Hadoop: What is it Good For? Absolutely … Something by James Kobielus is an interesting review of how to contrast Hadoop with an enterprise database warehouse (EDW).
From the post:
So – apart from being an open-source community with broad industry momentum – what is Hadoop good for that you can’t get elsewhere? The answer to that is a mouthful, but a powerful one.
Essentially, Hadoop is vendor-agnostic in-database analytics in the cloud, leveraging an open, comprehensive, extensible framework for building complex advanced analytics and data management functions for deployment into cloud computing architectures. At the heart of that framework is MapReduce, which is the only industry framework for developing statistical analysis, predictive modeling, data mining, natural language processing, sentiment analysis, machine learning, and other advanced analytics. Another linchpin of Hadoop, Pig, is a versatile language for building data integration processing logic.
Promoting Hadoop without singing Aquarius, promising us a new era in human relationships, or that we are going to be smarter than we were 100, 500, or even 1,000 years ago. Just cold hard data analysis advantages, the sort that reputations, businesses and billings are built upon. Maybe there is a lesson there for topic maps?