Multiobjective optimization for brokering of multicloud service composition

The choice of cloud providers whose offers best fit the requirements of a particular application is a complex issue due to the heterogeneity of the services in terms of resources, costs, technology, and service levels that providers ensure. This article investigates the effectiveness of multiobjective genetic algorithms to resolve a multicloud brokering problem. Experimental results provide clear evidence about how such a solution improves the choice made manually by users returning in real time optimal alternatives. It also investigates how the optimality depends on different genetic algorithms and parameters, problem type, and time constraints.