-
Simulation Modelling Practice and Theory 50 (2015) 96108Contents
lists available at ScienceDirect
Simulation Modelling Practice and Theory
journal homepage: www.elsevier .com/locate /s impatMobile
storage augmentation in mobile cloud computing:Taxonomy,
approaches, and open
issueshttp://dx.doi.org/10.1016/j.simpat.2014.05.0091569-190X/ 2014
Elsevier B.V. All rights reserved.
Corresponding author. Tel.: +60 1125273423.E-mail addresses:
[email protected] (N. Aminzadeh), [email protected] (Z.
Sanaei), [email protected] (S.H. Ab Hamid).Nazanin Aminzadeh ,
Zohreh Sanaei, Siti Hafizah Ab HamidFaculty of Computer Science and
IT, University of Malaya, Kuala Lumpur, Malaysiaa r t i c l e i n f
o
Article history:Available online 15 June 2014
Keywords:Mobile cloud computingCloud computingBig dataMobile
dataMobile data storageTaxonomya b s t r a c t
Worldwide employment of mobile devices in various critical
domains, particularlyhealthcare, disaster recovery, and education
has revolutionized data generation rate. How-ever, rapidly rising
data volume intensifies data storage and battery limitations of
mobiledevices. Mobile Cloud Computing (MCC) as the state-of-the-art
mobile computing aims toaugment mobile storage by leveraging
infinite cloud resources to provide unlimitedstorage capabilities
with energy-dissipation prevention. Researchers have
alreadysurveyed varied MCC aspects and its challenges, but
successful futuristic Mobile StorageAugmentation (MSA) approaches
demand deep insight into the current storage augmenta-tion
solutions that highlights critical challenges, which are lacking.
This paper thoroughlyinvestigates the main MSA issues in three
domains of mobile computing, cloud computing,and MCC to present a
taxonomy. Also, it examines several credible MSA approaches
andmechanisms in MCC, classifies characteristics of cloud-based
storage resources, andpresents open issues that direct future
research.
2014 Elsevier B.V. All rights reserved.1. Introduction
Mobile data volume has been increased drastically in recent
years. Analysys Masons report [1] forecasts 6.3 times growthof
mobile data volume traffic amid 2013 and 2018. One reason, among
the bevy of inducements is the exploitation of mobiledevices in
various critical domains of healthcare [2], disaster recovery [3],
and education [4]. The voluminous and rapidlygenerated data in
various modalities, which is known as big data [5], demands high
capacity, flexible, and reliable storageinfrastructure.
However, mobile devices are characterized by storage
constraints, limited processing, and a short span battery.
Storagelimitation and energy consumption are critical factors for
resource constrained non-stationary computing devices,
especiallysmartphones. High performance and energy-efficient data
storage ensures the battery lifes augmentation. Despite
advance-ments for augmenting a mobile devices storage including
employment of flash and Secure Digital (SD) cards, current
richmobile applications [6] demand higher storage capacity.
Reducing the effects of mobile devices deficiencies and
unreliablewireless connections in comparison with wired networks,
are the ultimate goal in the plethora of efforts [7,8] and research
torealize end-users demand.
The emergence of the cloud as a rich resource with unlimited
computing and storage capacities can be considered as agood
solution to mobile devices resource constraints. The use of a
wireless medium, the offloaded data, the dependency
http://crossmark.crossref.org/dialog/?doi=10.1016/j.simpat.2014.05.009&domain=pdfhttp://dx.doi.org/10.1016/j.simpat.2014.05.009mailto:[email protected]:[email protected]:[email protected]://dx.doi.org/10.1016/j.simpat.2014.05.009http://www.sciencedirect.com/science/journal/1569190Xhttp://www.elsevier.com/locate/simpat
-
Table 1List of acronym definitions.
Acronym Description
ACID Atomicity, Consistency, Isolation, DurabilityDBMS Data Base
Management SystemDSP Decryption Service ProviderESP Encryption
Service ProviderGPS Global Positioning SystemsI/O Input/OutputIaaS
Infrastructure as a ServiceICN Information Centric NetworkiSCSI
Internet Small Computer System InterfaceLMH Large Mobile HostMANET
Mobile Ad-hoc NetworkMCC Mobile Cloud ComputingMNO Mobile Network
OperatorsMSA Mobile Storage AugmentationPC Personal ComputerPDA
Personal Digital AssistantsPP-CP-ABE Privacy Preserving-Ciphtertext
Policy-Attribute
Based EncryptionSaaS Storage as a ServiceSAL Storage Abstraction
LayerSD Secure DigitalSIM Subscriber Identity ModuleSLA Service
Level AgreementSMH Small Mobile HostSOA Service Oriented
ArchitectureVM Virtual MachineWAN Wide Area NetworkWi-Fi Wireless
FidelityWLAN Wireless Local Area Network
N. Aminzadeh et al. / Simulation Modelling Practice and Theory
50 (2015) 96108 97on a specific vendor, and data replication are
among several challenges in this domain. Comprehensive studies
[911] havereviewed and tried to address challenges in this area.
However, the latest endeavor is to deploy cloud resources to
augmentcomputing [12] and storage [13] capabilities for a multitude
of mobile devices which leads to the state-of-the-art MCC.
The MCC paradigm combines cloud computing, mobile computing, and
networking [14] to enhance the performance andcapacity of mobile
devices. It is characterized by inherited mobility and rich
services from mobile and cloud computingwhere a resource poverty
(storage, computation, and battery) can impede the vision of time-,
location-, and system type-freeubiquitous computing [15].
In the previous works [1419], the MCC domain have been
comprehensively investigated from various perspectives. In[19],
authors presented an extensive survey of heterogeneity in the MCC
domain, presented an MCC definition, identifiedmajor MCC
challenges, devised its taxonomy, and highlighted several crucial
open issues that help to identify future researchdirections.
Cloud-based augmentation [16] surveys the recent mobile
augmentation efforts that employ cloud computinginfrastructures to
enhance computing capabilities of resource-constraint mobile
devices, especially smartphones. To theextent of our knowledge,
investigation of storage augmentation issues in the MCC domain is a
nascent literature and requirescomprehensive study and
analysis.
In this paper, we comprehensively analyze MSA issues in the
context of mobile computing, cloud computing, and MCCwhere each
domains issues are investigated. Based on the investigated issues,
we classify MSA issues into three classesof mobile device,
cloud-based and converged issues. Based on a review of prominent
MSA approaches in MCC, we classifycloud-based storage
characteristics in a taxonomy that encompasses architecture,
capacity, tiering, mobility, location, andback-end connectivity.
The paper highlights several open issues in MCC for MSA to pave the
way for future efforts. Mobiledevices and smartphones are used
interchangeably in this paper. Table 1 provides a list of acronyms
used throughout thepaper.
The remainder of this paper is presented as follows. Section 2
presents the motivation for MSA in MCC based on a devisedtaxonomy
of issues. Section 3 reviews current approaches for MSA in MCC.
Section 4 provides a taxonomy of cloud-basedstorage characteristics
in MCC. Open issues are highlighted in Sections 5 and 6 concludes
the paper.2. Motivation
Contemporary smartphones are dominant mobile devices that not
only provide the basic telephony features of traditionalcellphones,
but also incorporate the functionalities of several other digital
devices, particularly Personal Digital Assistants(PDA), Global
Positioning Systems (GPS), sound recorders, and digital cameras
[20] and are contributing to rapid digital datageneration rate by
producing data files, including emails, spreadsheets, bank
statements, and multimedia files (i.e., video,
-
98 N. Aminzadeh et al. / Simulation Modelling Practice and
Theory 50 (2015) 96108music, images, and any location related
information) [21]. However, such rapid data generation by mobile
devices namedmobile data demands extremely large elastic storage on
mobile devices which is beyond their intrinsic capabilities.
