In this thesis, a new stack architecture for sensor networks is proposed. The stack consists of the following layers: application, query, aggregation, network, MAC and physical. Various algorithms are implemented using this stack and it is shown that this stack is modular. Following an overview of sensor networks, the previous protocol stack suggestions for sensor networks are examined. Sensor network algorithms that can be classified as sensor data management systems are surveyed and compared with each other. Four of the surveyed algorithms, namely, TAG, Synopsis Diffusion, Tributary-Delta and Directed Diffusion are implemented using the introduced stack. The implementation is performed using a sensor network model developed with OMNeT++ simulator. The simulation results are compared to the original results of these algorithms. Obtaining similar results, the stack and algorithm implementations are validated, moreover, it is shown that the stack does not induce any performance degradation. Using the implementation details of the algorithms, the modularity of the suggested stack is demonstrated. Finally, additional benefits of the stack are discussed.