A Tensor-Based Holistic Edge Computing Optimization Framework for Internet of Things

Balancing the costs of different objectives in EC requires comprehensive and global analysis. This article investigates the holistic EC optimization problem for IoT. First, a triple-plane EC architecture for IoT is proposed including the edge device plane, edge server plane, and cloud plane, respectively, which is conducive to collaboratively accomplishing the EC applications. Then five tensor-based representation models are constructed to represent the complex relationships and resolve the heterogeneity of different devices. Afterward, we construct a generalized and holistic EC optimization model based on the constructed tensors including energy consumption, execution time, system reliability, and quality of experience. Finally, a customized optimization framework is proposed in which the optimization objectives can be arbitrarily combined according to practical applications. A case study is conducted to evaluate the performance of the proposed scheme; results demonstrate that it significantly outperforms the state-of-the-art cloud-assisted mobile computing scheme and holistic mobile cloud computing scheme.