Towards Cloud Service Customization: A State of the Art

Cristian Paez, Oscar Avila and Dario Correal

Department of Systems and Computing Engineering, School of Engineering, Universidad de Los Andes, Bogota, Colombia

Keywords:

Cloud Computing, Cloud Customization, Personalized Services, Conﬁgurable Services, SaaS, PaaS, IaaS.

Abstract:

Cloud computing is becoming an important tool that is offering multiple opportunities to organizations. Ho-

wever, cloud providers can not always offer personalized services that fully match customers’ needs. This

occurs for several reasons, the most important is the level of investment required to design and operate perso-

nalized services considering the great number of customers and the diversity of their needs. In this context,

this work presents a literature review to identify what are the main contributions made in the research literature

to carry out such assessment. This review is made by deﬁning three context and ﬁve customization research

questions, making an analysis of each one of them in a ﬁltered set of customization papers and ﬁnding families

or categories that can answer the research questions proposed. The objective is to provide an overview of this

research area and help researchers to develop new contributions for cloud service customization.

1 INTRODUCTION

Cloud computing is becoming an important tool capa-

ble of providing multiple ﬁnancial, social and techno-

logical opportunities to institutions in all industrial

sectors (Armbrust et al., 2010; Zhang et al., 2010).

The possibility of taking such advantage of this mo-

del depends directly on the customization capabilities

that cloud services provide to clients. Such capabi-

lities are essential in order to adapt the service cha-

racteristics to customer needs. Customization needs

can range from changes on the graphical user inter-

face (GUI) to workﬂow and business process logic

modiﬁcation, creating a large range of options when

we refer to cloud customization. However, providers

can not always offer personalized services that fully

match the needs of their clients. This occurs for se-

veral reasons, among which are: (i) the investment

level required to design personalized services taking

into account the number of clients and the diversity

of their needs and (ii) the investment level required to

manage personalized services during the operation ta-

king into account the differences in its design and the

differences in the request for change raised by custo-

mers (Khan and Jiang, 2017).

In this context, this research presents a literature

review to identify and analyze research contributions

to cloud service customization. The main objective of

this review is to provide an updated overview of the

main existing works in the area by identifying their

main characteristics, like customization aspects, met-

hods, techniques, and so on. Other objectives of the

review are: on the one hand, this review will ease re-

searchers search for relevant studies in the area and

help them to constitute a base for further research. On

the other hand, structuring the literature in a detailed

and systematic manner also clariﬁes which issues are

not well covered.

This work is organized as follows: Section two

describes the review method and the applications of

its two ﬁrst steps (planning, and conducting and ma-

terial collection). Section three presents the applica-

tion of the last step of the method, which consists in

a synthesis for each of the evaluation categories used.

Finally, section four presents the conclusions and a

discussion of the results, future work and limitations

of the study.

2 STUDY METHOD AND

DESCRIPTION OF SELECTED

WORKS

The literature review process includes the following

steps: (i) planning, (ii) conducting and material col-

lection and (iii) reporting (Bosse et al., 2014; Seuring

and M

uller, 2008). (i) Planning. It consists on the

deﬁnition of the criteria to conduct the search and va-

lidate the selected works. It aims at identifying the

relevant contributions in the research ﬁeld. (ii) Con-

ducting and Material Collection. This step is related

Paez C., Avila O. and Correal D.

Towards Cloud Service Customization: A State of the Art.

DOI: 10.5220/0006916903710378

In Proceedings of the 15th International Joint Conference on e-Business and Telecommunications (ICETE 2018), pages 205-212

ISBN: 978-989-758-319-3

205

to perform an exhaustive search for primary approa-

ches and validates and assesses the found approaches

with respect to the research questions. The goal of this

step is to select a ﬁnal set of works. (iii) Reporting

phase. This step presents a synthesis of the answers

for each research question, besides a material analysis

and interpretation is also presented.

2.1 Planning

The review is aimed at ﬁnding out the main techni-

ques and methods used for cloud customization. To

perform this analysis, two evaluation categories are

deﬁned: (i) Context to analyze the domain in which

the reviewed works were originally intended to, as

well as their goal and scope and (ii) customizing met-

hod to identify the characteristics of their customiza-

tion process, as well as the customizable aspects, met-

hods for customizing, models for customizing, type

of architecture and customization strategy. Every ca-

tegory includes evaluation criteria that in turn have

related questions to analyze the contribution of each

work to the criteria. The elements previously descri-

bed are shown in Table 1.

2.2 Conducting and Material Collection

With the purpose of ﬁnding potential research works

answering the research questions preciously announ-

ced, the Scopus database was used by introducing the

following criteria:

• Cloud Services Search Terms: (IT Service) or

(ICT service) or (Cloud Service) (Cloud Compu-

ting) or (Cloud Management) or (Cloud Sourcing)

or (Software as a Service) or (Platform as a Ser-

vice) or (Infrastructure as a Service).

• User Customization Terms: Customizing or Per-

sonalize or Customized or Adapted or Adjusted

or Custom-build or Custom-made or Tailor-made

or Made-to-order or Tailored or Conﬁgurable or

Conﬁguration

• Search Area: Computer Science.

