EEO: an Efficient MDS-like RAID-6 Code for Parallel Implementation

Description:

Extension of the Even-Odd (EEO):  RAID-6 Code for Parallel Implementation

Description:

Data storage technology has been recognized as one of the main dimensions of modern information technology. Traditionally, Redundant Array of Independent Disks (RAID) systems provide fault tolerance and capacity for recovery by making use of data redundancy. Among conventional RAID systems, RAID-6 based on the "Exclusive Or" (XOR) operation has better performance than others.

Binghamton University’s Extension of the EVEN-ODD (EEO) coding scheme is an efficient Maximum Distance Separable (MDS) coding scheme that significantly improves the performance of decoding procedure with negligible overhead. Based on extensive testing, Binghamton University’s EEO scheme was proven to be the fastest double-disk failure recovering algorithm implemented in RAID-6 systems to date.

Potential applications:

Companies running large data centers can use this EEO technology to dramatically boost the availability and the performance of their system.

Advantages:

The EEO coding scheme provides several advantages over current technology:

1. A stable speed boost up to four times faster than any commercially available technology.

2. First technology of its class that can be implemented in systems using parallel computing and quad-processors.

3. Storage systems utilizing the EEO coding scheme will benefit further as performance improvements in I/O throughput technology reaches the market.

4. Can be applied to any current large-scale distributed-data storage system based on RAID-6.

Development status:

Pre-alpha stage. Tests were performed using three computers, running on a single-, a dual-, and a quad-processor.

Patent status:

US Patent pending.

Publication:

J. Feng, Y. Chen, D. Summerville, "EEO: an Efficient MDS-like RAID-6 Code for Parallel Implementation," the 33rd IEEE Sarnoff Symposium, Apr. 12-14, 2010, Princeton, NJ, USA http://pods.binghamton.edu/~ychen/PID1141099.pdf

 

Keywords:

XOR, RAID-6, Fault-tolerant, Storage System, Cloud Storage

Inventors:

Jun Feng is a Ph.D. candidate of the Department of Electrical and Computer Engineering at the State University of New York (SUNY) at Binghamton. His research interests include information assurance, fault tolerant data recovery, security in Cloud Computing, and network security. He is a member of the IEEE.

Dr. Yu Chen is an Assistant Professor of Electrical and Computer Engineering at the State University of New York – Binghamton. He received the Ph.D. in Electrical Engineering from the University of Southern California (USC) in 2006 under the supervision of Professor Kai Hwang. His research interests span network security; distributed/grid computing; and, computer architectures. Particularly, his work covers security in networks and large distributed/grid computing systems; internet security protocols; infrastructure security; trust, security and privacy in wireless, embedded and pervasive computing. He has authored or coauthored more than 50 scientific papers in refereed journals, conferences, and book chapters. He is a member of ACM, IEEE (Computer & Communication Societies) and SPIE. Jun Feng is a Ph.D. candidate of the Department of Electrical and Computer Engineering at the State University of New York (SUNY) at Binghamton. His research interests include information assurance, fault tolerant data recovery, security in cloud computing, and network security. He is a member of the IEEE.

Additional Information :

More information regarding Professor Chen and his research program is available at:  http://pods.binghamton.edu/~ychen/ and http://www2.binghamton.edu/watson/programs/academic-departments/electrical-and-computer-engineering/people/faculty_staff_profiles/yu-chen.html

Patent Information:
For Information, Contact:
Scott Hancock
Director, IP Management and Licensing
Binghamton University
(607) 777-5874
shancock@binghamton.edu
Inventors:
Yu Chen
Jun Feng
Keywords:
© 2017. All Rights Reserved. Powered by Inteum