The continuing explosion in data sources and volumes strains and exceeds the scalability of traditional data management and analytic systems. At the same time, a new generation of analytics has emerged with requirements that are difficult or impossible for traditional architectures to meet: analysis of large data volumes, ultra-fast results, and deep data exploration through ad-hoc and interactive analysis. The demand for advanced analytic applications including fraud detection, customer behavior analysis, forecasting, scenario modeling and deep click-stream analysis requires a solution that overcomes the scalability, performance and complexity limitations of traditional approaches.