Three-dimensional error correcting codes for volume holographic data storage

Abstract

Volume holographic data storage represents a potential technology path to meet the storage intensive application needs of the information age. Important features of these systems are large storage capacity and fast transfer rate. These storage systems store and retrieve data as two-dimensional data pages and are also referred to as page oriented memories. Conventional storage systems such as magnetic and optical disks store and retrieve data serially. Noise sources encountered in holographic data storage include cross talk, optical and electrical noise sources. During data storage and retrieval, these noise sources interact to produce several error types such as random single bit errors, burst errors and cluster errors. Two-dimensional and three-dimensional error correcting codes have been proposed to perform the data encoding and decoding required by page oriented memories. In this dissertation, we evaluate three-dimensional error correcting codes based on array codes for volume holographic data storage. This three-dimensional encoding scheme, also called z-encoding, provides error detection and correction for error events which may span multiple pages. The attributes of two-dimensional array codes are reviewed and adapted for use as three-dimensional array codes. These codes are analyzed regarding hardware requirements, timing delay and code performance. The row and column array code is found to be effective on random single bit errors. This code is adapted for use as a burst error correcting code and adapted again for use as a cluster error correcting code. Interactions between array code attributes, storage capacity and decode time are analyzed. The attributes of array codes which allow it to be used either as an equal error protection code or an unequal error protection code are evaluated. An algorithm for converting an equal error protection code to an unequal error protection code is described.