In the noise section, Razavi has mentioned we are calculating the random swings by computing the average power of the noise and he is considering the noise to be a periodic signal...I am a bit skeptical, since it is random, isn't it better to consider it as a non periodic signal