Energy Efficient Resource Allocation in Vehicular Cloud based Architecture

TL;DR

This paper proposes an energy-efficient resource allocation architecture for vehicular clouds, integrating metro fog nodes and central cloud, using MILP to optimize power consumption and improve resource utilization.

Contribution

It introduces a novel architecture combining vehicular clouds with metro fog and cloud, and develops an MILP model for energy-efficient task assignment.

Findings

Traffic demands significantly impact power consumption.

Integrating metro fog nodes can save up to 54% power.

Distributing tasks among multiple vehicles increases power savings by 12%.

Abstract

The increasing availability of on-board processing units in vehicles has led to a new promising mobile edge computing (MEC) concept which integrates desirable features of clouds and VANETs under the concept of vehicular clouds (VC). In this paper we propose an architecture that integrates VC with metro fog nodes and the central cloud to ensure service continuity. We tackle the problem of energy efficient resource allocation in this architecture by developing a Mixed Integer Linear Programming (MILP) model to minimize power consumption by optimizing the assignment of different tasks to the available resources in this architecture. We study service provisioning considering different assignment strategies under varying application demands and analyze the impact of these strategies on the utilization of the VC resources and therefore, the overall power consumption. The results show that traffic demands have a higher impact on the power consumption, compared to the impact of the processing demands. Integrating metro fog nodes and vehicle edge nodes in the cloud-based architecture can save power, with an average power saving up to 54%. The power savings can increase by 12% by distributing the task assignment among multiple vehicles in the VC level, compared to assigning the whole task to a single processing node.