New Coding Paradigms for Data Storage

Yuval Cassuto
Hitachi Global Storage Technologies

Thusrday, May 27, 2010 at 1:00pm
Engr IV Faraday Room 67-124

Abstract
Information theory, as a field of study, was presented by Shannon as a theory of communication. And indeed, the two most studied information theoretic channels - the binary symmetric channel (BSC) and binary erasure channel (BEC) - provided suitable abstraction of communication channels that led to highly effective coding schemes. An important component in the theory of error-correcting codes is the Hamming distance metric, which naturally captures the equivocation induced by both the BSC and BEC. In comparison, data-storage channels tend to be more complex than these classical channels, and require developing coding theories beyond the Hamming metric.

In the talk we will discuss two contemporary storage channels that open rich coding-theoretic research pastures. One is multi-level Flash memories that lead to a theory of asymmetric limited-magnitude error-correcting codes. The other is patterned magnetic storage media that motivate a new study of codes for channels whose outputs are overlapping symbol-pairs. The talk will have non-negligible components on both the theoretical and practical axes, so audience with both inclinations should be comfortable in attendance.

Biography
Yuval Cassuto is a Research Staff Member at Hitachi Global Storage Technologies, San Jose Research Center. His research focuses on error-correcting codes for data-storage channels, algebraic and combinatorial coding theory, and information theory. His research also spans to the architecture domain, with interests in storage performance, reliability and security.

He received the B.Sc degree in Electrical Engineering, summa cum laude, from the Technion, Israel Institute of Technology, in 2001, and the MS and Ph.D degrees in Electrical Engineering from the California Institute of Technology, in 2004 and 2008, respectively.

From 2000 to 2002, he was with Qualcomm, Israel R&D Center, where he worked on modeling and analysis of physical-layer communication principles.

Dr. Cassuto was co-awarded the 2001 Texas Instruments Worldwide DSP and Analog Challenge $100,000 award, as well as the Powell and Atwood graduate research fellowship awards.