Mobiledevices are characterized by storage constraints, limited
processing, and a short span battery. Hence, storage limitation
andenergy consumption are critical factors for resource constrained
non-stationary computing devices. Moreover, issues, partic-ularly
mobility, limited wireless throughput, and a small user interface
are among real challenges that affect data availabilityin mobile
computing [16]. High performance energy-efficient storage ensures
the battery life prolonging.
Despite the momentous role of mobile devices in various aspects
of our daily life, resource poverty is a mobile devicesmajor
deficiency that impedes an end-users rich experience and demands.
The incorporation of SD cards to expand storagecapacity in addition
to the available built-in and on-SIM (Subscriber Identity Module)
storage on smartphones as an instanceof mobile devices, cannot
compete with end-users demand. To address poor storage capacity, a
plethora of research hasfocused on two varied hardware and software
level approaches; shrinking big data volume and augmenting
storagecapacities.
To shrink big data, researchers in [22,23] try to apply quantum
theory to optimize the use of current storage by up to 80%.Quantum
data represent bits of data that use quantum bits or qubits to
reduce the amount of bits used by each data in thestorage.
Currently, a bit uses 12 atoms; however, in 2012, IBM [24]
demonstrated feasibility of using only eight atoms tostore one bit
data. Quantum theory can realize the possibility of dedicating only
one atom to each bit to ensure a tremendousreduction in storage
utilization (up to 80%). However, the fully realization of this
vision demands further extensive research.Alternatively, storage
augmentation approaches aim to alleviate storage shortcomings by
leveraging lightweight methods.
The review of storage augmentation approaches in mobile devices,
cloud resources, and MCC has led us to devise ataxonomy of issues
as depicted in Fig. 1. The taxonomy encompasses three main classes
of mobile device, cloud-based,and converged issues, which are
described in more details as follows.2.1. Mobile device issues
In the work [16], a taxonomy of augmentation motivation has been
presented based on the intrinsic deficiencies of mobiledevices.
Storage augmentation approaches mainly facing issues pertaining to
intrinsic and non-intrinsic deficiencies ofmobile devices. Hence,
the same taxonomy can be applied for the classification of storage
augmentation issues. The process-ing power of mobile devices,
energy resources, local storage, visualization capabilities, data
safety, security, and privacy arethe vital issues that pose
challenges and encumber efficient and effective mobile data
storage.
Mobile devices are characterized by limited processing power,
which implies restricted cache equipment. Restrictedcache imposes
more I/O operation and increased energy consumption and processing
time for data-intensive operations.Data transmission in
intermittent wireless networks is an energy-consuming task.
Intermittency due to the unreliabilityof a wireless link or a
planned disconnection to save battery life are inevitable [25].
Therefore, data availability on mobiledevices becomes a significant
research direction. However, being nomadic, mobile devices cannot
benefit from the vast datastorage that is available to their
immobile counterparts, which confines the amount of locally-stored
data. Moreover, amobile DataBase Management Systems (DBMS)
functionality is influenced by the limited processing power and
memoryof the host mobile devices. The size-constraint screens and
the gracile keyboards of mobile devices as visualization
limita-tions are other impediments in the design of storage
augmentation approaches.
Researchers endeavor to address aforementioned issues by major
MSA approaches including caching [26], hoarding, andreplication
mechanisms [27] as well as broadcasting [28] and summarization [25]
for mobile environment data.Fig. 1. Taxonomy of MSA issues in
MCC.
-
N. Aminzadeh et al. / Simulation Modelling Practice and Theory
50 (2015) 96108 992.2. Cloud-based issues
Storage augmentation issues are not merely peculiar to the
mobile environment. Advancements in data storage technol-ogy from
magnetic tapes and disks, optical/magneto-optical storage media,
and flash memory to a storage area network,network attached
storage, and even storage virtualization approaches [29] cannot
compete with the drastic growth of digitaldata to terabytes and
petabytes [9], which calls for additional data-intensive computing
and management mechanisms.
The advent of the cloud as a rich computing resource [30] has
provided virtually infinite data storage to the industry
andindividuals. The cloud as proposed in [30] is a type of parallel
and distributed system consisting of a collection of
intercon-nected and virtualized computers dynamically provisioned
and presented as one or more unified computing resources basedon
service-level agreements established through negotiation between
the service provider and consumers. Cloud comput-ing exploits cloud
resources to provide ubiquitous, convenient, on-demand network
access to a shared pool of configurablecomputing resources (e.g.,
networks, servers, storage, applications, and services) that can be
rapidly provisioned and releasedwith minimal management effort or
service provider interaction [31].
However, reliability, Service Level Agreement (SLA),
performance, portability, security, and data confidentiality [9]
can beconsidered major cloud-related storage augmentation issues.
Reliability can be realized with redundancy, which may jeop-ardize
guarantees for Atomicity, Consistency, Isolation, and Durability
(ACID) transactions. Active standardization to mitigatea customers
data extraction as an approach for the data lock-in and portability
issue, the concurrent execution of an appli-cations treads,
optimized online scheduling [32] and the rich encryption as
mechanisms for performance and security areamong some of the
effective approaches for addressing the aforementioned issues
[33].
2.3. Converged issues
The power of MCC originates from the convergence of mobile and
cloud computing, which merges mobility, richresources, and
functionality as renowned advantages of each domain [15]. Resource
poverty processing power, storagecapacity, and battery power are
inborn characteristics of mobile devices, and the consequent mobile
device storageaugmentation issues can be alleviated by leveraging
MCC. MCC is a rich mobile computing technology that leverages
theunified elastic resources of various clouds and network
technologies toward an unrestricted functionality, storage,
andmobility to serve a multitude of mobile devices anywhere and
anytime through the channel of the Ethernet or Internetregardless
of the heterogeneous environments and platforms based on the
pay-as-you-use principle [19].
However, MCC is heterogeneous by nature because of inhomogeneous
converged computing and networking technolo-gies [19]. MCC
heterogeneity is a consequence of diverse hardwares, operating
systems, development languages, datastructures and network
technologies. Although mobile computing is augmented through rich
resources of the cloud, theheterogeneity-based issues pose real
challenges for storage amelioration in MCC. Data portability and
interoperability,energy efficiency, long WAN latency, and data
fragmentations are among the major issues of storage augmentation
inMCC as identified in our proposed taxonomy. Addressing mobile
data storage issues in MCC has leaded to several MSAapproaches. In
the next section, we critically review the state-of-the-art
approaches.
3. Mobile-cloud storage augmentation approaches
In this section, we review credible efforts that aimed to
address storage amelioration and the I/O performance of
mobiledevices by leveraging cloud resources and lightweight
methods. We analyze and synthesize the approaches from
severalaspects, including mobility, energy efficiency, latency,
reliability, and architecture which are presented in Table 2.
Attribute-Based Data Storage (ABDS)The ABDS system [34] is part
of the security framework presented in [34]. The framework
encompasses two prominentcomponents: namely, a privacy preserving
CP-ABE (PP-CP-ABE) and an ABDS scheme. By leveraging PP-CP-ABE, the
cloudhandles encryption and decryption tasks that entail an
excessive load on resource-poor mobile devices. Two core parts,the
Encryption Service Provider (ESP) and the Decryption Service
Provider (DSP), provide data confidentiality and securityservices
without being aware of the encryption key and data contents.
