Workflow Scheduling in the Cloud-Edge Continuum

Scheduling in the cloud-edge continuum is a challenging problem. In fact, scheduling has to cope with the peculiarities of these complex ecosystems and satisfy at the same time the desired service levels. In this paper, we investigate the benefits of the cloud-edge continuum for deploying workflows with different characteristics, e.g., computation or communication-intensive. In detail, we formulate a multi-objective optimization problem solved using a Genetic Algorithm. This problem is aimed at identifying the scheduling plans that minimize two conflicting objectives, namely, the expected workflow execution time and monetary cost associated with the cloud and edge resources to be provisioned. Our experiments have shown that the plans that exploit both cloud and edge resources represent a good tradeoff between the two objectives. In addition, the workflow characteristics strongly influence these plans. Similarly, the uncertainties that might affect the infrastructure performance are responsible of significant changes in the corresponding Pareto fronts.