In this chapter, we study robustness of resource allocation policies. DR allows us to evaluate the mean response time of a broad class of threshold-based policies under a wide range of load conditions, and this leads to lessons useful in designing robust resource allocation policies.