Half-Product Codes for Flash Memory

Abstract: A product code is a set of codeword matrices such that the each row is a codeword of a row code and each column is a codeword of a column code. They have been studied for several years in the literature and they possess computationally efficient encoding and decoding algorithms and they can also have reasonably low error floors. In this talk, we will consider a variant of product codes called half product codes which are obtained by enforcing symmetry to the codeword matrices. In contrast to product codes, half-product codes have not been studied extensively in the literature. In this talk, we will review the construction of half-product codes and present efficient algorithms for encoding and decoding such codes. We will present results on the minimum distance, stopping set sizes, and an estimate of error floor for these codes. These results show that Half product codes can perform better than product codes for the same length and dimension for some range of parameters.

Bio: Krishna Narayanan is a professor in the ECE department at Texas A&M University. His interests are in coding and signal processing for data storage and wireless communications. He is a Fellow of the IEEE and a recipient of the 2006 best paper award in data storage from IEEE comsoc. He served as the area editor for coding theory and its applications for the IEEE Transactions on Communications from 2007-2011.