Performance Evaluation of Placement Policies for Cloud-Edge Applications

The applications commonly deployed in cloud-edge environments consist of multiple inter-dependent modules organized according to the Distributed Data Flow model. Decisions about the mapping between these modules and the available resources are quite difficult because of the resource-constrained nature of devices at the edge of the network and the timing requirements of the applications. In this paper we investigate the problem of application placement by proposing a lightweight heuristic that takes into account the volume of data exchanged between modules. In detail, among all possible devices, the proposed policy allocates a given module to the “best” device, that is, the device that ensures the minimum network delay. The policy has been evaluated in a simulated cloud-edge environment based on the iFogSim toolkit. We consider workloads consisting of applications with different demands in terms of processing and data exchanged between modules. The simulation results are promising and indicate that our policy offers competitive advantages when compared to other heuristics.