We created a demo, http://demo.snaplogic.org/griddemo.html of the data mashup displayed in the Dojo Grid that uses a large data set and manipulates it entirely via SnapLogic framework. Data residing on the server or multiple servers is fed into SnapLogic's easy-to-create - either in GUI designer or Python - Pipelines that allow various manipulations and transformations (filtering, sorting, joining, etc.) and expose the results via RESTful interface.