No idea which engine yet, we are just investigating ways to improve the speed of the result set.

Anything from NoSQL to the use of CUDA programming.