Skip to main content

Advances, Systems and Applications

Journal of Cloud Computing Cover Image

Table 3 Host overload detection algorithms

From: Proactive dynamic virtual-machine consolidation for energy conservation in cloud data centres

Algorithms Characteristics
Static utilization threshold [21, 100]   Depend on the mean of the last CPU used.
Unsuitable for dynamic and unpredictable workload
Adaptive utilization threshold [18] Median Absolute Deviation Depend on statistical dispersion
Similar to the static threshold
Inter-quartile range [130] Same as Median but compare third and first quartiles
Provide poor prediction of host overloading
Regression based algorithms [1, 17] Local regression algorithms By fitting simple models to CPU utilization
regression robust Use regression to predict the future CPU utilization
Markov overload detection Add constraint in estimation increase the computation