The Evolution of Hadoop and Spark

Matt: I like the idea of you have a single data store, and then you just apply different engines to it. And the engine that you use the engine that's most appropriate for the job is going to be different than the engines for analytics. So if we can do that all in one platform, it just makes your life so much easier and you get away from the evil of moving data. Tim Bergwind: It more came down to how programmable do you need your queries to be? Can you use a set oriented language sequel coming from math? Do you need something that's highly programmable where you can program your own steps in a bag and then you need to go to spark