Coding Techniques for Data-Storage Systems

  • Yuval Cassuto | Department of Electrical Engineering California Institute of Technology

Traditional Coding Theory, inspired by Shannon’s motivations in communication problems, has been very successful in proposing meaningful abstractions (e.g Hamming distance, code rate, decoding complexity) and perfecting code construction for those abstractions. Data-storage applications, however, exhibit unique behaviors and constraints that are often addressed by either inefficient adaptations of known codes, or by ad-hoc solutions. In this lecture, examples of more methodical treatments of data-storage coding techniques are presented. In one part of the lecture, the unique error characteristics of Multi-Level Flash Memories are used to abstract a new error model; codes that are efficient in both redundancy and implementation are constructed, and their performance is validated using analytical tools and experiments on real floating gate arrays. Another part of the lecture introduces codes for a new well-motivated failure model of disk arrays: “Clustered Failures”, which are failure combinations that fall into a limited number of contiguous clusters. Other code constructions for storage systems will be shared with the audience as time permits.

Speaker Details

Yuval Cassuto is a recent PhD graduate of the department of Electrical Engineering, Caltech. His main research area is error-correcting codes for storage applications. Other research interests include algebraic and combinatorial coding theory and Network Coding. Before joining Caltech he held a System Engineer position in Qualcomm. His undergraduate project “Real-Time Digital Watermarking for Audio Signals” has won the first prize in the 2001 Texas Instruments $100,000 Worldwide DSP Challenge.

Watch Next