Cost optimization heuristics for deadline constrained workflow scheduling on clouds and their comparative evaluation

摘要

Nowadays, cloud service providers usually offer users virtual machines with various combinationsrnof configurations and prices. As this new service scheme emerges, the problem of choosingrnthe cost-minimized combination under a deadline constraint is becoming more complex for users.rnThe complexity of determining the cost-minimized combination may be resulted from differentrncauses: the characteristics of user applications and providers' setting on the configurations andrnpricing of virtual machine. In this paper, we proposed an algorithm with two variants to help thernusers to schedule their workflow applications on clouds so that the cost can beminimized and therndeadline constraints can be satisfied. The proposed algorithm is evaluated by extensive simulationrnexperiments with two realistic workflows.