CWE-203: Observable Discrepancy

Description

The product behaves differently or sends different responses under different circumstances in a way that is observable to an unauthorized actor.

Potential Impact

Confidentiality, Access Control

Read Application Data, Bypass Protection Mechanism

Confidentiality

Read Application Data

Demonstrative Examples

The following code checks validity of the supplied username and password and notifies the user of a successful or failed login.

Bad

my $username=param('username');
                  my $password=param('password');
                  
                  if (IsValidUsername($username) == 1)
                  {
                  if (IsValidPassword($username, $password) == 1)
                  {
                  print "Login Successful";
                  }
                  else
                  {
                  print "Login Failed - incorrect password";
                  }
                  }
                  else
                  {
                  print "Login Failed - unknown username";
                  }

In the above code, there are different messages for when an incorrect username is supplied, versus when the username is correct but the password is wrong. This difference enables a potential attacker to understand the state of the login function, and could allow an attacker to discover a valid username by trying different values until the incorrect password message is returned. In essence, this makes it easier for an attacker to obtain half of the necessary authentication credentials.

While this type of information may be helpful to a user, it is also useful to a potential attacker. In the above example, the message for both failed cases should be the same, such as:

Result

"Login Failed - incorrect username or password"

In this example, the attacker observes how long an authentication takes when the user types in the correct password.

When the attacker tries their own values, they can first try strings of various length. When they find a string of the right length, the computation will take a bit longer, because the for loop will run at least once. Additionally, with this code, the attacker can possibly learn one character of the password at a time, because when they guess the first character right, the computation will take longer than a wrong guesses. Such an attack can break even the most sophisticated password with a few hundred guesses.

Bad

def validate_password(actual_pw, typed_pw):
		 
                   if len(actual_pw) <> len(typed_pw):
		   return 0
                   for i in len(actual_pw):
		   if actual_pw[i] <> typed_pw[i]:
		   return 0
                   
                   return 1

Note that in this example, the actual password must be handled in constant time as far as the attacker is concerned, even if the actual password is of an unusual length. This is one reason why it is good to use an algorithm that, among other things, stores a seeded cryptographic one-way hash of the password, then compare the hashes, which will always be of the same length.

Non-uniform processing time causes timing channel.

Bad

Suppose an algorithm for implementing an encryption routine works fine per se, but the time taken to output the result of the encryption routine depends on a relationship between the input plaintext and the key (e.g., suppose, if the plaintext is similar to the key, it would run very fast).

In the example above, an attacker may vary the inputs, then observe differences between processing times (since different plaintexts take different time). This could be used to infer information about the key.

Good

Artificial delays may be added to ensure that all calculations take equal time to execute.

Suppose memory access patterns for an encryption routine are dependent on the secret key.
An attacker can recover the key by knowing if specific memory locations have been accessed or not.  The value stored at those memory locations is irrelevant.  The encryption routine's memory accesses will affect the state of the processor cache.  If cache resources are shared across contexts, after the encryption routine completes, an attacker in different execution context can discover which memory locations the routine accessed by measuring the time it takes for their own memory accesses to complete.

Mitigations & Prevention

Architecture and Design

Compartmentalize the system to have "safe" areas where trust boundaries can be unambiguously drawn. Do not allow sensitive data to go outside of the trust boundary and always be careful when interfacing with a compartment outside of the safe area. Ensure that appropriate compartmentalization is built into the system design, and the compartmentalization allows for and reinforces privilege separation functionality. Architects and designers should rely on the principle of least

Implementation

Ensure that error messages only contain minimal details that are useful to the intended audience and no one else. The messages need to strike the balance between being too cryptic (which can confuse users) or being too detailed (which may reveal more than intended). The messages should not reveal the methods that were used to determine the error. Attackers can use detailed information to refine or optimize their original attack, thereby increasing their chances of success. If

Real-World CVE Examples

CVE ID	Description
CVE-2020-8695	Observable discrepancy in the RAPL interface for some Intel processors allows information disclosure.
CVE-2019-14353	Crypto hardware wallet's power consumption relates to total number of pixels illuminated, creating a side channel in the USB connection that allows attackers to determine secrets displayed such as PIN
CVE-2019-10071	Java-oriented framework compares HMAC signatures using String.equals() instead of a constant-time algorithm, causing timing discrepancies
CVE-2002-2094	This, and others, use ".." attacks and monitor error responses, so there is overlap with directory traversal.
CVE-2001-1483	Enumeration of valid usernames based on inconsistent responses
CVE-2001-1528	Account number enumeration via inconsistent responses.
CVE-2004-2150	User enumeration via discrepancies in error messages.
CVE-2005-1650	User enumeration via discrepancies in error messages.
CVE-2004-0294	Bulletin Board displays different error messages when a user exists or not, which makes it easier for remote attackers to identify valid users and conduct a brute force password guessing attack.
CVE-2004-0243	Operating System, when direct remote login is disabled, displays a different message if the password is correct, which allows remote attackers to guess the password via brute force methods.
CVE-2002-0514	Product allows remote attackers to determine if a port is being filtered because the response packet TTL is different than the default TTL.
CVE-2002-0515	Product sets a different TTL when a port is being filtered than when it is not being filtered, which allows remote attackers to identify filtered ports by comparing TTLs.
CVE-2002-0208	Product modifies TCP/IP stack and ICMP error messages in unusual ways that show the product is in use.
CVE-2004-2252	Behavioral infoleak by responding to SYN-FIN packets.
CVE-2001-1387	Product may generate different responses than specified by the administrator, possibly leading to an information leak.

Showing 15 of 23 observed examples.

Taxonomy Mappings

PLOVER: — Discrepancy Information Leaks
OWASP Top Ten 2007: A6 — Information Leakage and Improper Error Handling
OWASP Top Ten 2004: A7 — Improper Error Handling

Frequently Asked Questions

What is CWE-203?

CWE-203 (Observable Discrepancy) is a software weakness identified by MITRE's Common Weakness Enumeration. It is classified as a Base-level weakness. The product behaves differently or sends different responses under different circumstances in a way that is observable to an unauthorized actor.

How can CWE-203 be exploited?

Attackers can exploit CWE-203 (Observable Discrepancy) to read application data, bypass protection mechanism. This weakness is typically introduced during the Architecture and Design, Implementation phase of software development.

How do I prevent CWE-203?

Key mitigations include: Compartmentalize the system to have "safe" areas where trust boundaries can be unambiguously drawn. Do not allow sensitive data to go outside of the trust boundary and always be careful when interfaci

What is the severity of CWE-203?

CWE-203 is classified as a Base-level weakness (Medium abstraction). It has been observed in 23 real-world CVEs.