On the Next Generations of Infrastructure-as-a-Services

Dana Petcu

, Maria Fazio

, Radu Prodan

, Zhiming Zhao

and Massimiliano Rak

Institute e-Austria and Computer Science Department, West University of Timis¸ oara, 4 V. P

arvan, Timis¸oara, Romania

Dicieama Department, University of Messina, 98166 Sant Agata, Messina, Italy

Institute for Computer Science, University of Innsbruck, 21a Technikerstrasse, Innsbruck, Austria

System and Network Engineering, University of Amsterdam, 904 Science Park, Amsterdam, Netherlands

Department of Industrial Engineering and Information, Second University of Naples, 29 Sede di via Roma, Aversa, Italy

Keywords:

Infrastructure-as-a-Service, Surveys, Previsions.

Abstract:

Following the wide adoption by industry of the cloud computing technologies, we can talk about a second

generation of cloud services and products that are currently under design phase. However, it is not yet clear

how the third generation of cloud products and services of the next decade will look like, especially at the

delivery level of Infrastructure-as-a-Service. In order to answer at least partially to such a challenging question,

we initiated a literature overview and two surveys involving the members of a cluster of European research

and innovation actions. The results are interpreted in this paper and a set of topics of interest for the third

generation are identiﬁed.

1 INTRODUCTION

After a decade in which cloud computing has domi-

nated the interest of the communities involved in dis-

tributed systems, we ask ourself what can bring more

this paradigm in the next decade.

The ﬁrst generation of cloud computing products,

following the stages of technology triggers, peaks of

inﬂated expectations, as well as through of disillu-

sionment, is based on the concept of centralized pool

of services, the expectation that one-size-ﬁts-all, and

intensive use of virtualization technologies. The de-

velopment focus was put on building the proper tools

to offer the provider-side service management at costs

affordable by their clients.

Triggered by various factors, like user’s special re-

quirements, integration with other technologies, con-

cerns like vendor-lock-in, interoperability or security,

we assist nowadays to the raise of a second-generation

of more mature cloud services, technologies and pro-

ducts. This generation is characterized by a go-to

model for enterprises, born-in-the-cloud applications,

decentralization tendencies, integration with Internet

of Things, Big Data or mobile technologies, user-

driven, software-deﬁned or energy-efﬁcient service

management and orchestration, and even the shrink-

ing or dissolution of the virtual machines with the up-

rise of containers or unikernels. We look forward to

see what the third generation of Cloud products will

bring out-of-the-box, as well when the user adoption

and acceptance is world-wide spread so that the cloud

can be declared ubiquitous. However, the research ac-

tivities that are supporting the new generation should

start today and challenges in front should be identi-

ﬁed as soon as possible. This paper is aligned to this

desire. It is trying to provide a peek into a possible

version of the future, by following the opinions of the

specialists working currently in research activities re-

lated to Infrastructure-as-a-Service (IaaS). The result

is a list of topics of interest for future research.

The paper is organized as follows. The next sec-

tion describes the methodology that was used. Sec-

tion 3 discusses several topics that are currently on

the research desks. Section 4 issues some hypothesis

about the topics of interest for the third generation.

Conclusions can be found in the last section.

2 METHODOLOGY

The sources of information and the methodology used

in this paper to elaborate on future research directions

are discussed on what follows.

Sources of Information and Stages. In a ﬁrst

stage of the study we have have appealed to the fol-

320

Petcu, D., Fazio, M., Prodan, R., Zhao, Z. and Rak, M.

On the Next Generations of Infrastructure-as-a-Services.

In Proceedings of the 6th International Conference on Cloud Computing and Services Science (CLOSER 2016) - Volume 1, pages 320-326

ISBN: 978-989-758-182-3

lowing sources: (1) research papers indexed in Sci-

enceDirect and IEEE Xplore catalogs from the pe-

riod 2014-2016 which are identifying challenges and

open issues related to the evolution of IaaS; (2) cloud-

related blogs and news provided by cloud service

owners or industry-related journals; (3) the text of the

most recent call for research and innovation actions in

the ﬁeld of cloud computing (European Commission,

2015), in the frame of H2020-ICT, as result of a long

process of specialists and decision-makers consulta-

tions; plus, occurrences of its keywords in SCOPUS

catalog of scientiﬁc articles and in the searches done

using Google Trends. The result is a list of topics

related to IaaS supposed to reﬂect the present, near

future and far future activities in the ﬁeld.

