ad-assurance - [AD-Assurance] Suggested compensating controls for passwords
Subject: Meeting the InCommon Assurance profile criteria using Active Directory
List archive
- From: "Capehart,Jeffrey D" <>
- To: "" <>
- Subject: [AD-Assurance] Suggested compensating controls for passwords
- Date: Tue, 28 May 2013 17:35:50 +0000
- Accept-language: en-US
- Authentication-results: sfpop-ironport04.merit.edu; dkim=neutral (message not signed) header.i=none
Last week on our call there was some discussion about passwords as Authentication Secrets versus the “Challenge/Response” protocol and what is being sent over the network. SP800-63 breaks these parts into “long term authentication secrets”
and “short term authentication secrets”. However, Silver does not appear to do that in its definition of Authentication Secret. I came across a good review of some compensating controls for passwords from the NIST publication SP800-82
GUIDE TO
INDUSTRIAL
CONTROL
SYSTEMS
(ICS) SECURITY. These compensating controls may be considered when the full set of (FISMA) controls and control enhancements cannot be met. http://csrc.nist.gov/publications/nistpubs/800-82/SP800-82-final.pdf These same ideas could potentially be used as a basis for developing ALTERNATIVE MEANS for Active Directory or other “password” store, providing the context is similar. The considerations specifically state: Passwords should not be sent across any network unless protected by some form of FIPS-approved encryption or salted cryptographic hash specifically
designed to prevent replay attacks. It is assumed that the device used to enter a password is connected to the network in a secure manner.
NOTE: A salted cryptographic hash DESIGNED to prevent replay attacks --- if FIPS-approved encryption doesn’t work. See more info in these extracts: ICS Specific Recommendations and Guidance
Computer systems in ICS environments typically rely on traditional passwords for authentication. Control system suppliers often supply systems with default passwords. These passwords are factory set and are often easy to guess or are changed
infrequently, which creates additional security risks. Also, protocols currently used in ICS environments generally have inadequate or no network service authentication. There are now several forms of authentication available in addition to traditional password
techniques being used with ICS. Some of these, including password authentication, are presented in the following sections with discussions regarding their use with ICS.
6.3.1.1 Password Authentication
Password authentication technologies determine authenticity based on testing for something the device or human requesting access
should know, such as a PIN number or password. Password authentication schemes are thought of as the simplest and most common forms of authentication.
Password vulnerabilities can be reduced by using an active password checker that prohibits weak, recently used, or commonly used passwords. Another weakness is the ease of
third-party eavesdropping. Passwords typed at a keypad or keyboard are easily observed or recorded, especially in areas where adversaries could plant tiny wireless cameras or keystroke loggers. Network service authentication often transmits passwords as plaintext
(unencrypted), allowing any network capture tool to expose the passwords. Some ICS operating systems make setting secure passwords difficult, as the password size is very small and the system allows only group passwords at each level of access, not individual
passwords. Some industrial (and Internet) protocols transmit passwords in plaintext, making them susceptible to interception. In cases where this practice cannot be avoided, it is important that users have different (and unrelated) passwords for use with encrypted
and non-encrypted protocols. The following are general recommendations and considerations with regards to the use of passwords.
·
The length, strength, and complexity of passwords should balance security and operational ease of access within the capabilities of the software and underlying OS.
·
Passwords should have appropriate length and complexity for the required security. In particular, they should not be able to be found in a dictionary or contain predictable sequences of numbers or
letters.
·
Passwords should be used with care on operator interface devices such as control consoles on critical processes. Using passwords on these consoles could introduce potential safety issues if operators
are locked out or delayed access during critical events. Physical security should supplement operator control consoles when password protection is not feasible.
·
The keeper of master passwords should be a trusted employee, available during emergencies. Any copies of the master passwords must be stored in a very secure location with limited access.
·
The passwords of privileged users (such as network technicians, electrical or electronics technicians and management, and network designers/operators) should be most secure and be changed frequently.
Authority to change master passwords should be limited to trusted employees. A password audit record, especially for master passwords, should be maintained separately from the control system.
·
In environments with a high risk of interception or intrusion (such as remote operator interfaces in a facility that lacks local physical security access controls), organizations should consider
supplementing password authentication with other forms of authentication such as challenge/response or multi-factor authentication using biometric or physical tokens.
·
For user authentication purposes, password use is common and generally acceptable for users logging directly into a local device or computer. Passwords should not be sent across any network unless
protected by some form of FIPS-approved encryption or salted cryptographic hash specifically designed to prevent replay attacks. It is assumed that the device used to enter a password is connected to the network in a secure manner.
·
For network service authentication purposes, passwords should be avoided if possible. There are more secure alternatives available, such as challenge/response or public key authentication.
6.3.1.2 Challenge/response Authentication
Challenge/response authentication requires that both the service requester and service provider know a “secret” code in advance.
When service is requested, the service provider sends a random number or string as a challenge to the service requester. The service requester uses the secret code to generate a unique response for the service provider. If the response is as expected, it proves
that the service requester has access to the “secret” without ever exposing the secret on the network.
Challenge/response authentication addresses the security vulnerabilities of traditional password authentication. When passwords (hashed or plain) are sent across a network,
a portion of the actual “secret” itself is being sent. Authentication is performed by giving the secret to the remote device. Jeff Capehart, CISA |
- [AD-Assurance] Suggested compensating controls for passwords, Capehart,Jeffrey D, 05/28/2013
Archive powered by MHonArc 2.6.16.