Similar to Phoenix [35], ABDS exploits the datapartitioning
mechanism. However, ABDS differs from Phoenix by outsourcing the
encryption and decryption tasks tothe public cloud. The ABDS scheme
partitions data to a number of blocks and manages the encryption
operations onthe required blocks independently. Consequently, when
modifying a file on the cloud, the update and encryption tasksare
merely applied on specific blocks, which balance the storage and
communication overhead through data storageand retrieval.Fig. 2
depicts ABDS framework. The data owner refers to the mobile nodes
in the framework. The data owner accesses thestorage service
provider services in a secure, energy-efficient, and optimized
manner by utilizing outsourced services oftwo prominent components
(i.e., DSP and ESP).Encryption through a data partitioning approach
minimizes the data management and the cloud vendors charged
costs.However, encryption requires additional control information
for data blocks, which imposes extra overhead for the data
-
Table 2Comparison of cloud-based storage augmentation
approaches.
ABDS SmartBox WhereStore Wukong Phoenix E-DRM EECRS MOMCC
UbiqStor SAMI
Mobility Immobile Immobile Immobile Immobile Mobile Mobile
Mobile Mobile Immobile ImmobileStorage augmentation High High High
High Low Low Low Low High HighEnergy Efficiency Low Low Low Low
High High High High Medium MediumArchitecture ClientServer
ClientServer ClientServer ClientServer P2P P2P P2P P2P Hybrid
HybridTiering Single-tier Single-tier Single-tier Single-tier
Single-tier Single-tier Single-tier Single-tier Multi-tier
Multi-tierWAN Latency High High High High Low Low Low Low Hybrid
HybridData Blocking Low Low Low Low High High High High Medium
MediumReliability High High High High Low Low Low Low Medium
MediumAvailability High High High High Low Low Low Low Medium
MediumSecurity and Trust High High High High Low Low Low Low Medium
MediumData Safety High High High High Low Low Low Low Medium
MediumLocality Low Low Low Low High High High High Medium MediumI/O
Performance Varied Varied Varied Varied Low Low Low Low Medium
MediumBack-end Connectivity Wired Wired Wired Wired Wireless
Wireless Wireless Wireless Wired WiredCommunication Platform
Internet Internet Internet Internet Ethernet Ethernet Ethernet
Ethernet Internet/EthernetContext Awareness Low Low Low Low High
High High High Medium MediumClient Networking Cellular/WLAN
100N.A
minzadeh
etal./Sim
ulationModelling
Practiceand
Theory50
(2015)96
108
-
Fig. 2. ABDS framework.
N. Aminzadeh et al. / Simulation Modelling Practice and Theory
50 (2015) 96108 101management tasks. Moreover, the data block size
should be identified optimally to satisfy the goal of minimizing
datamanagement overhead tasks.SmartBoxSmartBox [21] is a solution
for mobile storage capacity expansion through cloud resources. It
employs cloud spacereferred to as shadow storage for storing and
retrieving personal data by assigning a unique account for each
mobiledevice. SmartBox is a rich personal repository and provides a
public space as common storage for sharing data amongassociated
mobile devices. Data access and navigation is provided via
traditional hierarchical namespace and employsan attribute-based
method for semantic query, which uses publisher-provider
metadata.Nevertheless, the SmartBox design conforms to write once
read many policy, which impedes its application as an enter-prise
solution. In addition, it requires a permanent connection to access
data on the cloud, which is not feasible in manycases.WhereStoreA
location-aware data storage approach is considered by WhereStore
[36] to enhance smartphones I/O performance byusing cloud storage
resources. The prediction and caching of adjacent place information
through the replication of prom-inent data retrieved from the cloud
is proposed in this solution to optimize data transfer time and to
conserve energy.Although several human mobility models and routing
techniques [37] are proposed to figure out movement plan ofmobile
users, predicting the users destination efficiently and specifying
the appropriate state and time for caching arenon-trivial tasks,
which require further investigation.WukongWukong [13] is an
additional I/O performance optimization and storage enlargement
effort and is a cloud-based file ser-vice that provides
user-friendly, transparent data access to inhomogeneous remote
cloud storage services for mobiledevices. By leveraging the Storage
Abstraction Layer (SAL) and plugin procedures, cloud data and
services, namely Drop-box and Amazon S3 can be accessed
transparently and remotely via mobile applications as if they are
available locally.Wukong realizes the provision of a uniform
interface to access various cloud services simultaneously. The
optimizationstack, cache management, and pre-fetch mechanisms are
utilized to improve the throughput and minimize long WANlatency.
Moreover, the application of encryption and compression mechanisms
enhances data security and communica-tion, respectively, and
completes the contribution of this approach.However, limited
wireless bandwidth and I/O operation costs are impediments that
hinder Wukong in minimizing longWAN latency. Moreover, the
employment of efficient compression methods for multimedia files is
required for thesuccess of this proposal. The proposed compression
method is more beneficial for minimizing a text files size.
These
-
102 N. Aminzadeh et al. / Simulation Modelling Practice and
Theory 50 (2015) 96108objects encompass file types with a high
ratio of compression. In addition, the prefetching method is more
efficient insequential reading, which requires an improvement in
random reading to enhance the performance of the work. How-ever, in
current prototype, defining large intervals among the open and read
processes would save time during readoperations.PhoenixPhoenix [35]
is a distributed communication and storage protocol that considers
the power-aware storage sharing ofeager nearby mobile devices in
the cloud for one-hop networks. Unlike E-DRM [38], which utilizes a
periodic broadcastof buffered data, Phoenix ensures data
availability on participant mobile nodes by breaking each data
content to a numberof blocks and copying them on at least two or
more mobile devices. Phoenix provides edge nodes with a true number
ofdata blocks copied to maintain data availability and longevity in
the self-organized cloud of mobile devices. Authors focuson the
extending the data survival time by minimizing disconnections, by
administrating node movement, and by leavingthe support for the
data update tasks to databases built on Phoenixs platform.In
addition, autonomous management and maintenance is realized by
utilizing an advertising model. When a nodeparticipates, it starts
an asynchronous advertisement timer to decide whether to broadcast
a block as a winner or tobecome a follower contributor as a loser.
The Phoenix performance is evaluated through testbed and simulation
exper-iments. Simulation is performed via TOSSIM in TinyOS. The
number of data block copies as well as the effect of nodesfailure
and mobility are among the main factors examined in 100
simulations. The results show Phoenix support for dataavailability
because the number of available block copies is rarely less than
the number of nodes in the network wherenodes are immobile. In the
case of node mobility, it ensures data longevity by maintaining at
least two copies of blocks inthe network. However, the current
evaluation considers only one-hop networks. The simulation in
multi-hop networks[39] can improve Phoenixs applicability.Phoenixs
dependency on defining the number of participant nodes, the lack of
authentication over a wireless network,and the overhead imposed on
the network, in addition to the security, data safety, and privacy
issues are among thechallenging tasks that can hinder Phoenixs
application in a real scenario.Eager replication extended Database
State Machine (E-DRM)An E-DRM [38] is an energy-efficient
replication approach for a Mobile Ad-hoc NETwork (MANET) [40]. A
MANET encom-passes a network of wireless mobile servers and
clients. The nodes of a MANET are nomadic and have no
permanentaccess to electricity. Hence, MANETs are restricted from
an energy source. Eager replication maintains data consistencyvia
the guarantee of an identical value for all copies on nodes. Unlike
Phoenix [35] that focuses on preserving a true num-ber of copies in
the network, E-DRMminimizes the number of message broadcasts to
conserve energy. In E-DRM, authorspropose an eager replication
scheme for application on a Database State Machine (DSM) approach
introduced in [41]. TheDSM supports atomic broadcasts in which
nodes consent to the order and the set of delivered
transactions.E-DRM considers two mobile node types in the
architecture of a MANET. Small Mobile Hosts (SMH) have
restrictedresources, and Large Mobile Hosts (LMH) have a rich
resource capacity. SMHs contain only a part of the database
toaccomplish query tasks. Complete databases reside on LMHs and are
responsible for broadcasting, certifying, and updat-ing tasks.