In a second stage, the opinion of the specialists

involved in 20 actions

partially funded be the Eu-

ropean Commission in the frame of FP7-ICT, CIP

and H2020-ICT programmes was requested. Their

research subjects are varying from heterogeneous

clouds or DevOps for multiple clouds, virtualization

technologies or security in clouds to IoT-cloud inte-

gration, cloud-based Big Data, or mobile cloud. Two

surveys were pursued respectively in October and

November 2015, to identify the followings: (1) the

on-going research activities that are expected to sup-

port the second generation services (with 16 contribu-

tors, 62% from academia and 38% from entreprises);

(2) the foreseen research challenges in front of the

third generation services (with 42 contributors, from

which 20% developers, 24% software architects, 81%

researchers, 27% managers). The results are pre-

sented in two on-line documents available on the web

site (InfraCluster, 2015). In this paper, we are com-

menting, linking and interpreting further the results.

Complementarity with other Studies. A com-

plementary set of topics for future research was iden-

tiﬁed by (Jeferry et al., 2015) under the title of tech-

nological challenges and goals for the evolution of

cloud services and environments. The considered ap-

proach is different from ours as it starts from a number

of application scenarios for which the requirements

cannot be met by current cloud technologies, but can

drive the evolution of future cloud technologies. The

AppHub: apphub.eu.com, ARCADIA: arcadia-

framework.eu, BEACON: beacon-project.eu, CloudLight-

ning: loudlightning.eu, CloudSpaces:/cloudspaces.eu,

ClouT: clout-project.eu, CloudWave: cloudwave-fp7.eu,

DICE: dice-h2020.eu, ENTICE: entice-project.eu, iiKaaS:

www.ikaas.com, INPUT: input-project.eu, IOStack:

iostack.eu, MCN: mobile-cloud-networking.eu, MIKE-

LANGELO: mikelangelo-project.eu, Mo-Bizz: mobizz-

project.eu, MODAClouds: modaclouds.eu, MUSA:

musa-project.eu, RAPID: rapid-project.eu, SPECS:

specs-project.eu, SWITCH: switch-project.eu

identiﬁed scenarios are referring to: joint collabora-

tive business intelligence platforms with multiple data

sources; born-digital data interoperability through se-

mantic technologies; awareness of the application

characteristics enabling the infrastructure to provide

resources that accommodate the requirements that are

evolving during runtime; model-driven integration in

cloud IDEs; joint application and infrastructure con-

trollability; quality-of-service/experience-critical ap-

plications on the cloud; adaptable parametric applica-

tions; generic application templates.

The industry decision makers are trying also to

ﬁgure out where cloud computing is going in the near

future. Articles like (Yousif et al., 2014) are reﬂec-

ting very well the main current concerns that should

be remove in the near future (again a different ap-

proach from that of the current paper). The present

is driven by the requirements like: single catalog or

service-management interfaces allowing smooth tran-

sitions between clouds; systems of engagements for

agile type of application and application develop-

ment; short development cycles and continuous deli-

very model (DevOps); compliance, security, and con-

trol measures; spanning from customer premises to

the public cloud (including hybrid cloud). The line

between the infrastructure and the application is ex-

pected to blur, as the line between development and

quality assurance and operations, as well as the line

between compute, networking, and storage.

We will try in what follows to not touch the sub-

jects already tackled by the above two cited studies.

3 THE SECOND GENERATION:

NEAR FUTURE RESEARCH

We focus in this section on the expectations of the

second generation in terms ranging from challenges

to be solved to technologies under use.

Integration Challenges. Examining the recent

research reports, we observed that a large number

of activities are focusing on the integration of cloud

computing with other technologies. We are pointing

here towards several surveys on challenges.

Vertical Tailored for Speciﬁc Industry. The paper

(James and Chung, 2015) is revealing a new direction

currently followed by cloud service providers to tai-

lor their services speciﬁcally for their customer and

their industrial needs. The main challenges are re-

lated to the provision of tailored solution to various

levels, from infrastructure services, growth patterns,

or software functionality, to privacy and security, as

well as to follow the requirement to interoperate with

third party community tools.

