datanami.com
While I admit that I have mostly moved on to Spark, for many years Hadoop was a reliable way to process lots of data inexpensively.
Reply on news.ycombinator.com