• Document type: Conference paper or Journal pa-

per.

• Search ﬁeld type: Title.

Thus, the query introduced in the database is:

TITLE ( ( ”information and comunication technology service”

OR ”information technology service” OR ”IT service” OR ”ICT

service” OR ”cloud service” OR ”saas” OR ”paas” OR ”iaas” OR

”cloud computing” OR ”cloud management” OR ”cloud sourcing”

OR ”software as a service” OR ”platform as a service” OR ”infra-

structure as a service” OR ”software-as-a-service” OR ”platform-

as-a-service” OR ”infrastructure-as-a-service” ) AND ( ”multi te-

nant” OR multitenant OR multi-tenant OR customizing OR cus-

tomising OR personalize OR personalise OR customized OR per-

sonalised OR adapted OR adjusted OR custom-build OR custom-

made OR tailor-made OR made-to-order OR ”made to order” OR

tailored OR conﬁgurable OR conﬁguration ) )

With this criteria, the Scopus search engine retur-

ned 279 candidate articles. To reduce the number of

articles included in the analysis, ﬁrstly, a review of

the article titles was carried out. This ﬁlter reduced

the number to 89. Secondly, a reading of articles ab-

stracts was undertaken to ﬁlter those works that do

not present evidence of answering any of the analysis

questions. This ﬁlter limited the number of articles

to 31. Thirdly, a complete reading of the articles was

performed to select the ﬁnal works set, made up of

25 articles which were identiﬁed and included in the

analysis. The article set includes relevant information

about cloud customization and how different techni-

ques were used to achieve it.

3 REPORTING

After the 25 papers that were more closely related

to cloud customization were selected, we realized an

analysis of each and every one of them under the re-

search questions presented in Table 1. After an indi-

vidual analysis, we performed a group analysis of all

papers under each research question in order to ﬁnd

relationships between the papers that could lead into

conclusions for each research question.

3.1 Context

3.1.1 To Which Domain is the Research Work

Applied?

During our analysis of the knowledge in cloud cus-

tomization, we divided the papers into domain cate-

gories to analyze the context of the work done in this

ﬁeld. Of the 25 papers analyzed, we found that 21 of

them were oriented to the academia, while the remai-

ning 4 were oriented to the development of the indu-

stry. References to these papers can be found in table

2, under the ”Papers Domain Classiﬁcation” section.

3.1.2 What is the Main Objective of the

Research Work?

After the analysis of Cloud Customization related

papers, we found three categories of objectives that

ICETE 2018 - 15th International Joint Conference on e-Business and Telecommunications

206

Table 1: Categories, criteria and research questions.

Category Criteria Question

Context

Domain To which domain is the research work applied?

Goal What is the main objective of the research work?

Scope For what service (SaaS, PaaS, IaaS) and deployment (private, public, community,

hybrid) models is the research work intended?

Customizing

Method

Customizable Aspects Which aspects of the service can be customized? How is the customization ad-

dressed ?

Method for customi-

zing

Which are the methods employed to customize the service?

Models for Customi-

zing

Which are the models, languages or notations employed to customize the ser-

vice?

Type of Architecture What Architecture style was used?

Customization Stra-

tegy

In which degree the customization is made in an automatic, semiautomatic or

manual way?

could cover the core goals of all of them. First, we

identiﬁed the “Architectures based on Customization”

category. In this category we can found papers whose

goal is to propose or deﬁne architectures with custo-

mization in mind, using different techniques like (i)

the development of new architectures that have cu-

stomizable modules that can be selected by the te-

nant (Tizzei et al., 2017; Kriouile et al., 2015; Wang

and Zheng, 2010; Jumagaliyev et al., 2016; Walra-

ven et al., 2014), (ii) the development of a framework

to create applications with run time customizations of

user interfaces and business logic using inheritance

and polymorphism (Lee and Choi, 2012) or archi-

tectures built with conﬁgurable options (Zhou et al.,

2016; Cao et al., 2015; Kumara et al., 2013; Kha-

lil et al., 2013; Tsai and Sun, 2013; Schroeter et al.,

2012a; Kang et al., 2011; Mietzner et al., 2008; Khan

et al., 2015; Correia et al., 2013; Jamshidi and Pahl,

2015; Schroeter et al., 2012b)

Second, we found the “Optimization of Customi-

zation Features” category, which covers papers that

address the performance of conﬁguration processes

and implement techniques to increase their efﬁciency.

In this group we can ﬁnd papers that address the area

of tension between cost efﬁciency and customizabi-

lity (M

uller et al., 2009) or that develop a framework

to identify the proper amount of customization nee-

ded by a cloud application (Walraven et al., 2015).

Finally, in the context of cloud application with al-

ready developed features that can be implemented to a

base application by the tenant, we identiﬁed the ”Con-

nection of Features” category. This category contains

papers that help the connection and conﬁguration of

these features with the core application. In this ca-

tegory we found papers that address this issue using

primarily two techniques:

• Develop matching techniques that ﬁnd the featu-

res selected by the tenant, merge their core appli-

cation with the services selected and help to cus-

tomize them (Hajlaoui et al., 2017).

