An Efficient Sensitivity Analysis Method for Large Cloud Simulations
Kevin L. Mills
We describe Koala, an infrastructure Cloud simulator inspired by the Amazon Elastic Compute Cloud (EC2). We conduct a sensitivity analysis of Koala, revealing eight behavioral dimensions, which are influenced significantly by six parameters. Our findings, which set the stage for subsequent experiments, show that proposed resource allocation algorithms for on-demand infrastructure Clouds can be evaluated by simulating 64 parameter configurations, which can be reduced further by using orthogonal fractional factorial experiment designs. We expect that our results could establish baseline configurations against which proposed resource allocation algorithms can be evaluated.
IEEE Cloud 2011, The 4th International Conference on Cloud Computing