E-DRM minimizes message broadcasting overhead by buffering and
periodic broadcasting at the beginning ofthe broadcast cycle. The
broadcast cycle commences when a client requests a SMH to update or
read data. SMHs committhe transaction to a LMH to be processed and
broadcasted to other LMHs. LMHs commit or abort the transaction and
cer-tify it as a correct transaction. The cycle ends when a LMH
broadcasts the result to a SMH to be transmitted to the client.Fig.
3. E-DRM broadcast cycle.
-
N. Aminzadeh et al. / Simulation Modelling Practice and Theory
50 (2015) 96108 103Fig. 3 shows the broadcast cycle in detail. The
E-DRM approach considers the start of the broadcast cycle for
dispatchingtransactions to conserve energy. However, in real-time,
all LMHs validate the transaction locally and conserve the
energythat would be consumed for message broadcasting and
consistency management.EECRSEECRS [42] is an energy-aware content
retrieval scheme for mobile cloud. It proposes a
directive-selective forwardingscheme for broadcasting interest
packets in a MANET to obtain the contents. The EECRS approach is
considered in anInformation Centric Network (ICN). The contents of
an ICN are characterized by a hierarchical-based naming
strategy.Hence, names identify the requested contents. The EECRS
improves energy efficiency and scalability by eliminatinginterest
packets duplication to diminish traffic load. In addition, it
searches multiple caches in parallel to fill the missinggap of the
networks node where interest packets fail to reach.Nonetheless, the
inability of the EECRS to guarantee a 100% hit rate through the
proposed approach hinders its full appli-cation. Authors develop an
intelligent broadcast approach for content inquiry in case of
missing nodes. Although thismethod ensures a content node hit, the
imposed traffic overhead decelerates its performance and
energy-efficiency.Market-Oriented Mobile Cloud Computing
(MOMCC)MOMCC [43] is a service-oriented architecture-based mobile
application development framework. The prominent engag-ing entities
of the MOMCC architecture are the central supervisory entity as
governor, service programmer, mobile host,and requester mobile
node. In MOMCC, service developers create services and upload them
in a central database for pub-lic availability. Individual mobile
device owners who are interested in sharing their resources with
nearby resource-poormobile devices register their interest with the
governor. The governor authenticates and authorizes the registered
mobiledevices that enter a request for hosting specific services.
Moreover, the governor finds a secure proximate host for
serviceexecution in response to the service requester. The MOMCC
overcomes the imposed overhead of code offloading in [44]by
restricting hosting to services executed without offloading. The
participating hosts as well as the governor, service pro-grammer,
and application developer will earn money on publishing, governing,
hosting, and sharing services. The lack ofdirect attachment between
the service developer and the requester enhances privacy in the
proposed framework.The promise of this approach is the execution of
services on a multitude of nearby mobile devices under the
administra-tion of a centrally trusted governing entity. The
employment of adjacent mobile devices improves the availability
ofresources, which can be leveraged to alleviate mobile devices
storage limitations. Moreover, the communication ofmobile devices
over a WLAN minimizes latency and conserves the energy, which would
otherwise be dissipated oncellular networks. However, the MOMCC
architecture shortcomings, such as the restricted computing of host
devices,the necessity for fine-grained services, and the dependency
on networking impede its employment for enterpriseapplications and
in offline mode.UbiqStorUbiqStor [45] is an internet Small Computer
System Interface (iSCSI) cache server that reduces the response
delay timebetween mobile users and storage servers. The access to
remote fixed storage servers via the nearest immobile
UbiqStorserver as a hybrid, multi-tier solution provides mobile
clients energy-efficient access to rich storage capacity by
eliminat-ing long WAN latency in case of direct access to
non-proximate storage servers. Initiator, target, and block
managementmodules are components of UbiqStor, which handles iSCSI
read and write mobile clients requests through a
cachingmechanism.Despite the enhancement of transfer latency and
response delays, wireless network intermittency and cache
serverproximity factors can affect UbiqStors performance and can
hinder its successful application.Service-based Arbitrated
Multi-tier Infrastructure (SAMI)SAMI [15] is an Infrastructure as a
Service (IaaS) hybrid model for mobile cloud computing that
addresses storageaugmentation issues in addition to its main
purpose of data-intensive computations in MCC. It promotes the I/O
perfor-mance as well as the storage capacity of resource-poor
mobile devices through a hybrid architecture that encompassesthree
layers of an infrastructure, an arbitrator and a Service-Oriented
Architecture (SOA). Mobile device users can benefitfrom SAMIs
multi-tier infrastructure, which provides access to distant fixed
clouds, proximate MNOs, and adjacent MNOFig. 4. Cloud-based storage
taxonomy.
-
104 N. Aminzadeh et al. / Simulation Modelling Practice and
Theory 50 (2015) 96108authorized dealers. SAMIs Multi-tier
infrastructure provides decreased communication latency and energy
efficiency as aconsequence, to latency-sensitive applications by
MNO authorized dealers. In case of demand for
security-sensitiveservices, MNOs are beneficial by providing access
to their resources via more secure cellular network in comparison
toInternet channel. When MNOs resources are incapable of meeting
data-intensive services demand or not accessible afteroperating
hours, cloud infrastructures are exploited.Reducing latency,
establishment of trust, alleviating portability, and preventing
energy dissipation are among promisingadvantages of SAMI. Similar
to UbiqStor [45], SAMI exploits a multi-tier architecture in its
design. However, SAMI pro-vides higher trust to mobile service
consumers by leveraging resources of MNOs beside cloud and adjacent
MNO dealers.
Furthermore, Table 2 summarizes the comparison results of the
reviewed approaches based on the defined categories inthe taxonomy
presented in Fig. 1. Table 2 advocates differentiations between
mobile and immobile storage augmentationapproaches. We deduce from
Table 2 that the mobility factor plays an important role and
affects other specified factors.Employing mobile cloud resources
enhances energy efficiency, WAN latency, locality, and context
awareness, while utilizingan immobile cloud back-end improves
storage augmentation, data blocking, reliability, availability,
data safety, security, andtrust.
4. Cloud-based storage characteristics
According to existing MSA approaches including the reviewed
solutions in the previous section, we have devised acloud-based
storage taxonomy. This classification is based on the prominent
common characteristic and major differencesbetween various cloud
storage resources including the architecture, capacity, tiering,
mobility, locality, security, back-endconnectivity, and utilization
cost factors. Fig. 4 depicts the cloud-based storage taxonomy and
will be explained in thissection.
