In the von Neumann architecture, program and data are stored in the same memory and managed by the same information-handling subsystem. In the Harvard architecture, program and data are stored and handled by different subsystems. This is the essential difference between the two architectures.
However, in some niches, particularly certain embedded applications where the program is more-or-less hard wired, task requirements are such that the Harvard architecture can provide distinct operational advantages. Under certain conditions, a Harvard computer can be much faster than a von Neumann computer because data and program do not contend for the same information pathway, and storing the program in an immutable read-only memory can result in vast reliability improvements.