Harvard architecture is the architecture that provides separate buses for program and data memory.

Explanation:
The Harvard architecture is a computer architecture that separates the storage and manipulation of program instructions and data. It was named after the Harvard Mark I computer, which introduced this architecture in the 1940s.

In the Harvard architecture, the program memory and data memory are physically separate and have separate buses for accessing them. This separation allows the processor to fetch instructions and data simultaneously, which can lead to improved performance.

Key Points:
- Program Memory: In the Harvard architecture, program instructions are stored in a separate program memory. This memory is dedicated solely to storing the instructions that the processor will execute.
- Data Memory: Data memory, on the other hand, is used to store data that the processor manipulates during its execution. It is also separate from the program memory.
- Separate Buses: The Harvard architecture provides separate buses for accessing program memory and data memory. This means that the processor can fetch instructions and data simultaneously, improving performance.
- Advantages: The separation of program and data memory in the Harvard architecture allows for more efficient memory access. It also reduces the chances of conflicts between instructions and data, as they are stored in separate memories.
- Disadvantages: One disadvantage of the Harvard architecture is that it requires more physical memory, as both program memory and data memory need to be implemented separately. This can increase the cost and complexity of the system.

Conclusion:
In summary, the Harvard architecture provides separate buses for program and data memory. This separation allows for simultaneous access to instructions and data, improving performance. However, it also requires more physical memory, which can increase cost and complexity.