Bandwidth availability, device count increases and new applications such as machine learning have resulted in a steady growth of cloud computing needs, that data centers are required to answer. Among the available axes of improvement, data center scheduling acts as a virtually free technique to optimize resource utilization. Sparrow, proposed in 2013, is a distributed scheduler with on the fly decisions that adapts the power of two choices for load balancing to scheduling. This report aims at reproducing some simulation results from Sparrow and to discuss the authors’ adjustment of the power of two choices for their implementation.