Demand side management has gained increasing importance as the penetration of renewable energy grows. Based on a Markov jump process model of a group of thermostatic loads, this paper proposes a two level feedback system to govern the interactions between an independent system operator (ISO) and a number of regulation reserve providers such that two objectives are achieved: 1) the ISO can optimally dispatch regulation signals to multiple providers in real time in order to reduce the requirement for expensive spinning reserves, and 2) each reserve provider can control its thermostatic loads to respond to the ISO signal. It is shown that the amount of regulation reserve that can be provided is implicitly restricted by a few fundamental parameters of the provider itself, such as the allowable set point choice and its thermal constant. An interesting finding is that there is a trade-off between an appliance's capacity for providing a large sustained reserve and its capacity of rapid reserve ramping. Simulations are presented to verify theoretical results.