A Situation Enabled Framework for Energy-Efficient Workload Offloading in 5G Vehicular Edge Computing

A Situation Enabled Framework for Energy-Efficient Workload Offloading in 5G Vehicular Edge Computing

연구 분야: Networking

학회: 2020 IEEE World Congress on Services (SERVICES)

Cloud based vehicular edge computing is a promising technology to deliver quality of services for accessing public cloud from vehicles. As in-vehicle computers (e.g. onboard Android devices) are typically energy and resource constrained, they need to operate energy-efficiently. Due to the high cost in building sustainable infrastructure to support vehicular edge computing systems, edge services must also operate with high energy-efficiency for better acceptability and applicability in the market place. In this paper, we present a novel situation-enabled framework to enhance energy-efficiency of both in-vehicle computer applications and Mobile Edge Computing (MEC) services. The framework consists of three components. First, we collect a user's in-vehicle driving situation data and use the data to train Long Short Term Memory (LSTM) deep learning model. The model is used to predict the user's future situation, determine whether the requested application is allowable, and manage the allowable application pertaining to the predicted situation. Second, from vehicles to MEC servers, we optimize the energy consumption in request routing. Third, a Breadth First Search (BFS) based offloading algorithm to coordinate the placement of servers in the MEC servers to save the energy consumption in the server pool. Implementation of LSTM on top of a situation model and simulation of the vehicular edge offloading have been conducted to validate the energy efficiency of the proposed framework. The results show that the performance of our proposed offloading algorithm can be at par with the most aggressive energy saving variant.