• Develop a platform that allow tenants to switch

conﬁgurations of the features selected in a quick

and easy way (Bob

ak et al., 2015).

In our analysis, we could ﬁnd that 18 papers fell in

the “Architectures based on Customization” category,

of which 5 of them where papers that implemented

their architectures in the industry (Tizzei et al., 2017;

Cao et al., 2015; Kumara et al., 2013; Tsai and Sun,

2013; Walraven et al., 2014) and the rest 13 of them

in this category made their proposals in an academic

environment. We also found that 3 papers fell into

the “Optimization of Customization Features” cate-

gory and the remaining 4 papers where assigned to

the “Connection and Conﬁguration of Features” cate-

gory. References to papers from all categories of this

question can be found in table 2, under the ”Papers

goals classiﬁcation” section.

3.1.3 For What Service and Deployment Models

is the Research Work Intended?

Our research found that, from our ﬁnal set of 25 pa-

pers related to cloud customization, we found that

only one of them was oriented to Iaas and PaaS ap-

plications (Bob

ak et al., 2015) and another one was

oriented only to Iaas platforms (Zhou et al., 2016).

The rest of the papers were all focused on SaaS ap-

plications, while one of them was a SaaS variation

based on customized instances for each tenant (Khan

and Jiang, 2017). References to all papers focused on

SaaS applications can be found in table 2, under the

”Papers Scope Classiﬁcation” section.

Since the majority of work found was dedicated to

the academia and their proposals were usually orien-

ted towards SaaS, PaaS or IaaS in general, not all the

papers had an speciﬁc deployment model (private, pu-

blic, community or hybrid). We found that only 5 ar-

ticles that were focused in SaaS services were orien-

ted to a speciﬁc deployment model, of which 4 were

oriented to private SaaS (Tizzei et al., 2017; Cao et al.,

Towards Cloud Service Customization: A State of the Art

207

2015; Kriouile et al., 2015; Kumara et al., 2013) and

1 was oriented to public SaaS (Walraven et al., 2014).

3.2 Customizing Method

3.2.1 Which Aspects of the Service can be

Customized? How is the Customization

Addressed?

First, we deﬁned a group called “Selectable Modu-

les” category. This group has all the papers that try to

achieve customization by creating different, indepen-

dent and functional modules that can be integrated to

the main application according to user’s needs. This

process is usually made in a manual manner, by de-

veloping new functional modules that can be added

to the main application as the client requests them,

usually using modules made in the past for efﬁciency

(Tizzei et al., 2017; Schroeter et al., 2012a; Miet-

zner et al., 2008; M

uller et al., 2009; Jumagaliyev

et al., 2016; Correia et al., 2013; Walraven et al.,

2014; Schroeter et al., 2012b), but other papers use

an approach that allows the tenant to pick from a pre-

deﬁned list of already developed modules to use in

his application, making the assembling of modules in

real time (Liu et al., 2012; Wang and Zheng, 2010;

Jamshidi and Pahl, 2015). We found that the Selec-

table Modules category is the most common, with 11

papers related to it. References to all papers in this

category can be found in table 2, under the ”Papers

Customizing Aspects Classiﬁcation” section.

Second, we identiﬁed the ”Customization of Fi-

les” category. In this category we found techniques to

customize cloud applications by the modiﬁcation of

style ﬁles that deﬁne the aspect of the application, al-

lowing user to personalize logos and colors, and spe-

ciﬁc functionality methods, which allows to persona-

lize some speciﬁc functionalities. We found 2 papers

associated with this category in our study, references

to all papers in this category can be found in table

2, under the ”Papers Customizing Aspects Classiﬁca-

tion” section.

Third, we deﬁned the ”System Structure” cate-

gory, in which the articles that allow to personalize the

workﬂow and the structure of the application through

an interface are included. Such interface allows user

to select their preferences and reconﬁgure their sin-

gle instance according to their needs. In this category

we classiﬁed 2 papers which implement these techni-

ques, references to all papers in this category can be

found in table 2, under the ”Papers Customizing As-

pects Classiﬁcation” section.

Finally, we deﬁned the ”Feature Selection Aid”

category. In this category, cloud services help users

choose the features they need to get their desired ser-

vice, mapping those requirements with the services

needed. Some of these papers use matching algo-

rithms that search the features or hardware preferen-

ces and conﬁgure them to work with the tenant instan-

ced application without much more input for the user

(Hajlaoui et al., 2017; Zhou et al., 2016; Cao et al.,

2015), while others focus on the assistance of user in

the moment of selecting their desired service (Bob

et al., 2015; Kriouile et al., 2015; Kumara et al., 2013;

Tsai and Sun, 2013; Walraven et al., 2015; Sun et al.,

2008). This category has 8 papers that talk about

different ways to build the application that the user

needs, references to all papers in this category can be

found in table 2, under the ”Papers Customizing As-

pects Classiﬁcation” section.

3.2.2 Which are the Methods Employed to

Customize the Service?

