MSTR has extremely flexible schema support - in terms of generating SQL that will work - more restrictions come into play from a performance perspective when working with large data volumes. For example, if architected right, MSTR can support very normalized schema's. but from a performance perspective, price of high number of joins may be prohibitive.