On the Next Generations of Infrastructure-as-a-Services

321

CloudIoT Paradigm. Cloud and IoT are merged

together enabling a large number of application sce-

narios (Botta et al., 2016). Several new services are

added to the infrastructure level: Things, Sensing,

Sensing and Actuation, Sensor Event, Sensor, Video

Surveillance -as-a-Service. The main challenges are

related to privacy, legal and social aspects, large scale,

security, reliability, performance, heterogeneity. The

problems to be solved are referring to identiﬁcation,

addressing, naming, environmental context-based ser-

vices, large scale support for multi-networking, ﬂexi-

ble mechanisms for logically isolated network parti-

tions. convergence to a common open service plat-

form environment to develop applications.

Cloud Providing Capabilities for Big Data. Sev-

eral recent surveys are analyzing the approaches, en-

vironments, and technologies on areas that are key to

Big Data analytics capabilities and discuss how they

help building analytics solutions on clouds. In par-

ticular (Assuno et al., 2015) focuses on technical is-

sues of enabling cloud analytics and indicates a set of

gaps and recommendations on future research direc-

tions on cloud-supported Big Data.

Mobile Cloud Computing. Migrating an applica-

tion processing to the cloud data centers enables the

execution of resource-intensive applications on the

mobile devices. However, the resource-intensive mi-

gration approaches and the limitations of the wire-

less medium hinders the applications from attaining

optimal performance in the cloud. Executing the

application with low cost, minimal overhead, and

non-obtrusive migration is a challenge. The paper

(Ahmed et al., 2015) presents the state-of-the-art of

mobile application execution frameworks and dis-

cuss the optimization strategies facilitating effective

design, efﬁcient deployment, or application migra-

tion with optimal performance. The identiﬁed chal-

lenges are referring to: optimal application and execu-

tion framework design, efﬁcient deployment and user-

transparent execution, real time optimized manage-

ment of heterogeneous environments, automated ser-

vice provisioning, scalability, availability of services.

Secure the Cloud. The latest surveys, e.g. (Ali

et al., 2015) or (Tari et al., 2015), presents recent

solutions reported in the literature to overcome the

security issues. In particular, the latter explores the

challenges of cloud security, with a special focus on

data utilization management aspects, including data

storage, data analytics, and access control.

Cloud Networking. The functional problems as-

sociated with the evolution of cloud computing from

the network perspective are revealed in (Moura and

Hutchison, 2016). If the application could inform the

providers how its trafﬁc should be treated, an applica-

tion performance improvement would be possible.

User-driven Cloud Service Management. The pa-

per (Rehman et al., 2015) presents a comprehensive

discussion on the existing approaches to cloud ser-

vice management, and evaluates them against the fac-

tors required for the user to manage the cloud ser-

vice. While more automation for load management

and moving workloads around are achieved, under-

standing and controlling what is going-on gains in im-

portance (Yousif et al., 2014). Quality metrics are re-

deﬁned from an user perspective; for example (Nabi

et al., 2016) surveys the current mechanisms and me-

trics for availability and found that most solutions do

not provide protection against application failures.

Decentralized Models for Clouds. The paper

(Khan et al., 2015) studies alternative models of cloud

computing, like the community clouds at the edge,

volunteer-, edge-, social- or mobile-cloud computing.

The requirements for a community cloud system are

related to autonomy, security, self-management, ease

of use, social and economic mechanisms, support for

heterogeneity, standard middleware interfaces.

Software-deﬁned Environment. It enables the abs-

traction from the infrastructure, making the infras-

tructure programmable. It is the main driver for cloud

automation. OpenStack is a central element to this

concept, according (Yousif et al., 2014).

Energy-aware Services. The paper (Mastelic and

Brandic, 2015) surveys the current energy-efﬁciency

approaches, challenges, and future directions.