We manage to organize the different methods of cu-

stomization in three main categories. The ﬁrst group

was called “Software Product Line (SPL) related met-

hods and their variations”, in which all the methods

related to SPL or that use SPL techniques are grou-

ped. These techniques include the domain analysis of

the product, a development of assets that can be con-

nected to a base application and a set of rules for the

development and assembling of assets for new conﬁ-

gurations.

The second group we distinguished was the “De-

velopment of new Architecture”, in which the met-

hods that develop a new architecture with customi-

zation in mind can be found. This process usually

deﬁnes all the major components of the architecture

proposed depending on the focus of each architecture,

deﬁning their architectural components and general

workﬂow. This category is characterized by having

techniques that do not relly on independent modules,

since it is for the papers that focus on new architectu-

ral structures self made by the authors. These techni-

ques vary, but they usually deﬁne aspects like instan-

ces, data organization and user management for cus-

tomization.

Third, we deﬁned the “Function customization”

group, which locates the methods that modiﬁcate or

help the modiﬁcation of functions to customize user

experience. These techniques focus on the customi-

zation of short pieces of code (usually functions) that

allow them to customize speciﬁc functionalities for

the tenants.

Finally, we deﬁned the “Implementation of Algo-

rithms” category, that locates all the methods that run

algorithms that automate or aid users with their cus-

tomization process. The functionality of these algo-

ICETE 2018 - 15th International Joint Conference on e-Business and Telecommunications

208

rithms vary between papers, but they usually help the

user with the conﬁguration of desired features (Zhou

et al., 2016), connection with speciﬁc functionalities

(Hajlaoui et al., 2017; Kriouile et al., 2015) or auto-

matization (Tsai and Sun, 2013).

From the 25 papers analyzed, we found that the

“Software Product Line (SPL) related methods and

their variations” category had 10 papers related to it,

the “Development of new Architecture” category had

6, the “Function customization” category had 5 and

the “Implementation of Algorithms” category had 4.

References to these papers can be found in table 2, un-

der the “Papers Customizing Method Classiﬁcation”

section.

3.2.3 Which are the Models, Languages or

Notations Employed to Customize the

Service?

In our analysis we found that 9 papers used Flow

Charts, 9 papers used Feature Models, 7 papers used

the UML standard, 6 papers used Architectural mo-

dels, 3 papers used tree models, 3 papers used Com-

ponents Models and 2 used new models designed by

the authors, the reference of which can be found in ta-

ble 2 under the “Papers Customizing Models Classiﬁ-

cation” section. Feature Models were usually used to

represent variability in applications with independent

modules or that implemented SPL techniques, Tree

models were used to represent components or inheri-

tance relationships, Component Models were used in

architectures based in components and the new mo-

dels were used to represent characteristics of new ar-

chitectures that were not supported by traditional mo-

dels.

Since most papers used more than one model or

language, we could ﬁnd some relationships in our

analysis. We found that Feature models are used

with almost all the other type of models (Tizzei et al.,

2017; Hajlaoui et al., 2017; Cao et al., 2015; Kumara

et al., 2013; Tsai and Sun, 2013; Jumagaliyev et al.,

2016; Jamshidi and Pahl, 2015; Walraven et al., 2014;

Schroeter et al., 2012b), Tree models are usually used

with feature models (Hajlaoui et al., 2017; Tsai and

Sun, 2013) and Flow Charts usually are used alone

(Zhou et al., 2016; Bob

ak et al., 2015; Kriouile et al.,

2015; Khan et al., 2015) to describe the customization

methods in papers.

This analysis shows that Feature Models and Flow

Charts are the most commonly used to describe Cus-

tomization in cloud services (Tizzei et al., 2017; Ha-

jlaoui et al., 2017; Zhou et al., 2016; Bob

ak et al.,

2015; Cao et al., 2015; Kriouile et al., 2015; Kumara

et al., 2013; Tsai and Sun, 2013; Liu et al., 2012;

Kang et al., 2011; Mietzner et al., 2008; Jumaga-

liyev et al., 2016; Khan et al., 2015; Jamshidi and

Pahl, 2015; Walraven et al., 2014; Schroeter et al.,

2012b). However, since many of the papers focus in

new architectures, the Architectural models and UML

techniques are also used in many papers to describe

their proposals (Tizzei et al., 2017; Khan and Jiang,

2017; Khalil et al., 2013; Lee and Choi, 2012; Schroe-

ter et al., 2012a; Kang et al., 2011; Wang and Zheng,

2010; Jumagaliyev et al., 2016; Walraven et al., 2015;

Correia et al., 2013; Walraven et al., 2014). We can

also see that Tree and Component Models are com-

mon (Hajlaoui et al., 2017; Tsai and Sun, 2013; Lee

and Choi, 2012; Schroeter et al., 2012a; Sun et al.,

2008; Correia et al., 2013), which is not rare taking

into account the amount of papers that focus in modu-

lar architectures. Finally, we found that some papers

developed new models (M

uller et al., 2009; Schroeter

et al., 2012b), showing the necessity to explore new

ways to represent the concepts of cloud features in

some cases.

