iJoin: Importance-Aware Join Approximation over Data Streams

Abstract

We address approximate join processing over data streams when memory limitations cause incoming tuples to overflow the available memory, precluding exact processing. Moreover, in many real-world applications such as for news-feeds and sensor-data, different tuples may have different importance levels. Current methods pay little attention to load-shedding when tuples bear such importance semantics, and perform poorly due to premature tuple drops and unproductive tuple retention. We propose a novel framework, called iJoin, which overcomes these drawbacks, maximizes result importance, and has the best performance compared to earlier work.