A MODEL DRIVEN APPROACH SUPPORTING MULTI-VIEW

SERVICES MODELING AND VARIABILITY MANAGEMENT

Boutaina Chakir and Mounia Fredj

ENSIAS, Mohammed V Souissi University, Rabat, Morocco

Keywords: MDD, MDA, SOA, SOC, Services modeling, Separation of concerns, Variability.

Abstract: Service Oriented Architecture (SOA) is an architectural paradigm for defining how people, organizations

and systems provide and use services to achieve their business goals. Moreover, the growing of information

systems increases the need of agility which implies the ability of a system to be adaptable to the changes in

requirements and context of use. Managing variabily is considered as new leading edge concept for

improving interoperability and reuse. Indeed, variability refers to the characteristic of a system to adapt,

specialize and configure itself with the context of use. Several proposals have been proposed in this sense,

but they are still immature and incomplete. Consequently, in this paper we propose a model driven method

for managing variability in SOA based on MDA (Model Driven Architecture). In fact, through MD, the

method enables the automation of service’s realization regardless of supporting platforms. Our

representation of variability is based on the extension of SOAML which is the future standard for modeling

services. In addition, we adopt the separation of concerns theory by integrating modeling views, to better

organize the various modeling artifacts.

1 INTRODUCTION

Service Oriented Architecture (SOA) is considered

as another concept of maximizing agility and

interoperability between different organizations. It is

based on the description of services and their

interactions.

Moreover, while technology and standards are

important to support SOA, they are not sufficient on

their own. Indeed, with the growing of SOA and the

multiplicity of implementation platforms, the

emphasis is more and more on the services modeling

techniques. In fact, service’s modeling aims to

define the formalisms and notations needed to

describe SOA solutions regardless of the

technologies and standards. Numerous preliminary

methodologies for service’s modeling have been

proposed. However, they are still immature and

many issues haven’t been deeply addressed in these

approaches. In fact, with the continuing evolution of

information systems, the demanding customer

requirements increase the cost involved in designing

and generating variants of a service in an ad-hoc

manner. Hence, the need of mechanisms that support

changes and encourage reuse in SOA systems is

unavoidable. Techniques supporting reuse and

adaptability rely on identifying and managing

service’s variability. Indeed, the aim of service’s

variability is to provide one central technique for

better supporting the reusability of services in

different application scenarios and to simplify the

service consumption by consumers (Narendra,

Ponnalagu, Srivastava. And Banavar, 2008).

In this paper, we present our model driven

service modeling method. This method allows the

management of variability, by dividing modeling

process into two sub process. The first one is for

commonalties and the second is for the variability

aspects of a system. The proposed modeling method

is based on the model driven architecture (MDA),

witch enables the automation and the formalization

of the method. To model variability, we propose to

extend the SOAML profile (SOAML, 2009) by

variability’s stereotypes. Finally, in order to

facilitate and organize our modeling artifacts, we

divide the PIM layer of our method into multiple

modeling views for better separation of concerns.

The remainder of the paper is organized as

follows: Section 2 provides an overview of related

works. In Section 3, we present the modeling

approach and the different views integrated into our

method. A case study is presented in Section 4 to

illustrate our approach. The conclusion is reported in

Section 5.

247

Chakir B. and Fredj M..

A MODEL DRIVEN APPROACH SUPPORTING MULTI-VIEW SERVICES MODELING AND VARIABILITY MANAGEMENT .

DOI: 10.5220/0003495502470252

In Proceedings of the 13th International Conference on Enterprise Information Systems (ICEIS-2011), pages 247-252

ISBN: 978-989-8425-55-3

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

2 RELATED WORKS

In Service Oriented Computing (SOC) (Papazoglou,

2007), services modeling and designing approaches

are an active area of research that aims to specify

systems in a high level of abstraction regardless of

implementation technologies. Among the studied

approaches, we distinguish between those that use

existing development processes (like XP and RUP)

and those that propose their own processes. In the

first category, we include the SOUP method (Mittal,

2006), which proposes two slightly different

variations: one adopting RUP for initial SOA

projects and the other adopting a mix of RUP and

XP for the maintenance of existing SOA projects. In

the second category, we mention the following

methods: Thomas Erl’s (ERL, 2005), IBM Service-

Oriented Analysis and Design (SOAD)

(Zimmermann, 2004), IBM Service Oriented

Modeling and Architecture (SOMA) (Arsanjani,

2004), which act at provider side and support the

services analysis and design.

To provide agility and reuse in response to the

requirements and context changes in SOA solution,

the variability management in SOA is an important

issue. Some works are beginning to address this

aspect under two perspectives: i) the variability for

reuse, where we try to make explicit the variability

in design artifacts, to reuse the service at provider

level, ii) variability for adaptation, which focuses on

the resolution mechanisms of variability according

to the context of use. In this paper we were

interested by the representation of variability in the

purpose of improvement of services reuse. Among

the works reviewed, we cite Robak and

Franczyk

(2003), which introduce the concept of modeling the

variability of Web services using feature diagrams.

Also, the work of (Segura, Benavides, Ruiz-Cort´es

and Trinidad, 2008) focuses on the classification of

variation points for Web Service Flows as a starting

point for handling variability through services.

Besides, Chang and Kim (2007) give a general

classification of variation point in SOA solution:

Workflow, Composition, Interface and Logic

according to SOAD architecture’s layers. Narendra

and Ponnalagu (2007) propose an approach called

Variation-Oriented Requirements Analysis based on

a traceability model. All these approaches are

focused on the representation of variability at just

Figure 1: The proposed modeling method.

