Energy-Efficient Dynamic Offloading and Resource
Scheduling in Mobile Cloud Computing
Songtao Guo
∗
, Bin Xiao
†
, Yuanyuan Yang
‡
and Yang Yang
∗
∗
College of Electronic and Information Engineering, Southwest University, Chongqing, 400715 China
†
Department of Computing, The Hong Kong Polytechnic University, Kowloon, Hong Kong
‡
Department of Electrical & Computer Engineering, Stony Brook University, Stony Brook NY 11794, USA
Abstract—Mobile cloud computing (MCC) as an emerging
and prospective computing paradigm, can significantly enhance
computation capability and save energy of smart mobile devices
(SMDs) by offloading computation-intensive tasks from resource-
constrained SMDs onto the resource-rich cloud. However, how
to achieve energy-efficient computation offloading under the hard
constraint for application completion time remains a challenge
issue. To address such a challenge, in this paper, we provide
an energy-efficient dynamic offloading and resource scheduling
(eDors) policy to red uce energy consumption and shorten applica-
tion completion time. We first formulate the eDors problem into the
energy-efficiency cost (EEC) minimization problem whil e satisfying
the task-dependency requirements and the completion time dead-
line constraint. To solve the optimization problem, we then propose
a distributed eDors algorithm consisting of three subalgorithms
of computation offloading selection, clock frequency control and
transmission power allocation. More importantly, we find that
the computation offloading selection depends on not only the
computing workload of a task, but also the maximum completion
time of its immediate predecessors and the clock frequency and
transmission power of the mobile device. Fin all y, our experimental
results in a real testbed demonstrate that the eDors algorithm can
effectively reduce the EEC by optimally adjusting the CPU clock
frequency of SMDs based on the dynamic voltage and frequency
scaling (DVFS) technique in local computing, and adapting the
transmission power for the wireless channel condition s in cloud
computing.
Index Terms—Mobile cloud computing, energy-efficiency cost,
computation offloading, resource all ocation.
I. INTRODUCTION
Recently, smart mobile devices (SMD), e.g., smartphones and
tablet-PCs, are gaining enormous popularity due to their porta-
bility and compactness. As expected, SMDs are taken as the
dominant future computing devices for su pporting computation-
intensive applications, such as interactive g aming, image/video
processing, e-commerce, and o nline social ne twork services [1],
[2]. Such complex applications necessitate higher comp uting
power, memory and battery lifetime on SMDs [3]. Due to the
physical size constraint, however, mo bile devices are in general
resource- constrained. In particu la r, the limited energy supply
from the battery has been one of the most challenging d esign
issues for SMDs [5]–[ 7].
With the d evelopment of wireless communication technology
such as 3G, Wi-Fi and 4G, mobile cloud computing (MCC) is
envisioned as a pro mising approach to ad dress such a challenge.
The objective of MCC is to extends powerful c omputing capa-
bility of the resource-rich clouds to the resources constrained
SMDs so as to augment computing pote ntials of SMDs. To
achieve this objective, MCC nee ds to migrate resource -intensive
computa tions from SMDs to the cloud via wireless access,
referred to as computation offloading. A mobile application
in MCC needs to be partitioned in to a seq uence of tasks
that can b e executed on the mobile device, called local com-
puting/execution, or be executed on the clou d, named cloud
computing/execution. Clearly, MCC c an accommodate SMDs
to execute com plex applications which are impractical to run
solely on SMDs due to insufficient SMD resources, such as
perceptio n applications [4]. As a nother advantage, MCC can
improve the performance of mobile devices by selectively
offloading tasks of an application onto the c loud. Moreover,
MCC is beneficial to save energy in mobile devices and prolo ng
operation time.
Although the MCC based on computation offloading tech-
nique can significantly enhanc e computation capability of
SMDs, it still remains challenging to deve lop a reliable MCC
system. A key ch allenge is how to achieve an energy-efficient
computa tion offloading. In order to realize the prospective ben-
efits of MCC in en ergy saving and performance improvement
for mobile devices, we should consider the following questions:
(i) Whic h tasks of an application should be offloaded onto the
cloud? (ii) How much CPU clo ck fr equency should be assigned
for each task in local computing? (iii) How much transmission
power should be employed for offloading the ta sks in cloud
computing?
To answer the above questions, in this paper we foc us on the
joint computation offloading and resource scheduling prob le m,
in which there are three key issues to be addressed.
• What happens with computation offloading selection when
both task-depe ndency requirements and application com-
pletion time constraint are enforced? The enforcemen t is
necessary since there exist in general a ce rtain precedenc es
among the tasks and the ap plication completion time is a
hard constraint for latency-sensitive applications.
• Can the energy-efficiency cost (EEC), as defined in Defi-
nition 2, be minimized by optimally controlling the CPU
clock fr equency of the mobile device via the DVFS tech-
nique [8] in local execution?
• Can the E EC be minimized by optimally allocatin g the
data transmission power for each computation offloading
while satisfying the task-pr ecedence requirements in cloud
execution?
The objective of this paper is to provide an optimal ene rgy-
efficient dynamic offloading and resource scheduling (eDors)
IEEE INFOCOM 2016 - The 35th Annual IEEE International Conference on Computer Communications
978-1-4673-9953-1/16/$31.00 ©2016 IEEE