3.2.4 What Architecture Style was used?

Over the 25 papers analyzed, we identiﬁed four

groups that could represent the different types of

techniques used for architecture. First, we identiﬁed

the Components and Software Product Line related

techniques, which covers all the papers that use ar-

chitectures based on components or techniques simi-

lar to the ones used in SPL engineering, since SPL

techniques heavily rely on connection between com-

ponent models. Second, we deﬁned the Integrated

Architectures category, which covers the papers that

focus on architectures that do not use functional mo-

dules, and instead use different type of techniques like

layers, meta data or services to provide the customi-

zation. Third, we found the Architectures Based on

tree category, that deﬁnes architectures that change

their functions following a tree structure. Finally, we

deﬁned the Techniques for Customization category,

which covers the papers that do not deﬁne an architec-

ture scheme but use different techniques to facilitate

the customization for tenants.

In our analysis, we found that the Components and

Software Product Line related techniques had 11 pa-

pers related to it, the Integrated Architectures cate-

gory had 8 papers, the Architectures Based on Trees

had 3 papers and the Techniques for Customization

category had also 3 papers related to it. References to

these papers can be found in table 2, under the ”Pa-

pers Type of Architecture Classiﬁcation” section.

Towards Cloud Service Customization: A State of the Art

209

Table 2: Analysis of papers related to Cloud Customization.

Category Total Papers

Papers Domain Classiﬁcation

Academia 20 (Khan and Jiang, 2017; Hajlaoui et al., 2017; Zhou et al., 2016; Bob

ak et al., 2015; Kriouile et al.,

2015; Khalil et al., 2013; Lee and Choi, 2012; Schroeter et al., 2012a; Liu et al., 2012; Kang et al.,

2011; Wang and Zheng, 2010; Mietzner et al., 2008; M

uller et al., 2009; Jumagaliyev et al., 2016;

Walraven et al., 2015; Khan et al., 2015; Sun et al., 2008; Correia et al., 2013; Jamshidi and Pahl,

2015; Schroeter et al., 2012b)

Industry 5 (Tizzei et al., 2017; Cao et al., 2015; Tsai and Sun, 2013; Walraven et al., 2014)

Papers Goals Classiﬁcation

Architectures based on Custo-

mization

18 (Tizzei et al., 2017; Zhou et al., 2016; Cao et al., 2015; Kriouile et al., 2015; Kumara et al., 2013;

Khalil et al., 2013; Tsai and Sun, 2013; Lee and Choi, 2012; Schroeter et al., 2012a; Kang et al.,

2011; Wang and Zheng, 2010; Mietzner et al., 2008; Jumagaliyev et al., 2016; Khan et al., 2015;

Correia et al., 2013; Jamshidi and Pahl, 2015; Walraven et al., 2014; Schroeter et al., 2012b)

Optimization of Customiza-

tion Features

3 (M

uller et al., 2009; Walraven et al., 2015; Sun et al., 2008)

Connection and Conﬁguration

of Features

4 (Khan and Jiang, 2017; Hajlaoui et al., 2017; Bob

ak et al., 2015; Liu et al., 2012)

Papers Scope Classiﬁcation

SaaS 24 (Tizzei et al., 2017; Hajlaoui et al., 2017; Cao et al., 2015; Kriouile et al., 2015; Kumara et al., 2013;

Khalil et al., 2013; Tsai and Sun, 2013; Lee and Choi, 2012; Schroeter et al., 2012a; Liu et al., 2012;

Kang et al., 2011; Wang and Zheng, 2010; Mietzner et al., 2008; M

uller et al., 2009; Jumagaliyev

et al., 2016; Walraven et al., 2015; Khan et al., 2015; Sun et al., 2008; Correia et al., 2013; Jamshidi

and Pahl, 2015; Walraven et al., 2014; Schroeter et al., 2012b)

Iaas 2 (Zhou et al., 2016; Bob

ak et al., 2015)

Paas 1 (Bob

ak et al., 2015)

Private 4 (Tizzei et al., 2017; Cao et al., 2015; Kriouile et al., 2015; Kumara et al., 2013)

Public 1 (Walraven et al., 2014)

Papers Customizing Aspects Classiﬁcation

Selectable Modules 11 (Tizzei et al., 2017; Schroeter et al., 2012a; Liu et al., 2012; Wang and Zheng, 2010; Mietzner et al.,

2008; M

uller et al., 2009; Jumagaliyev et al., 2016; Correia et al., 2013; Jamshidi and Pahl, 2015;

Walraven et al., 2014; Schroeter et al., 2012b)

Customization of Files 2 (Khan and Jiang, 2017; Lee and Choi, 2012; Khan et al., 2015)

System Structure 2 (Khalil et al., 2013; Kang et al., 2011)

Feature Selection Aid 8 (Hajlaoui et al., 2017; Zhou et al., 2016; Bob

ak et al., 2015; Cao et al., 2015; Kriouile et al., 2015;

Kumara et al., 2013; Tsai and Sun, 2013; Walraven et al., 2015; Sun et al., 2008)

Papers Customizing Method Classiﬁcation

Software Product Line (SPL)

related methods and their va-

riations

10 (Tizzei et al., 2017; Cao et al., 2015; Kumara et al., 2013; Schroeter et al., 2012a; Mietzner et al.,

