A paper a day keeps the doctor away: Medians and Beyond: New Aggregation Techniques for Sensor Networks
The Internet of Things has spurred renewed interest in sensors and sensor telemetry. For each IoT application, modern sensors--ones that support sensing, computation, and communication, collect data about their environment, and relay it back to a "base station". The base station aggregates data from all the sensors to make sense of the telemetry. Since most of the sensors are deployed in settings that have severe constrains on power consumption and battery life, it is usually expensive to send all the raw telemetry data to the "base station" without aggregation or sampling at each sensor. For certain class of metrics data, such as counts, sums, and averages, the base station can combine the aggregates from each sensor, and come up with meaningful statistics overall. However for sensor aggregates such as median, mode, and percentiles, the job of the base station becomes harder. The authors of the paper " Medians and Beyond: New Aggregation Techn...