Architecture: The variety in the cloud resources architecture
arises from diverse roles of the cloud in augmentingresource-poor
mobile devices. In a clientserver architecture, a multitude of rich
stationary servers owned by vendorsand enterprises serve mobile
devices as their clients to enhance their storage capacity. Public
and private cloud resourcesfall into this category. Mobile Network
Operators (MNO), stationary computers in public places as well as
authorizedMNO dealers, computing kiosks, and mobile devices are
among other resources that have a server role and offers
storageservices to storage-constrained mobile devices [16]. Amazon
S3,1 Amazon EBS,2 HP Cloud Object Storage,3 and HP CloudBlock
Storage4 are well-known instances of storage services offered via
the public cloud.In peer-to-peer architecture, mobile devices are
both a provider and consumer of services in a decentralized
autonomousnetwork of proximate non-stationary devices that form a
mobile storage service. Conferences, coffee shops, and
universitycampuses are scenarios in which ad hoc networks can be
formed among participants that generally share common interests.In
this architecture, cloud storage services are exploited locally
from the network of available mobile devices in the vicinityfor a
period of time [35,38]. The storage capacity provided in a
peer-to-peer architecture is incapable of competing with infi-nite
storage that is accessible via the public cloud in a clientserver
architecture. Moreover, a peer-to-peer architectureleverages a
decentralized management, on the contrary to a clientserver
architecture with centralized management. How-ever, the
exploitation of edge devices storage in a peer-to-peer network
alleviates long WAN latency and device energy con-sumption.
Moreover, caching information of same interest on participants
devices is another advantage of this approach,which lessens the
traffic over a cellular network. Nonetheless, a guarantee of the
persistency and redundancy of data is achallenging task in this
architecture due to node mobility and a participants disconnection,
which may cause data loss whenthe participant leaves the network or
powers off the device [46]. The researchers efforts are toward the
application of power-aware, real-time-aware, or partition-aware
[38,47] approaches in wireless and ad hoc networks to conserve
mobile devicesenergy, to minimize communication volume, and to
maintain the availability of data in a well-timed manner.The hybrid
architecture includes approaches that exploit nearby mobile devices
that have a reduced latency as well as richservers supplying
services in public and private clouds and that have proximate
computers in the vicinity. Nevertheless, theadvantages of
unleashing hybrid architectures necessitate novel implementation,
administration, and resource schedulingprocedures that can lead to
new challenges in MCC realm.Capacity: Storage is one of the main
services provided by the cloud as Storage as a Service (SaaS). The
unlimited storagecapacity of the cloud can augment
resource-constrained mobile devices. SaaS capacitates keeping data
and applicationson the cloud resources and access them remotely.
These unified, elastic, reliable, secure resources are usually
located farfrom mobile nodes. Hence, mobile devices require
crossing an Internet channel via a wireless network to access the1
http://aws.amazon.com/.2 http://aws.amazon.com/ebs/.3
http://docs.hpcloud.com/object-storage/.4
https://docs.hpcloud.com/block-storage/.
http://aws.amazon.com/http://aws.amazon.com/ebs/http://docs.hpcloud.com/object-storage/http://https://docs.hpcloud.com/block-storage/
-
N. Aminzadeh et al. / Simulation Modelling Practice and Theory
50 (2015) 96108 105unlimited storage of a distant public cloud.
Dropbox,5 Google Docs,6 Amazon S3, iCloud,7 and SkyDrive8 are among
therenowned cloud storage services that facilitate the storage,
sharing, and synchronization of data for desktop and mobiledevice
users. Several efforts such as UbiqStor [45] and MiSC [48] tried
augmenting mobile devices using remote storage serv-ers. The
utilization of non-proximate cloud storage services over an
Internet channel imposes increased WAN latency anddecreased
security. An alternative solution is the use of constrained
transient storage capacity via nearby immobile cloudresources [49]
and mobile devices [35]. Although leveraging this approach supplies
limited storage to lightweight mobiledevices in comparison to
infinite storage capacity of the cloud, the reduced latency over
cellular networks and Wi-Fienhances performance and addresses
mobile devices energy deficiency problem.Tiering: Tiering refers to
the number of layers a mobile device passes over to exploit cloud
services. Approaches such as[45,15] are referred to as multi-tier.
For example, the three-tier infrastructure of SAMI [15] encompasses
distant cloudresources, proximate MNOs, and MNO-authorized dealers
in the vicinity. Conversely, the one-tier category
approachesleverage one of the cloud resources for storage
augmentation.Multi-tier infrastructure provides rich computing
beside energy- and latency-aware communication. It is
realizedthrough the provision of several options for execution of
services. While proximate cloud resources tier covers
therequirements of latency-sensitive services, the choice of
exploiting reliable and well known organizations resources tieris
beneficial to the services with security and trust requirements.
Service providers in the telecommunication are trustfuloptions for
the provision of resources over the secure cellular network as a
cloud. The public clouds as the equipped tierwith elastic, infinite
storage and computing resources are the choice for data-intensive
services when resources of trust-worthy, proximate tiers are
insufficient.Mobility: Cloud resources are either mobile or
immobile with respect to the mobility characteristic. When
immobile,cloud resources are delineated by a richer source of
computation and storage capacity and no energy shortcomings
com-pared with mobile resources. Public and private cloud servers,
MNOs, and fine-grained resources such as available desk-top
computers in coffee shops and airports and cloudlets are instances
of immobile storage resources. A cloudlet is aproximate immobile
cloud solution that concentrates on the computational augmentation
of non-stationary neighboringdevices. Similarly, it can be applied
for storage enhancement in which the VM-based offloading method of
the cloudlet isemployed for more storage capacity to proximate
mobile devices. On the contrary, mobile resources provide
context-awareness and lower latency, which can adversely impact the
energy consumption and transmission period.Location: Cloud storage
resources are either located in the proximity or are distant.
Approaches utilizing a multi-tierinfrastructure can benefit from
both and are referred to as hybrid resources. Distant resources,
particularly public cloudstorage, are accessible via an Internet
channel over a wireless medium. Long WAN latency and low security
are commonchallenges in wireless networks and are intensified
through an Internet channel. However, proximate cloud resources
areaccessible via both Ethernet and Internet channel. Utilizing
nearby mobile or fixed resources reduces latency, cost,
andcommunication overheads [50]. Stationary computers in public
places and nearby mobile devices specifically smart-phones,
laptops, and tablets are resources that can serve client mobile
nodes in the vicinity.Back-end Connectivity: This category
considers the connectivity of back-end storage resources to the
Internet. They areeither connected to a wired Internet connection
or utilize a wireless medium. Mobile devices that play the role of
serversto other mobile nodes or participate in a peer-to-peer
architecture are supported via a wireless connection to the
Internet.On the contrary, stationary servers, whether public or
private cloud resources, or immobile computers in the neighbor-hood
are equipped with a wired Internet connection.
5. Open issues
In this section, we present crucial open issues on MSA in MCC as
several thought-provoking future research directions.5.1.
Energy-awareness
Energy conservation is one of the ultimate goals in mobile
computing, cloud computing, and MCC domains. Although lim-itation
of resources, processing, storage, and battery life of mobile
devices can be alleviated through rich cloud-basedresources [16],
high pre- and post-offloading overheads of in/less-efficiently
designed MCC solutions can remarkably inten-sify the energy
limitations of mobile devices. Therefore, energy-awareness using
lightweight mobile augmentation frame-works [51],
communication-aware approaches [52], and fidelity adaptation [53]
solutions are essential to meet theenergy conservation in mobile
devices. Lightweight solutions could mitigate the offloading
overheads (processing andlatency) [54] and ensure that the
communication cost would not exceed the data computation benefits
of mobile devices.Similarly, optimal communication-aware and
fidelity adaptation solutions alleviate the energy efficiency issue
by leveragingheterogeneous wireless networks (i.e., cellular and
WLAN) for an energy-efficient trade-off for lightweight MCC
solutions.5 https://www.dropbox.com.6 http://docs.google.com.7
https://www.icloud.com.8 www.skydrive.com.
http://https://www.dropbox.comhttp://docs.google.comhttp://https://www.icloud.comhttp://www.skydrive.com
-
106 N. Aminzadeh et al. / Simulation Modelling Practice and
Theory 50 (2015) 961085.2. Data integrity
Although the anywhere, anytime, any device principal of MCC is
beneficial to mobile device users, the variation and inho-mogeneity
of storage services provided by varied cloud vendors cause data
integrity issue and necessitates additional inves-tigation of MCC
storage augmentation approaches. Data integrity issue jeopardizes
the reliability of cloud service providersand rises data retrieval
issue as a consequence. To mitigate data integrity issue, efforts
similar to [13] seek to provide a uni-form interface to access
various cloud storage services by exploiting middleware, SOA, and
domain specific languageapproaches. Moreover, management of data,
which are distributed in various cloud storage services [30] is
promising toovercome data integrity issues in MCC. However, further
research is required for a unified, standard storage
infrastructureand management of heterogeneity to alleviate data
migration.