2008; Jumagaliyev et al., 2016; Correia et al., 2013; Jamshidi and Pahl, 2015; Walraven et al., 2014;

Schroeter et al., 2012b)

Development of new Archi-

tecture

6 (Bob

ak et al., 2015; Khalil et al., 2013; Lee and Choi, 2012; Kang et al., 2011; Walraven et al., 2015;

Sun et al., 2008)

Function customization 5 (Khan and Jiang, 2017; Liu et al., 2012; Wang and Zheng, 2010; M

uller et al., 2009; Khan et al.,

2015)

Implementation of Algo-

rithms

4 (Hajlaoui et al., 2017; Zhou et al., 2016; Kriouile et al., 2015; Tsai and Sun, 2013)

Papers Customizing Models Classiﬁcation

Flow Charts 9 (Zhou et al., 2016; Bob

ak et al., 2015; Cao et al., 2015; Kriouile et al., 2015; Kumara et al., 2013;

Liu et al., 2012; Kang et al., 2011; Mietzner et al., 2008; Khan et al., 2015)

Feature Models 9 (Tizzei et al., 2017; Hajlaoui et al., 2017; Cao et al., 2015; Kumara et al., 2013; Tsai and Sun, 2013;

Jumagaliyev et al., 2016; Jamshidi and Pahl, 2015; Walraven et al., 2014; Schroeter et al., 2012b)

Component Models 3 (Schroeter et al., 2012a; Sun et al., 2008; Correia et al., 2013)

UML 7 (Tizzei et al., 2017; Khan and Jiang, 2017; Schroeter et al., 2012a; Wang and Zheng, 2010; Jumaga-

liyev et al., 2016; Correia et al., 2013; Walraven et al., 2014)

Architectural Models 6 (Khalil et al., 2013; Lee and Choi, 2012; Schroeter et al., 2012a; Kang et al., 2011; Wang and Zheng,

2010; Walraven et al., 2015)

Trees Models 3 (Hajlaoui et al., 2017; Tsai and Sun, 2013; Lee and Choi, 2012)

New Models 2 (M

uller et al., 2009; Schroeter et al., 2012b)

Papers Type of Architecture Classiﬁcation

Components and Software

Product Line related techni-

ques

11 (Tizzei et al., 2017; Kumara et al., 2013; Khalil et al., 2013; Schroeter et al., 2012a; Mietzner et al.,

2008; Jumagaliyev et al., 2016; Walraven et al., 2015; Correia et al., 2013; Jamshidi and Pahl, 2015;

Walraven et al., 2014; Schroeter et al., 2012b)

Integrated Architectures 8 (Zhou et al., 2016; Bob

ak et al., 2015; Kriouile et al., 2015; Lee and Choi, 2012; Liu et al., 2012;

Kang et al., 2011; Wang and Zheng, 2010; M

uller et al., 2009)

Architectures Based on Trees 3 (Hajlaoui et al., 2017; Cao et al., 2015; Tsai and Sun, 2013)

Techniques for Customization 3 (Khan and Jiang, 2017; Khan et al., 2015; Sun et al., 2008)

Papers Customizing Strategy Classiﬁcation

Manual 17 (Tizzei et al., 2017; Khan and Jiang, 2017; Hajlaoui et al., 2017; Cao et al., 2015; Kriouile et al.,

2015; Kumara et al., 2013; Khalil et al., 2013; Lee and Choi, 2012; Schroeter et al., 2012a; Mietzner

et al., 2008; M

uller et al., 2009; Jumagaliyev et al., 2016; Walraven et al., 2015; Khan et al., 2015;

Correia et al., 2013; Walraven et al., 2014; Schroeter et al., 2012b)

Semiautomatic 5 (Liu et al., 2012; Kang et al., 2011; Wang and Zheng, 2010; Sun et al., 2008; Jamshidi and Pahl,

2015)

Automatic 3 (Zhou et al., 2016; Bob

ak et al., 2015; Tsai and Sun, 2013)

ICETE 2018 - 15th International Joint Conference on e-Business and Telecommunications

210

3.2.5 In Which Degree the Customization is

Made is an Automatic, Semiautomatic or

Manual Way?

In our study, we analyzed 25 papers and found that

17 of them used manual, 5 used Semi-automatic and

3 used Automatic customization methods. When re-

lated with the Methods for Customization category,

the Manual strategy is mostly related to the “Software

Product Line (SPL) related methods and their variati-

ons” category (Tizzei et al., 2017; Cao et al., 2015;

Kumara et al., 2013; Schroeter et al., 2012a; Miet-

zner et al., 2008; Jumagaliyev et al., 2016; Correia

et al., 2013; Walraven et al., 2014; Schroeter et al.,

2012b), the Semiautomatic strategy is mostly related

to the “Development of new Architecture” (Liu et al.,

2012; Wang and Zheng, 2010) and “Function custo-

mization” (Kang et al., 2011; Sun et al., 2008) in the

same amount, and the automatic strategy is mostly re-

lated to the “Implementation of Algorithms” (Zhou

et al., 2016; Tsai and Sun, 2013). References to all

papers can be found in table 2, under the ”Papers Type

of Architecture Classiﬁcation” section.

