This paper explores flow in complex nano-sized channels by use of molecular dynamics. Due to the small nature of these channels and to better capture wall effects, non-equilibrium molecular dynamics simulations were performed. Straight, constricted and sawtooth channels were studied. The function used for modeling the particle interactions is the Lennard-Jones 6-12 potential. Stochastic boundary conditions are used in conjunction with periodic boundary conditions in a 3D domain. Computational enhancements including cell subdivision and neighbor listing provide increased efficiency. The channels were homogeneous in the depth dimension and the results were averaged in the depth direction in order to improve averages. Velocity profiles at several locations were computed and are presented in the paper. The eventual goal of this research is to study the effects of time-dependent inflow and pressure drops so as to understand the flow in nano channels in the human bone.