Introduction

Abstract

Recent advances in smartphone technologies and wireless communications have created a strong uptick in the use of smart applications over web-enabled, resource-constrained end devices. In order to support such computation and data intensive applications over wireless devices, two problems must be addressed: (1) the resource constraints on the device (e.g., battery power, memory, etc.) and (2) network constraints (e.g., capacity/spectrum and latency limitations). Multiple radio spectrum access technologies (multi-RAT) is becoming one of the ways to address some of the network related problems (e.g., capacity). The heterogeneous network (HetNets) paradigm, enabling multi-RATs is expected to become a mainstay of future wireless networks. Simultaneous access to multiple RATs or spectrum bands can be implemented at the transport layer, network layer, or PHY/MAC layer of wireless devices. The growth of mobile virtual network operators (MVNO) will also facilitate such multi-RAT opportunistic spectrum access. Google’s recent deal with Sprint and T-Mobile is an example in this direction.

At the device level, cloud offloading has emerged as an indispensable part of the solution to the device level constraints. However, offloading computations to a remote cloud can also place an additional burden on the already overburdened wireless backbone. The term “cloud offloading” can mean data flow offloading or offloading computationally intense tasks to the cloud along with the data resulting from the computations. Here we mean the latter. In this chapter, we discuss the factors affecting computation offloading and discuss the organization of the book.