Thrashing and Performance Analysis
Thrashing and Performance Analysis
Copyright(c) Management Analytics, 1995 - All Rights Reserved
Copyright(c), 1990, 1995 Dr. Frederick B. Cohen - All Rights Reserved
Problem:
Systems often perform poorly under load conditions they are not
configured for. For example, if the physical memory is small and there
are a lot of processes, there may be a lot of paging of memory to disk
and back. This is called thrashing. The same system may perform very
well with a slightly different load distribution. An attacker can exploit
this to cause denial of services.
Prevention:
This problem can only be prevented by providing excess resources,
which costs more money than a more prudent amount.
Detection:
Thrashing can be detected by slow system performance, and analyzed
by the performance analysis tools provided with your system.
Cure:
Adjustments are often required in order to keep performance
high. For example, by reducing the size of the internal tables, adding
more physical memory, getting a faster disk, modifying the usage
patterns, or modifying key system facilities, we might dramatically
change performance.