Jump to content United States-English
HP.com Home Products and Services Support and Drivers Solutions How to Buy
» Contact HP

HP.com home


Information Theory Seminar


printable version
» 

HP Labs

» Research
» News and events
» Technical reports
» About HP Labs
» Careers @ HP Labs
» People
» Worldwide sites
» Downloads
Content starts here

TITLE: Error Correction Coding for Flash Memories

SPEAKER: Eitan Yaakobi (UC San Diego)

DATE: 10:00 - 11:00 AM, Friday, August 14, 2009

LOCATION: Sigma, 1L

ABSTRACT:
Non-volatile, solid-state NAND-flash memory devices are finding use in an increasing number of computing and consumer electronic devices. They have replaced hard drives in many of these applications because of their high data-transfer rates, mechanical durability, and low power consumption. To date, flash memory devices have used only low-redundancy ECC codes with minimal error correction and detection capabilities, such as single bit-error correcting Hamming codes and basic error-detecting cyclic redundancy check (CRC) codes. The demand for increased storage capacity, coupled with the introduction of multilevel cell (MLC) flash technology, has created the need for more powerful ECC methods, such as BCH codes and Reed-Solomon (RS) codes.

In this work, we use an extensive empirical database of errors induced by write, read, and erase operations to develop a comprehensive understanding of the error mechanisms and error characteristics of flash memories. We also use the error measurements to compare the performance of BCH and RS codes over a range of redundancies and error-correction capabilities when applied at the page level. In addition, we examine the effectiveness of interleaving techniques and burst-error correcting codes, as well as asymmetric-error correcting schemes that can improve performance if there is a bias in the direction of errors (i.e., from '0' to '1' or from '1' to '0'). The potential benefit of coding across multiple pages within a block is also investigated. One of the distinctive properties of flash memory is its read-write asymmetry. That is, reading an individual page is easily implemented, whereas rewriting a previously stored page requires the erasure of the entire block containing the page, followed by the rewriting of the entire contents of the block, including the updated page. This erase-rewrite operation incurs a substantial cost in time and power consumption. Moreover, flash memory performance is degraded by repeated block erasures, thereby limiting the lifetime of the device. Nowadays, it is of high interest to design codes allowing information to be written multiple times before erasing the memory. We will survey some of the recent results related to this topic and discuss them from the practical and theoretical points of view.

BIOGRAPHY:
Eitan Yaakobi received the B.A. and M.Sc. degrees from the Technion - Israel Institute of Technology, Haifa, Israel, in 2005 and 2007 respectively, from the Computer Science Department. He is currently a Ph.D. candidate in Electrical and Computer Engineering at the University of California, San Diego, where he is associated with the Center for Magnetic Recording Research. His advisors are Prof. Paul Siegel, Prof. Alexander Vardy, and Prof. Jack Wolf. His research interests include algebraic error-correction coding, coding theory, and their applications for digital data storage, and in particular for flash memories.

Seminars

» Information Theory
» Publications
» People
» Discrete Universal Denoiser (DUDE)
» Elliptic Curve Cryptography
» Image Compression
» Seminars
» Related Links
This is a controller for a color printer. Each chip contains a compressor/decompressor based on an algorithm created by HP Labs.
Privacy statement Using this site means you accept its terms Feedback to HP Labs
© 2009 Hewlett-Packard Development Company, L.P.