Quantum key distribution (QKD) is believed to be the first realizable application of quantum information science.  In fact, over short distances, such QKD systems are commercially available.  Over long distances, however, the scenario is much different.  If we do not have access to a network of trusted nodes for key regeneration, the only known solution for long-distance QKD relies on entanglement swapping or quantum repeater systems.  Such systems face their own implementation challenges, including the need for quantum memory devices and highly efficient gates and detectors.  In this talk, I discuss a variety of physical requirements for quantum repeaters, compare different architectures for entanglement distribution, and address the prospect of developing these systems in the near future.