Mainframe Central Processor Unit (CPU) speed/performance has increased by more than two orders of magnitude since the 1970's. However, the I/O subsystem performance has only quadrupled in the same period of time. There is a need for a computer architecture with an effective storage hierarchy to complement improvements made in the CPU performance. This paper presents a technique to calculate the I/O performance of disk subsystems with and without disk cache buffers. Through a series of calculation it is shown that the best way to improve I/O performance is to increase the number of data paths and use disk memory caching.