Research Topics of Interest for the Near Fu-

ture. Taking into account that the key-words/phrases

of the next call for actions exposed by in the workpro-

gramme H2020-ICT (European Commission, 2015)

are the results of several rounds of involved commu-

nities’ consultancy, we consider them in this paper as

buzzwords for the research activities to be performed

in the near future, and, consequently, representative

for the second generation of products.

We estimate the volume of the activities around

these buzzwords as follows: (a) to reﬂect the current

interest, the volume of searches reported by Google

Trends using the keywords; (b) to reﬂect the solutions

already available, the volume of result reports regis-

tered in SCOPUS and related to the keywords; (c) to

reﬂect the incoming solutions, the on-going activities

related to the keywords in the cluster of projects men-

tioned earlier. The results are normalized relative to

the highest number of hits/values, in order to express

a relative importance provided to certain topics.

Figure 1 provides a graphical representation of the

result. Perfect match between the three percents as-

sociated with a key-word/phrase is not encountered,

as the subjects for the on-going research activities

CLOSER 2016 - 6th International Conference on Cloud Computing and Services Science

322

Figure 1: Interest in the current topics following the key-

words: in SCOPUS (work already done by researchers); in

Google Trends (interest for the future of a large commu-

nity); in the cluster of projects (on-going work).

Figure 2: Percents of respondents who are positioning the

selected topics in the present (green and magenta).

are evolving compared with the previous ones and do

not match exactly the hot topics depicted by peaks in

search volumes (subjects for the next generation).

Current technologies. Following the second sur-

vey, Figure 2 is pointing towards several subjects of

current research interest that are foreseen to lead to

second generation products in near future.

Containers. Are used for application distribution,

cost effectiveness, security and portability. It is ex-

pected to generate a new focus on automation tools,

including the interest to provide automated compo-

nent portability based on pre-deﬁned criteria. The

main expectation is to reduce complexity by using

container abstractions.

Continuous Development of Cloud-ready Appli-

cations. While the rapid development, agile or Dev-

Ops are now standard procedures, continuous appli-

cation development is an emerging trend. It involves

the constantly move of new releases into production,

while each iteration of an application is staged for

deployment once testing is completed. The steps in

the process are usually small and the iterations move

Table 1: Most used tools and standards.

Tools % Standards %

OpenStack 92.86 CloudML 28.57

KVM 50.00 TOSCA 21.43

Chef 42.86 Openﬂow 21.43

Docker 35.71 WS-Agreement 21.43

VMWare 21.43 OCCI 14.29

Eucalyptus 21.43 OVF 14.29

Puppet 21.43 ETSI GAL 14.29

through the development pipeline quickly. Therefore

the model accelerates the overall development pro-

cess and improves efﬁciency.

Web-powered Applications. Developing cloud-

based applications compatible with multiple plat-

forms is becoming a must. With new initiatives for

Internet protocols and scripting, the Internet becomes

the main platform for these applications.

Fat Cloud Server. Infrastructure services are

reshaped to support servers allowing larger instan-

ces or more virtual machines per server, in-memory

database instances or remote direct access memory.

BYOD. The ﬂexibility demanded by the mobile

workforce is the main reason for the integration be-

tween the personal cloud services for the companies

employees in a Bring your own device (BYOD) en-

vironment. End users are currently using mobile de-

vices to put their data into personal cloud services for

storage, streaming or syncing.

Tools and Standards in Use. Following the ﬁrst

survey, Table 1 provides an inside on the percent of

initiatives that are using todays tools and standards for

the development of the second generation products.

4 THE THIRD GENERATION:

CURRENT EXPECTATIONS

Following the second survey, the main topics selected

as being of interest for the research and innovation to

create the third generation are depicted in Figure 3.

Note that the complex topics related to security, soft-

ware engineering, multiple clouds were excluded.

Support for a New Software Generation. We

estimate that the IaaS will evolve towards its omni-

presence, freeing the cloud service consuming soft-

ware from the cloud services, by ensuring that the new

generation software, or existing modular or event-

reactive ones, will be able to be described to an ab-

stract level that is service agnostic, will be able to

form automatic and transparent combinations of hard-

ware and software resources according to its needs,

while resource provisioning, deployments, runtime

migrations, multi-tenancy with cost-effectiveness and

data protection, or the recovery from minute-to-

On the Next Generations of Infrastructure-as-a-Services

323

Figure 3: Percents of respondents positioning the selected topics in the future (red and blue): top-left, support for new software

generation; bottom-left, IaaS model evolution; top-right, IaaS market evolution; bottom-right, Datacenter services

minute failures will be managed automatically. More

details are provided bellow.

