CLARO: Modeling and Processing of High-Volume Uncertain Data Streams

Uncertain data streams, where data is incomplete, imprecise, and even misleading, have been observed in a variety of environments. Feeding uncertain data streams to existing stream systems can produce results of unknown quality, which is of paramount concern to monitoring applications. In this paper, we present the Claro system that supports uncertain data stream processing for data that is naturally captured using continuous random variables. The Claro system employs a unique data model that is flexible and allows efficient computation. Built on this model, we develop evaluation techniques for complex relational operators, including aggregates and joins, by exploring advanced statistical theory and approximation techniques. Our evaluation results show that our techniques can achieve high performance in stream processing while satisfying accuracy requirements, and these techniques significantly outperform a state-of-the-art sampling-based method. Furthermore, initial results of a case study show that our modeling and aggregation techniques can allow a tornado detection system to produce better quality results yet with lower execution time.