Regenerating codes for Distributed Storage

Seminar Speaker: Alex Dimakis, Wireless Foundations, EECS, UC Berkeley
11-12 pm November 20, 2007
Host: Bhaskar KrishnamachariRegenerating codes for Distributed StorageData centers and sensor networks require reliable information storage
over individually unreliable nodes. Storing a file using an erasure
code, in fragments spread across nodes, requires less redundancy than
simple replication for the same level of reliability. However, since
fragments must be periodically replaced as nodes fail, a key question
is how to generate a new encoded fragment in a distributed way while
transferring as little data as possible across the network.
We formulate code repair as an optimal flow problem on an infinite
graph and show how appropriate mixing of information (network coding)
can surprisingly reduce the repair bandwidth. We present both
information theoretic lower bounds and code constructions that achieve
optimal performance. Time permitting, some open problems related with
information storage in Data centers will be discussed.Bio: Alex Dimakis received the Diploma degree in Electrical and
Computer Engineering from the National Technical University of Athens
in 2003 and the M.S. degree in Electrical Engineering from UC Berkeley
in 2005. He is currently a Ph.D. candidate in the department of Electrical
Engineering and Computer Sciences at UC Berkeley.
His research interests include Communications, Signal processing, and
Networking with applications in distributed systems and sensor
networks. Mr. Dimakis has received two outstanding paper awards, the UC Berkeley
Departmental Fellowship in 2003, and the Microsoft Research Fellowship
in 2007.

Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248

Audiences: Everyone Is Invited

Contact: Shane Goodoff