Social-media Like Evolution. The applications

are expected to form automatic associations of hard-

ware and software resources according to their needs,

e.g. following a social-media style in which a

database can preferentially select a trusted server

or storage array (this approach requires that the re-

sources of a datacentre will mould itself around the

task required, rather than following the current oppo-

site way). Application developers will no more worry

about provisioning servers, storage, or communica-

tion as the provisioning process will happen automat-

ically.

Event-driven Architecture. This trend is reﬂected

in the recent Lambda architecture, in which code is

executed in response to an event, ensuring an impor-

tant step in moving away from server-centric design.

The cloud becomes a generic compute engine, and the

developers do not need to organize the resources as

they simple just run the code.

Failure Proof Systems. New software systems

need to be developed to deal with the likely minute-

to-minute failures of the consumed resources.

Thin Cloud Connection. Applications will no

more belong to only one cloud, but will be able to re-

side in various clouds and to use various on-premise

other applications. Different parts of applications will

be able to ﬂoat around in and out of current cloud, e.g.

following a policy-based SLA management plan.

Modular Cloud Applications. The software deve-

lopment process is expected to place an emphasis on

modular software, allowing their components to be

modiﬁed without shutting down the full.

Software Moving Away from Hardware. The soft-

ware is expected to ﬂoat away from the hardware

and software resources: the software will described

in abstracted space and will go through several adap-

tors before it interacts with hardware (being therefore

practically hardware agnostic).

Multi-tenancy in Software. The production of

multi-tenant software is challenging. When multiple

users access the software relying on one cloud, the

software developer needs to make sure that data is

kept separate and the associated costs are identiﬁed.

IaaS Model Evolution. The IaaS model will

evolve towards the integration with machine-to-

machine computing, solving the challenges of sepa-

rate technology stacks and dealing with limited me-

mory, storage and computation capacity of the edge

devices, speed of deployment, resource distribution,

cost-effective scalability, resilience, easiness of ma-

nagement or security.

Integration with Operational Technologies. A

challenging area is the merge of information techno-

logy with operational technology. The information

technology is established around the open-source, In-

ternet protocols, international standards. The opera-

tional technology is characterized by proprietary stan-

dards and non-cross functional equipment. This in-

cludes the hardware and the software that is used to

control and monitor the manufacturing equipment and

processes, and most communication is accomplished

between machines. The two groups have separate

technology stacks, network architectures, protocols,

standards, governance models, and organizations. By

CLOSER 2016 - 6th International Conference on Cloud Computing and Services Science

324

converging them in one solution, organizations can

create better products, achieve cost and risk reduc-

tions, improve performance, ﬂexibility and efﬁciency.

Fog Nodes Integration. The fog computing promi-

ses to bring small computation and storage capabili-

ties, supporting execution of applications that require

low latency interactions with sensors, actuators, end-

users. Fog nodes are implemented using embedded

systems, in industrial control boards or home routers.

Their limited memory, storage and computation is a

challenge in their integration into the cloud architec-

ture to enable the execution of application logic. Their

integration will leverage containers as main techno-

logy for application delivery and execution.

Smart Gateways. The success of fog computing

depends on the resilience of the smart gateways di-

recting tasks on an Internet teeming with IoT devices.

This smartening will rely on features such as out-of-

band access, automatic detection and recovery from

outages, cellular connectivity, or high-level security.

Moving the processing of data to the edge raises

the challenge to maintaining the availability of these

smart gateways and their communication path to the

cloud. Resilience is needed for business continuity,

with redundancy, security, monitoring of power and

cooling and failover solutions in place to ensure max-

imum uptime. Speed of deployment, cost-effective

scalability, and ease of management with limited re-

sources are also main concerns.

