Parallel Star Join Query Processing in Column-Oriented Data Storage

An increasing number of companies rely on the results of massive data analytics for critical business decision. Such analysis is crucial to improve service quality, increase revenue, support novel features, and detect changes in patterns over time. We have designed, implemented and evaluated a series of parallel and distributed algorithms that can process large volumes of data and provide ad-hoc real-time analysis without compromising performance. We have formulated the cost models for these algorithms and instantiated and evaluated them.