5.3. Trust
One of the most important concerns of the mobile end-users is to
establish trust for the largely growing cloud vendors.Although
several studies have been conducted for building trust in cloud
resources [15,5557], the transmission of user dataover insecure
wireless networks (where stern supervision is lacking) and the
Internet for storage of data in the cloud (whereusers do not have
any control) is inevitable. Heterogeneity of cloud infrastructures
as well as mobile devices resource pov-erty and intrinsic traits
intensify trust issue. Protection of user data by leveraging
encryption and decryption techniques [34],authorization, and
authentication are among the approaches try to overcome trust issue
in MCC. However, a necessity fornovel trust establishment
techniques similar to [58] exists which are required to be
lightweight with respect to limitedresources of mobile devices and
wireless communication in MCC domain.
5.4. Data portability
Harnessing cloud resources as a comparatively more reliable and
safe storage for data has alleviated the data lock-inproblem of
mobile device users. Hence, the migration of data for messages and
contact information between non-uniformmobile devices is
facilitated. However, data portability is still an issue for cloud
consumers and due to this problem, usersare unable to transfer
their data from Android-based to iOS-based mobile devices.
Moreover, the heterogeneity of cloud ser-vices offered by diverse
cloud vendors causes a vendor lock-in problem. Vendor lock-in is an
attractive issue in business [59],but a real concern for customers.
Vendor lock-in makes customers dependent on services provided by
specific cloud vendors,as the incompatibility with other services
or the consequence of cloud service providers policies.
Homogeneity between the existing cloud services architectures
and programming languages is desired by end-users toovercome code
and data portability concerns. Solutions including adapter and
particularly middleware [60] as well as stan-dardization such as
Open Cloud Computing Interface (OCCI)9 are among approaches
addressing portability issues in cloudcomputing. MCC data
portability issues grant for further research that considers
variant clouds policies, architectures, and stor-age structures to
alleviate imposed challenges on the migration of data between
dissimilar cloud storages and mobile devices.
6. Conclusions
This paper surveys the crucial intrinsic restrictions of mobile
devices and storage augmentation issues in three domains ofmobile
computing, cloud computing andMCC to devise a taxonomy of issues as
the motivation for the emergence of effectiveand efficient MSA
approaches in MCC. A number of approaches leverage data
partitioning whereas other approaches exploitdata replication,
cache management, or SOA. Based on a review of the credible MSA
approaches, the paper proposes a tax-onomy of cloud-based
storage.
A mobile devices local resource conservation, computational
augmentation, and storage extension have been realizedthrough the
convergence of mobile and cloud computing that has produced
state-of-the-art MCC. This association and allits inestimable
privileges originate multi-dimensional heterogeneity in various
domains, including platforms, operating sys-tems, networks, and
data structures. Pivotal sources of variations produce the
generation of several non-trivial challenges formobile data,
including portability, interoperability, integrity, and security.
Therefore, the need for lightweight, energy- andcommunication-aware
MSA approaches is vital for the successful adoption of mobile
computing. Mobility is one of the prom-inent factors affecting
other factors in taxonomizing cloud-based storage resources in a
positive or negative manner. Hence,additional efforts are required
to address a number of crucial MCC storage augmentation open
issues. Energy-awareness,data integrity, trust, and data
portability are open issues, which specify future research
directions in this area.
Acknowledgments
This work is funded by the University of Malaya Research Grant
RP005D13ICT and by the Malaysian Ministry of HigherEducation under
the University of Malaya High Impact Research Grant
UM.C/625/1/HIR/MOE/FCSIT/03.9 http://occi-wg.org/.
http://occi-wg.org/
-
N. Aminzadeh et al. / Simulation Modelling Practice and Theory
50 (2015) 96108 107References
[1] W. Rupert, Mobile Data will Grow 6.3 Times between 2013 and
2018 and Growth will Be Strongest Outside Europe and North America.
.
[2] J. Wan, C. Zou, C.-F. Lai, M. Zhou, X. Wang, Cloud-enabled
wireless body area networks for pervasive healthcare, IEEE Network
27 (5) (2013) 5661.[3] H.-Y. Kung, C.-H. Chen, H.-H. Ku, Designing
intelligent disaster prediction models and systems for debris-flow
disasters in Taiwan, Expert Syst. Appl. 39
(5) (2012) 58385856.[4] R. Ferzli, I. Khalife, Mobile cloud
computing educational tool for image/video processing algorithms,
in: IEEE Digital Signal Processing and Signal
Processing Education Meeting (DSP/SPE), 2011, pp. 529533.[5] M.
Chen, S. Mao, Y. Liu, Big data: a survey, Mobile Networks Appl. 19
(2) (2014) 171209.[6] S. Abolfazli, Z. Sanaei, A. Gani, F. Xia,
L.T. Yang, Rich mobile applications: genesis, taxonomy, and open
issues, J. Network Comput. Appl. 40 (2013) 345
362.[7] S.S. Ali Feizollah, Nor Badrul Anuar, Rosli Salleh,
Fairuz Amalina, Rauf Ridzuan Maarof, A study of machine learning
classifiers for anomaly-based
mobile botnet detection, Malaysian J. Comput. Sci. 26 (4) (2014)
251265.[8] N.B. Anuar, M. Papadaki, S. Furnell, N. Clarke, Incident
prioritisation using analytic hierarchy process (AHP): Risk Index
Model (RIM), Secur. Commun.
Networks 6 (9) (2013) 10871116.[9] S. Sakr, A. Liu, D.M.
Batista, M. Alomari, A survey of large scale data management
approaches in cloud environments, IEEE Commun. Surv. Tutorials
13
(3) (2011) 311336.[10] A. Ranabahu, A. Sheth, Semantics centric
solutions for application and data portability in cloud computing,
in: IEEE Second International Conference on
Cloud Computing Technology and Science(CloudCom), Dayton, OH,
USA, 2010, pp. 234241.[11] A. Sheth, A. Ranabahu, Semantic modeling
for cloud computing, part 1, IEEE Internet Comput. 14 (3) (2010)
8183.[12] M. Shiraz, S. Abolfazli, Z. Sanaei, A. Gani, A study on
virtual machine deployment for application outsourcing in mobile
cloud computing, J.
Supercomput. 63 (3) (2012) 946964.[13] H. Mao, N. Xiao, W. Shi,
Y. Lu, Wukong: a cloud-oriented file service for mobile Internet
devices, J. Parallel Distrib. Comput. 72 (2) (2012) 171184.[14] Z.
