The ultimate goal of systems biology is to attain a quantitative, predictive understanding of the behavior of a living system as a whole (i.e., physiology) from its molecular parts. A major obstacle to this endeavor is the enormous number of (mostly inaccessible) parameters underlying complex biological systems. My lab has tried to tackle this problem using a top-down approach starting from cellular physiology. In this talk, I will show how a simple phenomenological approach in the spirit of thermodynamics can provide quantitative, predictive understanding of the physiological behaviors of bacteria cells. I will present a number of linear relations describing the allocation of cellular resources for exponentially growing E. coli cells. With a few phenomenological parameters, these "growth laws" can be used to accurately predict the physiological responses to various perturbations, including the fitness effect of protein over-expression, and the abrupt cellular response to the application of antibiotics. Applying the top-down approach to the endogenous response of E. coli to different modes of nutrient limitation, we reveal key molecular interactions enabling seamless coordination between different branches of metabolism -- interactions which Monod pursued until the end of his life but remained elusive despite decades of extensive molecular studies.