4 CONCLUSIONS AND FUTURE

WORK

The conclusions regarding the ﬁrst category is descri-

bed as follows. The domain category shows an in-

clination towards the academia that could imply that

many research made in cloud customization has not

been implemented in the industry yet, and there is

ﬁeld for more research about the techniques imple-

mented in the industry. Regarding the goal category,

the Architectures based on Customization category

was the most popular, in which papers proposed new

architectures to this end. This inclination could indi-

cate that customization has to be a core requirement

during the construction of the architecture of a cloud

application. Regarding the scope, we found that a vast

majority of the work analyzed was oriented to SaaS

applications, and only two papers dedicated to IaaS

and PaaS service models. We believe that this incli-

nation is due to the difﬁculty that SaaS customization

represents.

The conclusions regarding the second category is

described as follows. Regarding the customizable as-

pects, we believe that the popularity of the Selecta-

ble Modules category relays on the amount of dif-

ferent services that an application with different se-

lectable modules can offer. Regarding the Methods

for Customizing, we found that the works in each

of the identiﬁed categories show that techniques to

achieve the above described customizable aspects can

vary between authors. However, we saw a clear incli-

nation to create architectures from scratch, by using

SPL techniques or new architectural styles. Regar-

ding the models for Customizing, a clear majority was

focused in the proposal of new architectures with dif-

ferent modelling techniques. Since the majority of

work is done in architectural proposals, models such

as Flow charts are the most used. Also, we found pa-

pers who created new modelling techniques either by

modifying already existing ones or creating their own

ones. Regarding the Type of Architecture, we found

that SPL and SPL derived techniques were popular in

cloud customization, but architectures that do not use

modular components are almost equally used. Finally,

regarding the Customization Strategy, we found that

the majority of the work analyzed made Manual Cu-

stomization. These results indicate that further work

is necessary for amortization. As future work, additi-

onal research is necessary to propose and structure a

maturity model for cloud service customization based

on the literature review carried out in this work. Such

model could provide elements to practitioners allo-

wing them to understand the maturity level in which

their organizations currently are and the path to move

towards higher maturity levels. In addition, further

research work is necessary in order to understand the

beneﬁts that cloud service customization can provide

to customers in several areas such as Business and IT

alignment, which is one of the today top priorities for

IT managers (Avila and Garc

es, 2014).

REFERENCES

Armbrust, M., Fox, A., Grifﬁth, R., Joseph, A., Katz, R.,

Konwinski, A., Lee, G., Patterson, D., Rabkin, A.,

Stoica, I., and M., Z. (2010). A view of cloud compu-

ting. 53(4):50–58.

Avila, O. and Garc

es, K. (2014). Change management

contributions for business-it alignment. In Abramo-

wicz, W. and Kokkinaki, A., editors, Business Infor-

mation Systems Workshops, volume 183, pages 156–

167, Cham. Springer International Publishing.

Bob

ak, M., Hluch

y, L., and Tran, V. (2015). Tailored plat-

forms as cloud service. In 2015 IEEE 13th Internatio-

nal Symposium on Intelligent Systems and Informatics

(SISY), pages 43–48.

Bosse, S., Splieth, M., and Turowski, K. (2014). Model-

based prediction of it service availability - a literature

review. Journal of Internet Services and Applications,

pages 141–148.

Cao, Y., Lung, C. H., and Ajila, S. A. (2015). Constraint-

based multi-tenant saas deployment using feature mo-

deling and xml ﬁltering techniques. In 2015 IEEE

Towards Cloud Service Customization: A State of the Art

211

39th Annual Computer Software and Applications

Conference, volume 3, pages 454–459.

Correia, A., Penha, J., and Da Cruz, A. (2013). An architec-

tural model for customizing the business logic of saas

applications. In ICSOFT 2013 - Proceedings of the 8th

International Joint Conference on Software Technolo-

gies, pages 162–168.

Hajlaoui, J. E., Omri, M. N., and Benslimane, D. (2017).

Multi-tenancy aware conﬁgurable service discovery

approach in cloud computing. In 2017 IEEE 26th In-

ternational Conference on Enabling Technologies: In-

frastructure for Collaborative Enterprises (WETICE),

pages 232–237.

Jamshidi, P. and Pahl, C. (2015). Orthogonal variability

modeling to support multi-cloud application conﬁgu-

ration. In Advances in Service-Oriented and Cloud

Computing, pages 249–261.

Jumagaliyev, A., Whittle, J., and Elkhatib, Y. (2016).

Evolving multi-tenant saas cloud applications using

model-driven engineering. CEUR Workshop Procee-

dings, 1706:60–64.

Kang, S., Kang, S., and Hur, S. (2011). A design of the

conceptual architecture for a multitenant saas applica-

tion platform. In 2011 First ACIS/JNU International

Conference on Computers, Networks, Systems and In-

dustrial Engineering, pages 462–467.

