Recent laboratory demonstrations have shown that time-domain persistent spectral holeburning memory can achieve record areal densities and density-bandwidth products. The same work has shown that many kilobits of data can be spectrally multiplexed within single spatial storage locations. These factors, taken together, open the door to an entirely new class of optical data storage, i.e., read/write Optical Random Access Memory (ORAM). ORAM is expected to provide 10-100 Gigabyte single-unit storage capacities, data transfer rates in the Gigabit/sec range, and achieve microsecond-scale random access latency through entirely non-mechanical operation. As the name implies, ORAM resembles semi-conductor memory in terms of its speed. However, ORAM resembles magnetic disk memory in terms of single-unit capacity. ORAM provides a unique joining of the attributes normally associated with semiconductor and magnetic disk storage, and, at the Gigabyte capacity levels of interest, ORAM should be at a distinct cost advantage relative to semiconductor RAM. ORAM offers the potential for development to speeds even beyond those afforded by semiconductor devices. Furthermore, the optical foundation of ORAM allows for the incorporation of holographic parallelism in later evolution of the technology. This proposal is aimed at the development of high speed (Gigabit/sec serial transfer rates) 10-100 Gigabyte capacity Optical Random Access Memory devices. These devices will greatly enhance the computational capabilities in cases where simultaneous speed and capacity are required.
Keywords: Optical Random Access Memory, Optical Memory, Gigabyte Storage, Spectral Holography, Random Access M
