Thursday, June 10, 2004
Noon - 1 pm
Hamerschlag Hall D-210
Efficient Byzantine-tolerant Erasure-coded Storage
In this talk, I describe a decentralized consistency protocol for
survivable storage that exploits local data versioning at each
storage-node. The protocol efficiently provides strong consistency
guarantees for reading and writing erasure-coded data in a Byzantine
environment. Measurements of a prototype storage system based on the
protocol demonstrate scalability in terms of the number of
storage-node failures tolerated and excellent throughput.
Jay is a 5th year Ph.D. student in the Parallel Data Laboratory. He is
interested in storage systems, distributed systems, scalability of
systems, and fault-tolerance.
Seminar Info Contact:
or visit http://www.pdl.cmu.edu/SDI/