Fog resource selection using historical executions

As an emergent technology, IoT promises to harness the computational and storage resources distributed across different remote clouds. Fog computing extends cloud computing by bringing the network and cloud resources closer to the network edge. Those resources are typically heterogeneous in nature requiring careful management. As the number of resources contributing to a cloud grows, so the problems associated with efficient and effective resource selection and allocation increase. In this paper, we introduce a fog resource selection algorithm (FResS) that enables automated fog selection and allocation for IoT systems. The proposed model collects and maintains a repository of fog performance data in the form of execution logs stored in standard formats. When a new task needs to be executed, prediction for its run-time is made by using these logs, which results in a realistic run-time estimate as well as best fog selection for task placement. Simulation results show a decrease in the overall end-to-end (e2e) latency of the system.