Queries against SnappyData can be written as conventional SQL or as Spark abstractions, so existing work done in both paradigms can be reused, alone or together, on the same data.