Probability calculations in fault-tolerant computing: simple RESTART, checkpointing and wallclock models
by Professor Soren Asmussen
Abstract: A job like the execution of a computer program or the transfer
of a file, with ideal execution time T, may fail and then needs
to be restarted. Thus the actual execution time X is at least T.
We look at the tail asymptotics for X both in the simple RESTART
setting just outlined, when checkpoints are build into the task,
and when failures are governed by the wallclock rather than
by the time since the last failure.
For More Information: contact: Owen Jones. email firstname.lastname@example.org