Energy limitation is the main concern of any wireless sensor network application. The communication between nodes is the greedy factor for the energy consumption. One important mechanism to reduce energy consumption is the in-network data aggregation. In-network data aggregation removes redundancy as well as unnecessary data forwarding, and hence cuts on the energy used in communications. Recently a new kind of applications are proposed which consider, in addition to energy efficiency, data latency and accuracy as important factors. Reducing data latency helps increasing the network throughput and early events detection. Before performing the aggregation process, each node should wait for a predefined time called WT (waiting time) to receive data from other nodes. Data latency (resp., accuracy) is decreased (resp., increased), if network nodes are well scheduled through optimal distribution of WT over the nodes. Many solutions have been proposed to schedule network nodes in order to make the data aggregation process more efficient. In this paper, we propose a taxonomy and classification of existing data aggregation scheduling solutions. We survey main solutions in the literature and illustrate their operations through examples. Furthermore, we discuss each solution and analyze it against performance criteria such as data latency and accuracy, energy consumption and collision avoidance. Finally, we shed some light on future research directions and open issues after deep analysis of existing solutions.