Sanaei, S. Abolfazli, A. Gani, R.H. Khokhar, Tripod of requirements
in horizontal heterogeneous mobile cloud computing, in: WSEAS
International
Conference on Computing, Information Systems and Communications
(CISCO12), Singapore, 2012.[15] Z. Sanaei, S. Abolfazli, A. Gani,
M. Shiraz, SAMI: service-based arbitrated multi-tier infrastructure
for mobile cloud computing, in: IEEE International
Conference on Communications in China Workshops (ICCC), Beijing,
China, 2012, pp. 1419.[16] S. Abolfazli, Z. Sanaei, E. Ahmed, A.
Gani, R. Buyya, Cloud-based augmentation for mobile devices:
motivation, taxonomies, and open challenges, IEEE
Commun. Surv. Tutorials 16 (1) (2013) 337368.[17] S. Abolfazli,
Z. Sanaei, A. Gani, Mobile cloud computing: a review on smartphone
augmentation approaches, in: WSEAS International Conference on
Computing, Information Systems and Communications (CISCO12),
Singapore, 2012.[18] M. Shiraz, A. Gani, R.H. Khokhar, R. Buyya, A
review on distributed application processing frameworks in smart
mobile devices for mobile cloud
computing, IEEE Commun. Surv. Tutorials 15 (3) (2013)
12941313.[19] Z. Sanaei, S. Abolfazli, A. Gani, R. Buyya,
Heterogeneity in mobile cloud computing: taxonomy and open
challenges, IEEE Commun. Surv. Tutorials 16
(1) (2013) 369392.[20] Gartner Reveals Top Predictions for IT
Organizations and Users for 2012 and Beyond, 2011. .[21] W. Zheng,
P. Xu, X. Huang, N. Wu, Design a cloud storage platform for
pervasive computing environments, Cluster Comput. 13 (2) (2010)
141151.[22] P. Shor, Algorithms for quantum computation: discrete
logarithms and factoring, in: Proceedings IEEE 35th Annual
Symposium on Foundations of
Computer Science, 1994, pp. 124134.[23] P.W. Shor,
Polynomial-time algorithms for prime factorization and discrete
logarithms on a quantum computer, SIAM J. Comput. 26 (5) (1997)
1484
1509.[24] C. Metz, IBM Busts Record for Superconducting Quantum
Computer, 2012. .[25] J.F.R. Darin Chan, Summarisation for mobile
databases, J. Res. Pract. Inf. Technol. 37 (3) (2005) 267284.[26]
Y. Jun, L. Shan, X. Lizhen, Research of cache mechanism in mobile
data management, in: Proceedings IEEE Ninth Web Information Systems
and
Applications Conference, 2012, pp. 159162.[27] H. Hpfner, E.
Mansour, D. Nicklas, Review of data management mechanisms on mobile
devices, Inf. Technol. 51 (2) (2009) 7984.[28] A. Vancea, B.
Stiller, CoopSC: a cooperative database caching architecture, in:
19th IEEE International Workshops on Enabling Technologies:
Infrastructures for Collaborative Enterprises, Larissa, 2010,
pp. 223228.[29] K. Goda, M. Kitsuregawa, The history of storage
systems, Proc. IEEE 100 (Special Centennial Issue) (2012)
14331440.[30] R. Buyya, C.S. Yeo, S. Venugopal, J. Broberg, I.
Brandic, Cloud computing and emerging IT platforms: vision, hype,
and reality for delivering computing as
the 5th utility, Future Gener. Comput. Syst. 25 (6) (2009)
599616.[31] P. Mell, T. Grance, The NIST Definition of Cloud
Computing, 2011. .[32] N. Cordeschi, M. Shojafar, E. Baccarelli,
Energy-saving self-configuring networked data centers, Comput.
Networks 57 (17) (2013) 34793491.[33] M. Armbrust, A.D. Joseph,
R.H. Katz, D.A. Patterson, Above the Clouds: A Berkeley View of
Cloud Computing, Tech. Rep. UCB/EECS-2009-28, EECS
Department, University of California, Berkeley, 2009.[34] Zhibin
Zhou, Dijiang Huang, Efficient and secure data storage operations
for mobile cloud computing, in: Network and Service Management
(CNSM),
2012 8th International Conference and 2012 Workshop on Systems
Virtualiztion Management (SVM), Las Vegas, NV, 2012, pp. 3745.[35]
R.K. Panta, R. Jana, F. Cheng, Y.-F.R. Chen, V.A. Vaishampayan,
Phoenix: storage using an autonomous mobile infrastructure, IEEE
Trans. Parallel Distrib.
Syst. 24 (9) (2013) 18631873.[36] P. Stuedi, I. Mohomed, D.
Terry, WhereStore: location-based data storage for mobile devices
interacting with the cloud, in: MCS 10 Proceedings of the
1st ACM Workshop on Mobile Cloud Computing & Services:
Social Networks and Beyond, New York, NY, USA, 2010.[37] B. Jedari,
F. Xia, A survey on routing and data dissemination in opportunistic
mobile social networks, IEEE Commun. Surv. Tutorials (99). doi:
10.1109/
SURV.2014.022714.00153.[38] A. Moon, H. Cho, Energy efficient
replication extended database state machine in mobile ad hoc
network, in: IADIS International Conference on Applied
Computing, LISBON, PORTUGAL, 2004.[39] E.R. Sanchez, L.M.
Murillo, B. Montrucchio, M. Rebaudengo, An adaptive power-aware
multi-hop routing algorithm for wireless sensor networks, in:
Eighth IEEE International Conference on Information Technology:
New Generations (ITNG), 2011, pp. 112116.[40] Jun-Zhao Sun, Mobile
ad hoc networking: an essential technology for pervasive computing,
in: IEEE International Conferences on Info-Tech and Info-
Net. Proceedings, vol. 3, Beijing, China, 2001, pp. 316321.[41]
F. Pedone, R. Guerraoui, The database state machine approach,
Distrib. Parallel Databases 14 (1) (2003) 7198.[42] Y. Lu, B. Zhou,
L.-C. Tung, M. Gerla, A. Ramesh, L. Nagaraja, Energy-efficient
content retrieval in mobile cloud, in: Proceedings of the Second
ACM
SIGCOMM Workshop on Mobile Cloud Computing MCC13, New York, USA,
2013.[43] S. Abolfazli, Z. Sanaei, M. Shiraz, A. Gani, MOMCC:
market-oriented architecture for mobile cloud computing based on
service oriented architecture, in:
1st IEEE International Conference on Communications in China
Workshops (ICCC), 2012, pp. 813.[44] M. Asif, S. Majumdar, R.
Dragnea, Hosting web services on resource constrained devices, in:
IEEE International Conference on Web Services (ICWS
2007), 2007, pp. 583590.[45] M. Ok, D. Kim, M.-s. Park,
UbiqStor: a remote storage service for mobile devices, Parallel
Distrib. Comput.: Appl. Technol. 3320 (2005) 685688.
