J2EE Misconfiguration: Insufficient Session ID Length

Session identifiers should be at least 128 bits long to prevent brute-force session guessing attacks.

The WebLogic deployment descriptor should specify a session identifier length of at least 24 bytes. A shorter session identifier leaves the application open to brute-force session guessing attacks. If an attacker can guess an authenticated user's session identifier, he can take over the user's session. The remainder of this explanation will detail a back-of-the-envelope justification for a 24-byte session identifier.

The session identifier is composed of a pseudorandom selection of the 62 alphanumeric characters, which means that if the string were composed in a truly random fashion each byte could yield a maximum of 6 bits of entropy.

The expected number of seconds required to guess a valid session identifier is given by the equation:

(2^B+1) / (2*A*S)

Where:

- B is the number of bits of entropy in the session identifier.

- A is the number of guesses an attacker may try each second.

- S is the number of valid session identifiers that are valid and available to be guessed at any given time.

The number of bits of entropy in the session identifier is always less than the total number of bits in the session identifier. For example, if session identifiers were provided in ascending order, there would be close to zero bits of entropy in the session identifier no matter the identifier's length. Assuming that the session identifiers are being generated using a good source of random numbers, we will estimate the number of bits of entropy in a session identifier to be half the total number of bits in the session identifier. For realistic identifier lengths this is possible, though perhaps optimistic.

If attackers use a botnet with hundreds or thousands of drone computers, it is reasonable to assume that they could attempt tens of thousands of guesses per second. If the web site in question is large and popular, a high volume of guessing might go unnoticed for some time.

A lower bound on the number of valid session identifiers that are available to be guessed is the number of users that are active on a site at any given moment. However, any users that abandon their sessions without logging out will increase this number. (This is one of many good reasons to have a short inactive session timeout.)

With a 64-bit session identifier, assume 32 bits of entropy. For a large web site, assume that the attacker may try 1,000 guesses per second and that there are 10,000 valid session identifiers at any given moment. Given these assumptions, the expected time for an attacker to successfully guess a valid session identifier is less than 4 minutes.

Now assume a 128-bit session identifier that provides 64 bits of entropy. With a very large web site, an attacker might try 10,000 guesses per second with 100,000 valid session identifiers available to be guessed. Given these assumptions, the expected time for an attacker to successfully guess a valid session identifier is greater than 292 years.

Working backwards from bits to bytes, now, the session identifier must be 128/6, which yields approximately 21 bytes. Furthermore, empirical testing has demonstrated that the first three bytes of the session identifier do not appear to be randomly generated, which means to achieve our desired 64 bits of entropy we need to configure WebLogic to use a session identifier 24 bytes in length.