A large fraction of new distributed generation will use renewable sources, and cannot be treated as conventional generators due to high variability. Energy storage is a natural mechanism for absorbing variability, and hence minimizing costly reliance on generator reserves. We consider a simple scenario in which energy imbalances are allocated to energy storage, representing either cooperation with a renewable producer or competitive operation in markets. The optimal storage scheduling strategy is shown to be an inventory control-like policy, which through mild approximation yields affine relationships with the state variables. We develop and numerically evaluate suboptimal policies that approximately incorporate the effects of correlation. Further approximation enables us to characterize the Nash equilibria arising between generation and storage, which we use to analyze strategic effects. Specifically, we find that inefficient storage can lead to non-socially optimal behavior.