DRESS Code for the Storage Cloud
Speaker: Dr. Salim El Rouayheb
Faculty Host: Professor Elza Erkip
Abstract
Cloud storage systems, such as Google data centers, Amazon S3, Wuala, etc. are now a growing paradigm for providing reliable and secure “on-demand” data storage services. Disk failures are frequent in these large-scale systems and constitute the norm rather the exception. Erasure codes can be used to guarantee the data reliability. However, a major challenge that arises here is how to repair the system and regenerate the lost data with a minimum system overhead in terms of bandwidth, latency, disk I/O calls, computational complexity, etc.
In this talk, I will address the above problem and present new efficient codes for distributed storage that we call Distributed REplication-based Simple Storage (DRESS) codes. DRESS codes have linear encoding/decoding complexity and permit fast and “uncoded” repair of failures with minimum bandwidth and disk reads overhead. The design of optimal DRESS codes translates into interesting combinatorial problems with many open questions. We present optimal code constructions based on projective planes and Steiner systems. We also propose simple randomized constructions for scalable DRESS codes using bins-and-balls models. When the security in the cloud is breached and some nodes start acting maliciously, our codes do not only guarantee data integrity, but also help catch the bad guys.
This is a joint work with Sameer Pawar, Nima Noorshams, and Kannan Ramchandran.
About the Speaker
Salim El Rouayheb is a postdoctoral research fellow with the Electrical Engineering and Computer Sciences (EECS) Department, University of California, Berkeley. His research interests lie in the broad area of communications with a focus on reliable an secure distributed information systems and on the algorithmic and information-theoretic aspects of networking.
He received his Diploma degree in electrical engineering from the Lebanese University, Roumieh, Lebanon, in 2002, and his M.S. degree in computer and communications engineering from the American University of Beirut, Lebanon, in 2004. He received his Ph.D. degree in electrical engineering from Texas A&M University, College Station, in 2009. During Summer 2006, he was an intern at the Mathematics of Communication Research Department at Bell Labs. He received the Charlie S. Korban award for outstanding graduate student, and the Texas Telecommunication Engineering Consortium (TXTEX) Graduate Fellowship.