Robust Distributed Data Aggregation

One application of aggregation is decentralized network size estimation algorithms. Systems, such as Skype, estimate the number of active users by centralized algorithms, enacting a high demand on central resources and creating single points of failure. Aggregation introduces scalable decentralized algorithms that allow counting (or summing in the general case) over a distributed network of nodes. A particular application is found on sensor networks, where data collection is often only practicable if aggregation is performed. Aggregation-based techniques can be used to detect how many sensors are deployed in a given area or what is the average temperature in the covered area. The current state of the art in this area depicts a rich variety of algorithms, showing different accuracies, time and communication tradeoffs. In particular, averaging algorithms are adequate for high precision estimation of aggregates but are highly sensitive to message loss.