Channel Coding Methods for Green Data Storage: Innovations Beyond the Hamming Metric

Lara Dolecek
Assistant Professor, Electrical Engineering, UCLA
Date: June 13, 2013

Abstract

Recent surge in large-scale data storage systems has created an immediate need to develop new coding methodologies to improve the capacity and reliability of such systems. We will discuss several recent results that offer significant performance improvements over the conventional methods by suitably departing from the ubiquitous Hamming metric. In particular, we will show how, at the system level, massive unnecessary redundancy can be eliminated using novel file synchronization methods. We will also demonstrate how, at the device level, new algebraic coding schemes specifically attuned to the physical properties of non-volatile memory devices can offer considerable improvements in lifetime and reliability. The talk will conclude with a discussion of several exciting open problems in this emerging research area.

Biography

Lara Dolecek is an assistant professor in the Electrical Engineering Department at UCLA where she heads the Laboratory for Robust Information Systems and co-directs UCLA Center on Development on Emerging Storage Systems (CoDESS). She received Intel Early Career Faculty Award, University of California Faculty Development Award and Early Career Okawa Research Grant all in 2013, NSF CAREER Award in 2012, Hellman Fellow award in 2011, and David J. Sakrison Award from the EECS Department at UC Berkeley in 2007. She is an Associate Editor for IEEE Transactions on Communications and IEEE Communication Letters, and is the lead guest editor for IEEE JSAC Special Issue on Communication Methodologies for the Next-Generation Data Storage Systems.