Cloud Offloading for Multi-Radio Enabled Mobile Devices

TL;DR

This paper proposes an energy-efficient computation offloading framework for multi-radio enabled mobile devices, optimizing data transfer and offloading to cloud to enhance performance and reduce energy consumption.

Contribution

It introduces a comprehensive energy consumption model and an iterative optimization algorithm for multi-radio offloading, achieving near-optimal energy savings.

Findings

The iterative algorithm consumes only 3% more energy than the optimal solution.

Simulations show significant energy efficiency improvements in multi-radio offloading.

The model effectively balances energy use and data transfer across multiple radio links.

Abstract

The advent of 5G networking technologies has increased the expectations from mobile devices, in that, more sophisticated, computationally intense applications are expected to be delivered on the mobile device which are themselves getting smaller and sleeker. This predicates a need for offloading computationally intense parts of the applications to a resource strong cloud. Parallely, in the wireless networking world, the trend has shifted to multi-radio (as opposed to multi-channel) enabled communications. In this paper, we provide a comprehensive computation offloading solution that uses the multiple radio links available for associated data transfer, optimally. Our contributions include: a comprehensive model for the energy consumption from the perspective of the mobile device; the formulation of the joint optimization problem to minimize the energy consumed as well as allocating the associated data transfer optimally through the available radio links and an iterative algorithm that converges to a locally optimal solution. Simulations on an HTC phone, running a 14-component application and using the Amazon EC2 as the cloud, show that the solution obtained through the iterative algorithm consumes only 3% more energy than the optimal solution (obtained via exhaustive search).