SQL-On-Hadoop Systems: State-Of-The-Art Exploration, Models, Performances, Issues and Recommendations

Analytics of tremendous big data generated from natural systems (e.g. tectonic plates' movement, atmospheric data, ...), engineered systems (e.g. servers, electronic devices, ...), and human activities (e.g. trajectories, web click-streams, health records, customers' transactions, user interactions in social networks, ...) require highly scalable data management systems with new capabilities in both algorithms and architectures. The focus in most of data management systems is on (1) horizontal scalability and high performance, (2) continuous availability, (3) non-structured data processing, and (4) real-time processing. Alternative systems such as SQL-on-Hadoop technologies are becoming mainstream for big data analytics.