Big interest, big money at stake
The Spark vision is certainly captivating and getting lots of attention. I'm anxious to talk to users about ease of use, consistency across the framework and library-by-library capabilities. Can it be best of breed in each category of analysis? I have my doubts. Expect commercial vendors, in particular, to start poking holes and leveling criticisms. How broad, standards-based and up to data is the SQL? How many algos? Is R highly scalable? I have no idea what the answers are, but a lot of money is at stake, so expect to hear lots of questions and assertions design to sow seeds of fear, uncertainty and doubt.