Abstract
Edge–cloud services provide heterogeneous virtual machine types to run various IoT workloads. Choosing the appropriate VM configuration for each workload can effectively improve performance and reduce costs. This article proposes ARVMEC, Adaptive Recommendation of Virtual Machines for IoT in Edge-Cloud Environment, which can always provide users with the best VM recommendation according to their own budget or deadline constraints. ARVMEC uses a tree-based ensemble learning algorithm to make accurate predictions on workload performance for all VM types. It can abstract user purposes in a more flexible and general mode, thus offer reasonable recommendations accordingly. Compared to state-of-art methods, ARVMEC can make better predictions with a 15% improvement in accuracy.
•We propose a Gene Generator to characterize IoT workloads in edge–cloud environment by collecting low-level metrics at the virtualized layer.•We build an XGBoost-based model which can make accurate performance prediction for enormous IoT workloads among various VM types in edge–cloud environment.•We abstract user purposes in a more flexible and general mode, thus adaptively make recommendations to meet diverse requirements.