“The key innovation of MapReduce is the ability to take a query over a data set, divide it, and run it in parallel over many nodes. This distribution solves the issue of data too large to fit onto a single machine.”