Harvard architecture is a modern alternative to von Neumann architecture which allows the computer to read data faster and more effectively, in a way that von Neumann architecture is incapable of. The main differences between the two types of computer architecture are related to the functioning of the central processing unit.
In von Neumann architecture, the CPU cannot read an instruction and perform a function at the same time, while Harvard architecture makes this possible. In a Harvard architecture-based computer system, instructions are typically stored in read-only memory, while von Neumann stores instructions and data in the same bus system.
Since this architecture stores information along the same memory pathway, von Neumann architecture obviously runs much more slowly than the Harvard equivalent. Harvard architecture, essentially, has a more complex circuitry. Its CPU is capable of performing more tasks at once and this is a distinct advantage over the slower, older von Neumann architecture.
Any discussion about computer architecture throughout history is incomplete without a reference to von Neumann architecture, since it was the very first format, and most computers have been built using this architecture. In the future, it is likely that more computers will be built with the Harvard architecture, as it is capable of more.