Optical Processing Techniques for High Speed Searching and Matching Applications
Project Description
Searching in its various forms has emerged as a fundamental and decisive operation in many of these applications including database processing, artificial intelligence, data mining, security, high-performance computing, and communications. Searching and table lookup can be significantly improved by the use of the content-addressable memory ( CAM ) processing model instead of the location-addressable memory model. CAM-based processing is not only more akin to the way search operations are interpreted in nature (in parallel) but it is also faster than location-addressing (RAM) schemes since the overhead cost of address computations is completely eliminated. A fast emerging technology that has the potential to provide adequate and efficient implementation of search operations in general, and CAM-based processing systems in particular, is optical information processing technology. In this research, we propose to investigate the development and detailed analysis of novel integrated guided-wave photonic search engines for parallel and fast search operations that can provide a quantum leap improvement in performance and size reduction. These engines can be considered as co-processing assists within a larger electronic processing node for computation and/or communication or as stand-alone photonic search accelerators.