A Gossip-based Churn Estimator for Large Dynamic Networks
Cristiano Giuffrida, S. Ortolani · VU Research Portal · 2010
Gossip-based aggregation is an emerging paradigm to perform distributed computations and measurements in a large-scale setting. In this paper we explore the possibility of using gossip-based aggregation to estimate churn in arbitrarily large networks. To this end, we introduce a new model to compute local estimates and formally prove how aggregated values closely match the real churn with high accuracy independently of the network setting. Experimental results confirm the viability of our approach