Inverse Cloud Models, like machine-to-machine

computing (M2M) or geo-distributed cloud, have

been proposed as an alternative to the bandwidth-

intensive cloud approach. Company branded ver-

sions, like fog computing, edge computing, in-

network distributed cloud, or Cloud 2.0 were pro-

posed in this context. M2M provides compute, sto-

rage, data or application services to client endpoints,

like the cloud computing model. Unlike it, M2M

distributes data transmission among endpoints and

routers, instead sending them to a server. However,

it has its challenges, including limited resources and

network capacity, security, or resource distribution.

IaaS Market Orientation. We estimate that the

IaaS will evolve towards diversiﬁcation, ensuring spe-

cial features like those sustaining user mobility, user

as service provider, service composition, personal

data service conﬁgurability, or speciality cloud ser-

vices. More details are provided bellow.

Personal Cloud – Conﬁgurable Services. The per-

sonal cloud providers are expected to focus on busi-

ness users along with offering superior value added

services for users, supplemented by the integration

of wearable technologies. The users will pay to ea-

sily deploy conﬁgurations and services they are ac-

quainted to use. The exploitation of containers in

embedded and resource constrained devices gives the

opportunity to encapsulate user preferences, applica-

tions and services in virtual environments and to de-

ploy them through the cloud in personal devices.

Users as Service Providers. Social tools will bring

increased collaboration to the cloud. As people be-

come more accustomed to these tools, customers use

them to communicate with fellows. The face of cloud

is expected to change into a solution which will be

delivered as a service with end users behaving more

like service providers as a result.

Specialized Clouds. Clouds will raise through

their ability to avoid specialized hardware acquisition

costs and outdated equipment. E.g. video editing will

create expanding markets for large instances and spe-

cialized GPU clouds. Running high-end graphics ap-

plications is requiring today substantial hardware in-

frastructure investment; newly cloud-based graphics

technologies are showing that end users can run high-

end graphic design applications with a web browser.

Personal Clouds – Data Services. Data genera-

tion activities among users and the need to access

data from anywhere using any device increases the

demand for personal cloud for real-time data access

and its sharing. Additionally, the growing need to cre-

ate a backup of critical data, along with provisions for

data recovery and planning for digital contingency, is

stimulating the personal cloud adoption.

Connecting Cloud Services. By connecting cloud

services, the companies should be able to effectively

manage their assets. However, a close look to which

providers offer cost-efﬁcient or reliable services is

needed. The ability to build agile hybrid clouds will

inﬂuence the design of the midleware.

Datacentre Services. We expect to see an evo-

lution of the software-deﬁned datacentres as ecosys-

tems, in which services are abstracted from infras-

tructure, changes and updating are done automatically

based on intelligent orchestration or new database

tools, security is software-deﬁned; such ecosys-

tem should enable warehouse-scale computing using

purpose-designed chips, new services like supercom-

puting on demand or massively federated, scalable

software architecture with orchestration through net-

work awareness. More details are provided bellow.

Warehouse-scale Computing. Low-power proces-

sors will be able to treat many workloads to support

massively federated and scalable software architec-

ture. A new generation of warehouse-scale comput-

ing is coming and custom chips for cloud partners and

purpose-designed chips as well as energy-aware ma-

nagement are expected to be supplied.

Ecosystems. The datacentres will be like biolo-

On the Next Generations of Infrastructure-as-a-Services

325

gical organisms: having different states, growing and

shrinking according to workloads, automatically cor-

rected and changed. An over-arching system will rule

equipment via software.

Software-deﬁned. The software-deﬁned datacen-

tres with all infrastructure virtualised and delivered

as a service is changing the way organisations view

the value of the underlying physical infrastructure.

Within the datacentre the services will be abstracted

from the infrastructure. Software-deﬁned security

will become part of the software-deﬁned datacentre.

Orchestration through Network Awareness. When

large-scale applications with thousands components

are deployed across multiple datacenters the SLAs

need to be enforced and software usage and faults

need to be monitored and managed. Cloud mana-

gement and orchestration must be aligned with var-

ious products and services. Consequently, everything

from orchestration to database tools will evolve. Dat-

acentre operators will add value to cloud orchestration

through network awareness and integration of cloud

orchestration with their management platforms.