http://www.analysysmason.com/About-Us/News/Insight/Mobile-data-Oct2013/http://www.analysysmason.com/About-Us/News/Insight/Mobile-data-Oct2013/http://refhub.elsevier.com/S1569-190X(14)00079-3/h0130http://refhub.elsevier.com/S1569-190X(14)00079-3/h0135http://refhub.elsevier.com/S1569-190X(14)00079-3/h0135http://refhub.elsevier.com/S1569-190X(14)00079-3/h0140http://refhub.elsevier.com/S1569-190X(14)00079-3/h0145http://refhub.elsevier.com/S1569-190X(14)00079-3/h0145http://refhub.elsevier.com/S1569-190X(14)00079-3/h0150http://refhub.elsevier.com/S1569-190X(14)00079-3/h0150http://refhub.elsevier.com/S1569-190X(14)00079-3/h0155http://refhub.elsevier.com/S1569-190X(14)00079-3/h0155http://refhub.elsevier.com/S1569-190X(14)00079-3/h0160http://refhub.elsevier.com/S1569-190X(14)00079-3/h0160http://refhub.elsevier.com/S1569-190X(14)00079-3/h0165http://refhub.elsevier.com/S1569-190X(14)00079-3/h0170http://refhub.elsevier.com/S1569-190X(14)00079-3/h0170http://refhub.elsevier.com/S1569-190X(14)00079-3/h0175http://refhub.elsevier.com/S1569-190X(14)00079-3/h0180http://refhub.elsevier.com/S1569-190X(14)00079-3/h0180http://refhub.elsevier.com/S1569-190X(14)00079-3/h0185http://refhub.elsevier.com/S1569-190X(14)00079-3/h0185http://refhub.elsevier.com/S1569-190X(14)00079-3/h0190http://refhub.elsevier.com/S1569-190X(14)00079-3/h0190http://www.gartner.com/newsroom/id/1862714http://refhub.elsevier.com/S1569-190X(14)00079-3/h0195http://refhub.elsevier.com/S1569-190X(14)00079-3/h0200http://refhub.elsevier.com/S1569-190X(14)00079-3/h0200http://www.wired.com/wiredenterprise/2012/02/ibm-quantum-milestone/http://www.wired.com/wiredenterprise/2012/02/ibm-quantum-milestone/http://refhub.elsevier.com/S1569-190X(14)00079-3/h0205http://refhub.elsevier.com/S1569-190X(14)00079-3/h0210http://refhub.elsevier.com/S1569-190X(14)00079-3/h0215http://refhub.elsevier.com/S1569-190X(14)00079-3/h0220http://refhub.elsevier.com/S1569-190X(14)00079-3/h0220http://csrc.nist.gov/publications/nistpubs/800-145/SP800-145.pdfhttp://refhub.elsevier.com/S1569-190X(14)00079-3/h0225http://refhub.elsevier.com/S1569-190X(14)00079-3/h0230http://refhub.elsevier.com/S1569-190X(14)00079-3/h0230http://refhub.elsevier.com/S1569-190X(14)00079-3/h0235http://refhub.elsevier.com/S1569-190X(14)00079-3/h0240
-
108 N. Aminzadeh et al. / Simulation Modelling Practice and
Theory 50 (2015) 96108[46] P. Padmanabhan, L. Gruenwald, A. Vallur,
M. Atiquzzaman, A survey of data replication techniques for mobile
ad hoc network databases, VLDB J. 17 (5)(2008) 11431164.
[47] K. Wang, B. Li, Efficient and guaranteed service coverage
in partitionable mobile ad-hoc networks, in: Proceedings of IEEE
INFOCOM, vol. 2, 2002, pp.10891098.
[48] J.-H. Kim, B.-S. Moon, M.-S. Park, MiSC: a new availability
remote storage system for mobile appliance, Lecture Notes in
Computer Science, vol. 3421,Springer, Berlin, Heidelberg, 2005, pp.
504520.
[49] M. Satyanarayanan, P. Bahl, R. Caceres, N. Davies, The case
for VM-based cloudlets in mobile computing, IEEE Pervasive Comput.
8 (4) (2009) 1423.[50] S. Abolfazli, Z. Sanaei, M. Alizadeh, A.
Gani, F. Xia, An experimental analysis on cloud-based mobile
augmentation in mobile cloud computing, IEEE
Trans. Consumer Electron. 60 (1) (2014) 146154.[51] M. Shiraz,
A. Gani, A lightweight active service migration framework for
computational offloading in mobile cloud computing, J. Supercomput.
68 (2)
(2014) 978995.[52] K. Kumar, Cloud computing for mobile users:
can offloading computation save energy?, Computer 43 (4) (2010)
5156[53] R. Balan, Powerful change part 2: reducing the power
demands of mobile devices, IEEE Pervasive Comput. 3 (2) (2004)
7173.[54] M. Sharifi, S. Kafaie, O. Kashefi, A survey and taxonomy
of cyber foraging of mobile devices, IEEE Commun. Surv. Tutorials
14 (4) (2012) 12321243.[55] M. Mowbray, S. Pearson, A client-based
privacy manager for cloud computing, in: Proceedings ACM of the
Fourth International ICST Conference on
COMmunication System softWAre and middlewaRE COMSWARE 09, New
York, USA, 2009.[56] Q. Wang, C. Wang, J. Li, K. Ren, W. Lou,
Enabling public verifiability and data dynamics for storage
security in cloud computing, in: ESORICS09
Proceedings of the 14th European Conference on Research in
Computer Security, Saint Malo, France, 2009, pp. 355370.[57] S.
Shamshirband, N.B. Anuar, M.L.M. Kiah, A. Patel, An appraisal and
design of a multi-agent system based cooperative wireless intrusion
detection
computational intelligence technique, Eng. Appl. Artif. Intell.
26 (9) (2013) 21052127.[58] W. Wang, Z. Li, R. Owens, B. Bhargava,
Secure and efficient access to outsourced data, in: Proceedings of
the 2009 ACM Workshop on Cloud Computing
Security CCSW 09, ACM Press, New York, USA, 2009, p. 55.[59]
S.S. Manvi, G. Krishna Shyam, Resource management for
Infrastructure as a Service (IaaS) in cloud computing: a survey, J.
Network Comput. Appl. 41
(2014) 424440.[60] G. Blair, P. Grace, Emergent middleware:
tackling the interoperability problem, IEEE Internet Comput. 16 (1)
(2012) 7882.
http://refhub.elsevier.com/S1569-190X(14)00079-3/h0245http://refhub.elsevier.com/S1569-190X(14)00079-3/h0245http://refhub.elsevier.com/S1569-190X(14)00079-3/h0250http://refhub.elsevier.com/S1569-190X(14)00079-3/h0250http://refhub.elsevier.com/S1569-190X(14)00079-3/h0250http://refhub.elsevier.com/S1569-190X(14)00079-3/h0255http://refhub.elsevier.com/S1569-190X(14)00079-3/h0260http://refhub.elsevier.com/S1569-190X(14)00079-3/h0260http://refhub.elsevier.com/S1569-190X(14)00079-3/h0265http://refhub.elsevier.com/S1569-190X(14)00079-3/h0265http://refhub.elsevier.com/S1569-190X(14)00079-3/h0270http://refhub.elsevier.com/S1569-190X(14)00079-3/h0275http://refhub.elsevier.com/S1569-190X(14)00079-3/h0280http://refhub.elsevier.com/S1569-190X(14)00079-3/h0285http://refhub.elsevier.com/S1569-190X(14)00079-3/h0285http://refhub.elsevier.com/S1569-190X(14)00079-3/h0290http://refhub.elsevier.com/S1569-190X(14)00079-3/h0290http://refhub.elsevier.com/S1569-190X(14)00079-3/h0290http://refhub.elsevier.com/S1569-190X(14)00079-3/h0295http://refhub.elsevier.com/S1569-190X(14)00079-3/h0295http://refhub.elsevier.com/S1569-190X(14)00079-3/h0300
Mobile storage augmentation in mobile cloud computing: Taxonomy,
approaches, and open issues1 Introduction2 Motivation2.1 Mobile
device issues2.2 Cloud-based issues2.3 Converged issues
3 Mobile-cloud storage augmentation approaches4 Cloud-based
storage characteristics5 Open issues5.1 Energy-awareness5.2 Data
integrity5.3 Trust5.4 Data portability
6 ConclusionsAcknowledgmentsReferences