Noise is ubiquitous in dynamical systems. Its impact on the dynamics 
can be quite versatile, ranging from bad signal-to-noise ratios to 
qualitatively new phenomena such as oscillations of populations, 
which would be absent without noise. Here "noise" stands for 
stochastic fluctuations of various origin. Of particular interest 
are stochastic fluctuations in the context of biology. What is 
their role in evolutionary processes or in the living cell as the 
basic unit of life? We shall focus on a dynamical unit which 
consists of only two species interacting in a non-linear way. 
This unit may be regarded as a coarse-grained description of a 
genetic circuit. For comparison we shall first discuss the unit 
in a deterministic description and then turn to a fully stochastic
 one. There we shall unravel the effect of demographic fluctuations 
and fluctuations in the reaction times. The power spectrum will 
show which source of stochastic behavior is dominant. We shall 
compare analytic predictions with Gillespie simulations.  The 
results raise an interesting generic question about oscillatory 
genetic systems observed in nature.