Information Technology Reference
In-Depth Information
8.3 Reliability of a simulation environment
Besides improving performance (i.e., reducing runtime) parallelization
may also be used for increasing reliability. The underlying computing
hardware is utilized by a simulation engine and
invisible
for the
actual model that is executed. Instead of using all available resources
for achieving maximum speedup, a simulation engine might exploit
some hardware resources for increasing reliability of the simulation
environment.
The importance of reliability increases with the amount of hardware
resources utilized by a simulation engine. Therefore, especially if
huge computing systems are considered (e. g., containing hundreds of
processors) dealing with reliability is essential. In the same way as a
simulation engine can use available hardware resources for performance
reasons, a simulation engine may as well utilize a portion of the
resources for reliability issues.
The actual strategies of a simulation engine for improving reliability
may be very different. Possible options include introducing redundancy
as well as providing fault-detection and fault-correction mechanisms
(cp. [49]).
