nearly 100% of their time hashing so hashing many times gives the appearance of slowing the attacker down by a factor of n while not noticeably affecting the typical user. A minimum of 1000 operations is recommended in RSA PKCS5 standard in 2000, a value that should be doubled every 2 years.
+
= Authors and Primary Editors =
  
== References ==
+
John Steven - john.steven[at]owasp.org
  
Cryptographic framework for password hashing is described in [http://www.rsa.com/rsalabs/node.asp?id=2127 PKCS #5 v2.1: Password-Based Cryptography Standard]. Specific secure password hashing algorithms exist such as [http://www.usenix.org/events/usenix99/provos/provos_html/node1.html bcrypt], [http://www.tarsnap.com/scrypt/scrypt.pdf scrypt]. Implementations of secure password hashing exist for PHP ([http://www.openwall.com/phpass/ phpass]), ASP.NET ([http://msdn.microsoft.com/en-us/library/ms998372.aspx#pagpractices0001_sensitivedata ASP.NET 2.0 Security Practices]), Java ([http://www.owasp.org/index.php/Hashing_Java OWASP Hashing Java]).
+
= Other Cheatsheets =
  
 
{{Cheatsheet_Navigation}}
 
{{Cheatsheet_Navigation}}
  
 
[[Category:Cheatsheets]]
 
[[Category:Cheatsheets]]

Revision as of 15:12, 20 March 2013

Contents

Introduction

Media covers the theft of large collections of passwords on an almost daily basis. Media coverage of password theft discloses the password storage scheme, the weakness of that scheme, and often discloses a large population of compromised credentials that can affect multiple web sites or other applications. This article provides guidance on properly storing passwords, secret question responses, and similar credential information. Proper storage helps prevent theft, compromise, and malicious use of credentials. Information systems store passwords and other credentials in a variety of protected forms. Common vulnerabilities allow the theft of protected passwords through attack vectors such as SQL Injection. Protected passwords can also be stolen from artifacts such as logs, dumps, and backups.

Specific guidance herein protects against stored credential theft but the bulk of guidance aims to prevent credential compromise. That is, this guidance helps designs resist revealing users’ credentials or allowing system access in the event threats steal protected credential information. For more information and a thorough treatment of this topic, refer to the Secure Password Storage Threat Model here http://goo.gl/Spvzs.

Guidance

Do not limit the character set or length of credentials

Some organizations restrict the 1) types of special characters and 2) length of credentials accepted by systems because of their inability to prevent SQL Injection, Cross-site scripting, command-injection and other forms of injection attacks. These restrictions, while well-intentioned, facilitate certain simple attacks such as brute force.

Do not apply length, character set, or encoding restrictions on the entry or storage of credentials. Continue applying encoding, escaping, masking, outright omission, and other best practices to eliminate injection risks.

Use a cryptographically strong credential-specific salt

A salt is fixed-length cryptographically-strong random value. Append credential data to the salt and use this as input to a protective function. Store the protected form appended to the salt as follows:

[protected form] = [salt] + protect([protection func], [salt] + [credential]);

Follow these practices to properly implement credential-specific salts:

  • Generate a unique salt upon creation of each stored credential (not just per user or system wide);
  • Use cryptographically-strong random [*3] data;
  • As storage permits, use a 32bit or 64b salt (actual size dependent on protection function);
  • Scheme security does not depend on hiding, splitting, or otherwise obscuring the salt.

Salts serve two purposes: 1) prevent the protected form from revealing two identical credentials and 2) augment entropy fed to protecting function without relying on credential complexity. The second aims to make pre-computed lookup attacks [*2] on an individual credential and time-based attacks on a population intractable.

Impose infeasible verification on attacker

The function used to protect stored credentials should balance attacker and defender verification. The defender needs an acceptable response time for verification of users’ credentials during peak use. However, the time required to map <credential> → <protected form> must remain beyond threats’ hardware (GPU, FPGA) and technique (dictionary-based, brute force, etc) capabilities.

Two approaches facilitate this, each imperfectly.

