Thursday, March 27 , 2008
12:00 pm - 1:00 pm
Wean Hall 8220
Reconfiguring a State Machine: A Tutorial
Replicated state machine is a well-known approach to fault tolerance in
distributed systems. Protocols that implement a replicated state machine
often assume a fixed configuration, where a configuration specifies the set
of processes executing the state machine. Reconfiguration can be performed
to reduce the vulnerability to further failures after a process has failed
or to replace hardware without shutting down the system.
The basic idea for state-machine reconfiguration is simple and known: use
the state machine itself to maintain the configuration; yet the implications
of this idea are not well understood. In this talk, we explain several
methods for reconfiguring a system implemented using the state-machine
approach, including some new ones.
This is based on joint work with Leslie Lamport and Dahlia Malkhi at
Microsoft Research Silicon Valley.
Lidong Zhou has been affiliated with Microsoft Research Silicon Valley since
2002. He obtained his M.S. and Ph.D. in Computer Science from Cornell
University. His main research interests are on theory and practice of
distributed system, system security, and computer networks.
or visit http://www.pdl.cmu.edu/SDI/