Spark and Hadoop - What's the Difference?

Spark is an evolutionary step on top of the MapRedis paradigm that was introduced by Yahoo and a bunch of universities when they created the Hadoop system. Instead of it being a disk-based operation system, Spark's effectively issuing mapping commands over files located in a file store. Because everything is in memory, this opened up amazing possibility for large-scale computing tasks. The optimization comes in when you're actually having to do things that are not a demonstration, or going through a tutorial, real-world problems.