Enterprise Hadoop Best Practices: Concrete Guidelines From Early Adopters In Online Services

Today's big data challenges require innovative approaches for extreme volumes, velocities, varieties, and variabilities of data for advanced analytics. Increasingly, high-tech early adopters are turning to Hadoop, despite challenges with the technology — in particular, its unfamiliarity, immaturity, complexity, lack of clear standards, and dearth of a wide range of commercial offerings. Best practices for Hadoop early adopters are still emerging, but focus on: 1) identifying advanced analytics applications that can benefit from petabyte scalability, 2) adopting and optimizing the open source Apache Hadoop codebase for specific requirements, 3) implementing Hadoop as a massive staging layer behind an enterprise data warehouse (EDW), and 4) licensing commercial Hadoop distributions for security, reliability, real-time performance, and other requirements that the open source distributions do not yet fully address.