Khalil, E., Enniari, S., and Zbakh, M. (2013). Cloud com-

puting architectures based multi-tenant ids. In 2013

National Security Days (JNS3), pages 1–5.

Khan, K. M. and Jiang, Z. (2017). Managing client-speciﬁc

customised functions in multi-tenant software-as-a-

service. In 2017 IEEE International Conference on

Software Quality, Reliability and Security Companion

(QRS-C), pages 108–113.

Khan, K. M., Nhlabatsi, A., and Khan, N. (2015). A process

model for customisation of software in multi-tenant

saas model. In 2015 IEEE/ACM 8th International

Conference on Utility and Cloud Computing (UCC),

pages 418–419.

Kriouile, H., Mcharﬁ, Z., and Asri, B. E. (2015). To-

wards a high conﬁgurable saas: To deploy and bind

auser-aware tenancy of the saas. In ICEIS 2015 - Pro-

ceedings of the 17th International Conference on En-

terprise Information Systems, Volume 2, Barcelona,

Spain, 27-30 April, 2015, volume 2, page 674–679.

Kumara, I., Han, J., Colman, A., and Kapuruge, M. (2013).

Runtime evolution of service-based multi-tenant saas

applications. In Service-Oriented Computing, pages

192–206.

Lee, W. and Choi, M. (2012). A multi-tenant web applica-

tion framework for saas. In 2012 IEEE Fifth Interna-

tional Conference on Cloud Computing, pages 970–

971.

Liu, G., Zhang, W., Li, D., Zhou, S., and Zhu, Z. (2012).

Bpel based service process conﬁguration in saas ap-

plication. In Advances in Computer, Communication,

Control and Automation, pages 157–164.

Mietzner, R., Leymann, F., and Papazoglou, M. P. (2008).

Deﬁning composite conﬁgurable saas application

packages using sca, variability descriptors and multi-

tenancy patterns. In 2008 Third International Confe-

rence on Internet and Web Applications and Services,

pages 156–161.

uller, J., Kr

uger, J., Enderlein, S., Helmich, M., and

Zeier, A. (2009). Customizing enterprise software as

a service applications: Back-end extension in a multi-

tenancy environment. In Enterprise Information Sys-

tems, pages 66–77.

Schroeter, J., Cech, S., G

otz, S., Wilke, C., and Aßmann,

U. (2012a). Towards modeling a variable architecture

for multi-tenant saas-applications. In Proceedings of

the Sixth International Workshop on Variability Mo-

deling of Software-Intensive Systems, VaMoS ’12, pa-

ges 111–120.

Schroeter, J., Mucha, P., Muth, M., Jugel, K., and Lochau,

M. (2012b). Dynamic conﬁguration management of

cloud-based applications. In Proceedings of the 16th

International Software Product Line Conference - Vo-

lume 2, SPLC ’12, pages 171–178.

Seuring, S. and M

uller, M. (2008). From a literature re-

view to a conceptual framework for sustainable supply

chain management. Journal of Cleaner Production,

16(15):1699 – 1710.

Sun, W., Zhang, X., Guo, C. J., Sun, P., and Su, H. (2008).

Software as a service: Conﬁguration and customiza-

tion perspectives. In 2008 IEEE Congress on Services

Part II (services-2 2008), pages 18–25.

Tizzei, L. P., Nery, M., Segura, V. C. V. B., and Cerqueira,

R. F. G. (2017). Using microservices and software

product line engineering to support reuse of evolving

multi-tenant saas. In Proceedings of the 21st Interna-

tional Systems and Software Product Line Conference

- Volume A, SPLC ’17, pages 205–214. ACM.

Tsai, W. T. and Sun, X. (2013). Saas multi-tenant applica-

tion customization. In 2013 IEEE Seventh Internatio-

nal Symposium on Service-Oriented System Engineer-

ing, pages 1–12.

Walraven, S., Borger, W. D., Vanbrabant, B., Lagaisse, B.,

Landuyt, D. V., and Joosen, W. (2015). Adaptive per-

formance isolation middleware for multi-tenant saas.

In 2015 IEEE/ACM 8th International Conference on

Utility and Cloud Computing (UCC), pages 112–121.

Walraven, S., Van Landuyt, D., Truyen, E., Handekyn,

K., and Joosen, W. (2014). Efﬁcient customiza-

tion of multi-tenant software-as-a-service applications

with service lines. Journal of Systems and Software,

91(1):48–62.

Wang, H. and Zheng, Z. (2010). Software architecture dri-

ven conﬁgurability of multi-tenant saas application. In

Web Information Systems and Mining, pages 418–424.

Zhang, Q., Cheng, L., and Boutaba, R. (2010). Cloud com-

puting: state-of-the-art and research challenges. Jour-

nal of Internet Services and Applications, 1(1):7–18.

Zhou, Y., Hoffmann, H., and Wentzlaff, D. (2016). Cash:

Supporting iaas customers with a sub-core conﬁgu-

rable architecture. In 2016 ACM/IEEE 43rd Annual

International Symposium on Computer Architecture

(ISCA), pages 682–694.

ICETE 2018 - 15th International Joint Conference on e-Business and Telecommunications

212