Simulation of a workflow execution as a real Cloud by adding noise.

•A new model and methodology to add a noise to workflow execution simulations.•Noisiness of tasks within a workflow and environmental noisiness over a workflow.•Up to 59 times simulation inaccuracy reduction of makespan’s mean value.•Instance type impacts the cloud performance instability.•Sensitivity analysis determines minimum data to be collected in Learning Phase in order to achieve accuracy and precision.