rand - pseudo-random number generator
#include <openssl/rand.h> int RAND_set_rand_engine(ENGINE *engine); int RAND_bytes(unsigned char *buf, int num); int RAND_pseudo_bytes(unsigned char *buf, int num); void RAND_seed(const void *buf, int num); void RAND_add(const void *buf, int num, double entropy); int RAND_status(void); int RAND_load_file(const char *file, long max_bytes); int RAND_write_file(const char *file); const char *RAND_file_name(char *file, size_t num); int RAND_egd(const char *path); void RAND_set_rand_method(const RAND_METHOD *meth); const RAND_METHOD *RAND_get_rand_method(void); RAND_METHOD *RAND_SSLeay(void); void RAND_cleanup(void); /* For Win32 only */ void RAND_screen(void); int RAND_event(UINT, WPARAM, LPARAM);
Since the introduction of the ENGINE API, the recommended way of controlling default implementations is by using the ENGINE API functions. The default RAND_METHOD, as set by RAND_set_rand_method() and returned by RAND_get_rand_method(), is only used if no ENGINE has been set as the default "rand" implementation. Hence, these two functions are no longer the recommended way to control defaults.
If an alternative RAND_METHOD implementation is being used (either set directly or as provided by an ENGINE module), then it is entirely responsible for the generation and management of a cryptographically secure PRNG stream. The mechanisms described below relate solely to the software PRNG implementation built in to OpenSSL and used by default.
These functions implement a cryptographically secure pseudo-random number generator (PRNG). It is used by other library functions for example to generate random keys, and applications can use it when they need randomness.
A cryptographic PRNG must be seeded with unpredictable data such as mouse movements or keys pressed at random by the user. This is described in RAND_add(3). Its state can be saved in a seed file (see RAND_load_file(3)) to avoid having to go through the seeding process whenever the application is started.
RAND_bytes(3) describes how to obtain random data from the PRNG.
The RAND_SSLeay() method implements a PRNG based on a cryptographic hash function.
The following description of its design is based on the SSLeay documentation:
First up I will state the things I believe I need for a good RNG.
A good hashing algorithm to mix things up and to convert the RNG 'state' to random numbers.
An initial source of random 'state'.
The state should be very large. If the RNG is being used to generate 4096 bit RSA keys, 2 2048 bit random strings are required (at a minimum). If your RNG state only has 128 bits, you are obviously limiting the search space to 128 bits, not 2048. I'm probably getting a little carried away on this last point but it does indicate that it may not be a bad idea to keep quite a lot of RNG state. It should be easier to break a cipher than guess the RNG seed data.
Any RNG seed data should influence all subsequent random numbers generated. This implies that any random seed data entered will have an influence on all subsequent random numbers generated.
When using data to seed the RNG state, the data used should not be extractable from the RNG state. I believe this should be a requirement because one possible source of 'secret' semi random data would be a private key or a password. This data must not be disclosed by either subsequent random numbers or a 'core' dump left by a program crash.
Given the same initial 'state', 2 systems should deviate in their RNG state (and hence the random numbers generated) over time if at all possible.
Given the random number output stream, it should not be possible to determine the RNG state or the next random number.
The algorithm is as follows.
There is global state made up of a 1023 byte buffer (the 'state'), a working hash value ('md'), and a counter ('count').
Whenever seed data is added, it is inserted into the 'state' as follows.
The input is chopped up into units of 20 bytes (or less for the last block). Each of these blocks is run through the hash function as follows: The data passed to the hash function is the current 'md', the same number of bytes from the 'state' (the location determined by in incremented looping index) as the current 'block', the new key data 'block', and 'count' (which is incremented after each use). The result of this is kept in 'md' and also xored into the 'state' at the same locations that were used as input into the hash function. I believe this system addresses points 1 (hash function; currently SHA-1), 3 (the 'state'), 4 (via the 'md'), 5 (by the use of a hash function and xor).
When bytes are extracted from the RNG, the following process is used. For each group of 10 bytes (or less), we do the following:
Input into the hash function the local 'md' (which is initialized from the global 'md' before any bytes are generated), the bytes that are to be overwritten by the random bytes, and bytes from the 'state' (incrementing looping index). From this digest output (which is kept in 'md'), the top (up to) 10 bytes are returned to the caller and the bottom 10 bytes are xored into the 'state'.
Finally, after we have finished 'num' random bytes for the caller, 'count' (which is incremented) and the local and global 'md' are fed into the hash function and the results are kept in the global 'md'.
I believe the above addressed points 1 (use of SHA-1), 6 (by hashing into the 'state' the 'old' data from the caller that is about to be overwritten) and 7 (by not using the 10 bytes given to the caller to update the 'state', but they are used to update 'md').
So of the points raised, only 2 is not addressed (but see RAND_add(3)).