Truly random numbers are generated with a minimum of extra hardware by taking advantage of the inherent noise in a communication channel. Random numbers can thus be generated without specialized manufacturing requirements and can be incorporated to conventional integrated circuits with minimal additional logic. The random number generation technique offloads the processor from performing extensive generation calculations without the use of the hardware accelerator. This random number generation technique may find application in, e.g., any network device that participates in a virtual private network or is used to implement electronic commerce.
A microelectronic apparatus and method for generating random binary words including at least one clocked pseudorandom binary number sequence generator normally operative to generate a cyclic output sequence of binary numbers, each number including a string of binary symbols, the cyclic output sequence including a basic sequence which is generated repeatedly, at least one bit stream generator generating a clocked bit stream including a stream of binary symbols of a first type occasionally interrupted by a binary symbol of a second type, wherein a first varying time interval between the occasional interruptions is intractably correlated to the output sequence of the number sequence generator, wherein each occurrence of an interruption of the stream of binary symbols of the first type by a binary symbol of the second type causes a pseudorandom modification of the cyclic output sequence of the number sequence generator and a sampling device operative to sample the cyclic output sequence of binary numbers thereby to generate a sampled output sequence including at least one sampled binary word.
A random number generator includes a noise signal generating device for supplying a noise signal, a sampling device connected between an output of the noise signal generating device and an output of the random number generator and sampling the noise signal in a sampling state and not sampling the noise signal in an idle state. The random number generator further includes a control oscillator for supplying a control alternating signal, the frequency of the control alternating signal of the control oscillator not being rigidly coupled to the frequency of the noise signal. The random number generator further includes an enabling device which puts the sampling device in the sampling state when the noise signal or the signal derived from the noise signal is in a first trigger state, and when, subsequently, the control alternating signal of the control oscillator is in a second trigger state. This is to ensure that the frequency of the output samples of the random number generator depends on the frequency of the noise signal, whereas the phase position is determined by the frequency of the control oscillator. Thus, the frequency of the output samples can be set and adjusted to existing requirements in a simple manner using the frequency setting functionality of the noise signal generating device.
