Energy-Efficiency of Aggregation in Wireless Sensor Networks - A Practical Evaluation

In wireless sensor networks (WSNs), data aggregation is often proposed as the solution against resource scarcity, especially in terms of energy. Because data aggregation reduces communication volume in typical monitoring sc enarios significantly, the same is assumed regarding energy consumption. We take a closer look at this assumption, analyzing how energy-efficient data aggregation really is. Because simulations often suffer from approximations, we make a real-world evaluation using real nodes and a dedicated measurement device. Our experiments show, that reducing packet size has far lower impact to energy consumption than often assumed. Other parameters, e.g., the overall number of transmissions, or the MAC layer's sleep cycle, have far more influence. We therefore suggest to use packet aggregation in many scenarios where it is almost as energy efficient as data aggregation but has far less disadvantages. We think, this has impact on many common problems in WSNs. For example, end-to-end security is nearly impossible to realize with data aggregation but feasible with simple packet aggregation.