Private Cluster of Demand. Technologies cur-

rently limited to supercomputing will make it into the

mainstream. The web-server-sized instances will be

replaced by on-demand private clusters. New cloud

services are expected to emerge, like supercomputers

on demand, high-performance storage, or new ways

of storing and processing data.

5 CONCLUSIONS

The paper touched the subject of predicting the un-

predictable, the far future of cloud, based on a limited

analysis of the trends in on-going research activities

in the European area and a small set of visionary pa-

pers. Based on them, we concluded that the evolution

of the IaaS model and market in the future will fol-

low four main directions: signiﬁcant changes in the

support of data application development, datacenters

reorganisation, model evolution towards smooth inte-

gration with other models, market orientation towards

specialization and personalization. Complex topics

like software engineering, security, multiple clouds

were intentionally excluded (future studies).

ACKNOWLEDGEMENTS

This research is partially supported by the grants RO-

PCE-0260-AMICAS and EC-H2020-ICT 643946-

CloudLightning, 644048-BEACON, 644179-EN-

TICE, 643963-SWITCH, 644429-MUSA.

REFERENCES

Ahmed, E., Gani, A., Sookhak, M., Hamid, S. H. A., and

Xia, F. (2015). Application optimization in mobile

cloud computing: Motivation, taxonomies, and open

challenges. Journal of Network and Computer Appli-

cations, 52:52 – 68.

Ali, M., Khan, S. U., and Vasilakos, A. V. (2015). Security

in cloud computing: Opportunities and challenges. In-

formation Sciences, 305:357 – 383.

Assuno, M. D., Calheiros, R. N., Bianchi, S., Netto, M. A.,

and Buyya, R. (2015). Big data computing and clouds:

Trends and future directions. Journal of Parallel and

Distributed Computing, 7980:3 – 15.

Botta, A., de Donato, W., Persico, V., and Pescap, A.

(2016). Integration of cloud computing and internet of

things: A survey. Future Generation Computer Sys-

tems, 56:684 – 700.

European Commission (2015). Topic ict-06: Cloud

computing. https://ec.europa.eu/research/ par-

ticipants/portal/desktop/en/opportunities/h2020/

topics/5091-ict-06-2016.html.

InfraCluster (2015). Novel approaches and tech-

nologies for resource and service management.

https://eucloudclusters.wordpress.com/new-

approaches-for-infrastructure-services.

James, A. and Chung, J.-Y. (2015). Business and industry

speciﬁc cloud: Challenges and opportunities. Future

Generation Computer Systems, 48:39 – 45.

Jeferry, K., Kousiouris, G., Kyriazis, D., Altmann, J., Ciuf-

foletti, A., Maglogiannis, I., Nesi, P., Suzic, B., and

Zhao, Z. (2015). Challenges emerging from future

cloud application scenarios. Procedia Computer Sci-

ence, 68:227 – 237.

Khan, A., Freitag, F., and Rodrigues, L. (2015). Cur-

rent trends and future directions in community edge

clouds. In 2015 IEEE 4th International Conference

on Cloud Networking (CloudNet), pages 239–241.

Mastelic, T. and Brandic, I. (2015). Recent trends in energy-

efﬁcient cloud computing. Cloud Computing, IEEE,

2(1):40–47.

Moura, J. and Hutchison, D. (2016). Review and analysis of

networking challenges in cloud computing. Journal of

Network and Computer Applications, 60:113 – 129.

Nabi, M., Toeroe, M., and Khendek, F. (2016). Availability

in the cloud: State of the art. Journal of Network and

Computer Applications, 60:54 – 67.

Rehman, Z., Hussain, O. K., and Hussain, F. K. (2015).

User-side cloud service management: State-of-the-art

and future directions. Journal of Network and Com-

puter Applications, 55:108 – 122.

Tari, Z., Yi, X., Premarathne, U., Bertok, P., and Khalil, I.

(2015). Security and privacy in cloud computing: Vi-

sion, trends, and challenges. Cloud Computing, IEEE,

2(2):30–38.

Yousif, M., Edsall, T., Krebbers, J., Pappe, S., and Khalidi,

Y. A. (2014). Cloud computing roundtable. Cloud

Computing, IEEE, 1(1):40–49.

CLOSER 2016 - 6th International Conference on Cloud Computing and Services Science

326