Leverage an adaptive one-way function

Adaptive one-way functions compute a one-way (irreversible) transform. Each function allows configuration of ‘work factor’. Underlying mechanisms used to achieve irreversibility and govern work factors (such as time, space, and parallelism) vary between functions and remain unimportant to this discussion.

Select:

  • PBKDF2 [*4] when FIPS certification or enterprise support on many platforms is required;
  • Scrypt [*5] where resisting any/all hardware accelerated attacks is necessary but support isn’t.

Example protect() pseudo-code follows:

return [salt] + pbkdf2([salt], [credential], c=10000);

Designers select one-way adaptive functions to implement protect() because these functions can be configured to cost (linearly or exponentially) more than a hash function to execute. Defenders adjust work factor to keep pace with threats’ increasing hardware capabilities. Those implementing adaptive one-way functions must tune work factors so as to impede attackers while providing acceptable user experience and scale.

Additionally, adaptive one-way functions do not effectively prevent reversal of common dictionary-based credentials (users with password ‘password’) regardless of user population size or salt usage.

Work Factor

Since resources are normally considered limited, a common rule of thumb for tuning the work factor (or cost) is to make protect() run as slow as possible without affecting the users' experience and without increasing the need for extra hardware over budget. So, if the registration and authentication's cases accept protect() taking up to 1 second, you can tune the cost so that it takes 1 second to run on your hardware. This way, it shouldn't be so slow that your users become affected, but it should also affect the attackers' attempt as much as possible.

While there is a minimum number of iterations recommended to ensure data safety, this value changes every year as technology improves. An example of the iteration count chosen by a well known company is the 10,000 iterations Apple uses for its iTunes passwords (using PBKDF2)[1](PDF file). However, it is critical to understand that a single work factor does not fit all designs. Experimentation is important.[*6]

Leverage Keyed functions

Keyed functions, such as HMACs, compute a one-way (irreversible) transform using a private key and given input. For example, HMACs inherit properties of hash functions including their speed, allowing for near instant verification. Key size imposes infeasible size- and/or space- requirements on compromise--even for common credentials (aka password = ‘password’). Designers protecting stored credentials with keyed functions:

  • Use a single “site-wide” key;
  • Protect this key as any private key using best practices;
  • Store the key outside the credential store (aka: not in the database);
  • Generate the key using cryptographically-strong pseudo-random data;
  • Do not worry about output block size (i.e. SHA-256 vs. SHA-512).

Example protect() pseudo-code follows:

return [salt] + HMAC-SHA-256([key], [salt] + [credential]);

Upholding security improvement over (solely) salted schemes relies on proper key management.

Design protection/verification for compromise

The frequency and ease with which threats steal protected credentials demands “design for failure”. Having detected theft, a credential storage scheme must support continued operation by marking credential data compromised and engaging alternative credential validation workflows as follows:

1) Protect the user’s account

 a. Invalidate authN ‘shortcuts’ disallowing login without 2nd factors or secret questions
 b. Disallow changes to account (secret questions, out of band exchange channel setup/selection, etc.)

2) Load and use new protection scheme

 a. Load a new (stronger) protect(credential) function
 b. Include version information stored with form
 c. Set ‘tainted’/‘compromised’ bit until user resets credentials
 d. Rotate any keys and/or adjust protection function parameters (iter count)
 e. Increment scheme version number

3) When user logs in:

 a. Validate credentials based on stored version (old or new); if old demand 2nd factor or secret answers
 b. Prompt user for credential change, apologize, & conduct OOB confirmation
 c. Convert stored credentials to new scheme as user successfully log in

Supporting workflow outlined above requires tight integration with Authentication frameworks and workflows.

References

Authors and Primary Editors

John Steven - john.steven[at]owasp.org

Other Cheatsheets

OWASP Cheat Sheets Project Homepage

Developer Cheat Sheets (Builder)

Assessment Cheat Sheets (Breaker)

Mobile Cheat Sheets

OpSec Cheat Sheets (Defender)

Draft Cheat Sheets