Collection of the sensed data in a wireless sensor network at one or more sink (s) is a well studied problem and there are a lot of efficient solutions for a variety of wireless sensor network configurations and application requirements. These methods are often optimized towards collection of the sensed data at a predetermined base station or sink. This inherently reduces the agility of the wireless sensor network as the flow of information is not easily changeable after the establishment of the routing and data collection algorithms. This paper presents an efficient data dissemination method based on the compressed sensing theory that allows each sensor node to take the role of a sink. Agile sink selection is especially advantageous in scenarios where the sink or the end user of the wireless sensor network is mobile. The proposed method allows availing the global state of the environment by fetching a small set of data from any arbitrary node. Our evaluations prove the better performance of our technique over existing methods. Also a comparison with an oracle-based approach gives sufficient experimental evidences of a nearly optimal performance of our method. © 2014 IEEE.