Theoretical and experimental evidence has currently been accumulating that chaotic dynamics is real and probably ubiquitous in natural systems.  The implications of these phenomena are huge for the inherent predictability and interpretation of computer simulations and experiments in soil ecosystems.  Initially, deterministic chaos was thought by many to be mainly a mathematical phenomenon.  However, in our opinion 3 papers (Becks et al, 2005, Nature Letters; Graham et al. 2007, Int. Soc Microb. Eco. J.; Beninca et al., 2008, Nature Letters) have brought together, using experimental studies and relevant mathematics, a breakthrough discovery that deterministic chaos is present in relatively simple biochemical systems.  Two of us (Faybishenko and Molz, 2013, Procedia Environ. Sci)) have numerically analyzed a mathematical model of rhizosphere dynamics (Kravchenko et al., 2004, Microbiology) and detected patterns of nonlinear dynamical interactions supporting evidence of synchronized synergetic oscillations of microbial populations, carbon and oxygen concentrations driven by root exudation. These interactions were inherent in the processes, but not realized previously from experimental nor modeling studies. To better understand experimental results, we have developed a new mathematical model to describe the Becks et al. (2005) chemostat experiment, with numerical solutions being evaluated at the present time.  Based on the experimental and theoretical results, we suggest that nonlinear dynamics and deterministic chaotic processes are essential characteristics that may lead to the sustainability and resilience of ecosystems.  In many mechanical systems chaos is viewed as a type of instability, appearing when energy is added.  However, based on the analysis of both the Becks et al. experiments, and our theoretical and computational results, we hypothesize that chaos in the soil-plant-microbe ecosystem is driven by a decrease in the food supply (chemical energy).  Somewhat paradoxically, this, in turn, may support a long-term system stability, suppressing extinction or extreme overgrowth of any component.