ICEIS 2011 - 13th International Conference on Enterprise Information Systems

248

one level: requirement or analysis and not at all

stages of services development lifecycle. To address

this issue, Narendra et al. (2008) propose an end-to-

end approach for variability modeling called (VOE),

which consists of three steps: Variation-Oriented

Analysis (VOA), which is concerned with analyzing

the solution with respect to its static and changing

parts; Variation-Oriented Design (VOD), where a

variation model for the solution design is

instantiated based on the results of VOA; and

Variation-Oriented Implementation (VOI), which is

responsible for producing an implementation based

on the results of VOD. But this approach is

orthogonal to service design, and also it is still not

detailed and immature.

By reviewing these works, we found the absence

of a comprehensive approach for modeling services,

supporting the variability in different stages of

lifecycle development process. Also, with the

increasing number of platforms supporting SOA, it

is interesting to use MDA in the development

approaches. This justifies our approach, that

consists of proposing a services modeling method,

managing variability and following the MDA

approach.

3 TOWARDS A MULTI VIEW

MODEL DRIVEN METHOD

FOR SERVICES MODELING

3.1 Overview of the Method

The main steps of the method are the following (see

figure 1):

• Requirements specification: it consists of

the identification of requirements sources,

the elicitation and analysis of requirements.

Also, it allows the identification of business

process. Finally, it allows the development

of the features model that contains common

features and variable ones. The fixed

elements trigger the base process of the

method and the variable elements are

the

subject of the variability process;

• Base process: the main activities are:

services analysis and design, services

realization and services implementation;

• Variability process: the main activities of

the process are: variability analysis and

design, variability realization and

variability implementation.

3.2 Extension of SOAML for

Variability Modelling

For services modeling, we adopt the SOAML

language, because it is the first standard language

proposed by the OMG (OMG, 2011)for modeling

SOA solutions. However, in order to model

variability, it is important to provide mechanisms

that permit the representation of variation point and

variant which are important for the variability

identification (Jacobson, Griss and Jonsson, 1997).

Hence, we propose to enhance the SOAML profile

by stereotypes that represent variability in SOAML

models. For this purpose we use the UML package

merge relationship to create our profile called

VarSOAML.

So, our profile is a package merging SOAML

(SOAML, 2009) with extensions defined in the

variability package (see Figure 2).

<<merge>>

SOAML:L0

VarSOAML VariabilityModel

Figure 2: Variability profile.

The VariabilityModel package is composed of the

following stereotypes (see Figure 3):

• variableOperation: expresses a variant for

the variation point representing a service

operation;

• VariableMessage: used at the SOAML

messageType level;

• ComplexVariableType: used at the

DataType level, that contains attributes;

• SimpleVariableType: applied at the

DataType level that does not contains

attributes.

The application of these stereotypes is explained

in (Chakir and Fredj, 2011).

Moreover, nowadays, design approaches

increasingly use the software engineering theory

known as the separation of concerns. This theory

asserts that a major problem is best solved when

decomposed into a series of small problems or

concerns. Thus, to facilitate the modeling of

services, we adopt the separation of concerns in our

method. In this perspective, we divide our PIM

A MODEL DRIVEN APPROACH SUPPORTING MULTI-VIEW SERVICES MODELING AND VARIABILITY

MANAGEMENT

249

Figure 3: Variability stereotypes.

Figure 4: Multi views modeling integration.

(Platform Independent Model) layer into several

modeling views.

3.3 Integration of Multi View Modeling

in Our Approach

In order to adopt the separation of the concerns in

our method, we divide our model PIM layer into two

sub layers, which are in turn divided into several

modeling views (see Figure 4):

• Business architecture layer: it captures the

business requirements as well as it allows

the identification of services from business

processes. The models produced in this

layer can be divided into two modeling

views:

o Business view: contains the various

business processes models of the

system, and the entities diagrams that

capture the semantics of the solution;

o Service View: allows the identification

of services. It contains the architecture,

contract diagrams and capability

diagrams.

• System architecture layer: develops

business architecture layer models. It is

composed of the following views:

o Functional view: contains services interface

diagrams and messages exchanged between

services;

o Interaction view: allows the representation

of services compositions. among the

diagrams used in this view, we mention the

service choreography diagram that models

the interactions between services;

o No functional view: represents the non-

functional properties applied to services.

4 CASE STUDY

To validate our approach, we use the application

proposed by (WS-I, 2007), describing "Supply

Chain Management System” (SCMS). SCMS

denotes the process of planning, implementing, and

controlling the operations of the supply chain with

the purpose to satisfy efficiently the customer

requirements. In the requirements specification

stage, we identify the following participants: Client,

Retailer, Warehouse, Shipper and Manufacturer. To

illustrate the functional variability of the SCSM

services, we consider that in the case of a foreigner

client, some retailers may offer the possibility of

checking whether the customer is eligible or not. By

ICEIS 2011 - 13th International Conference on Enterprise Information Systems

250

Shopping Serv ic es Archite cture

Ship Statut:Ship Statut

ship se rv ic e :S hip se rv ice

m a nif a c t urin g s e rv ic e :m a nuf a ct u ring se rv ice

good requesting:Requesting goods service

purchasing service:Purchasing service

ship Info

enquire

supplyer

from

requester

agent

buyer seller

requester

r:Retailer:

Clie nt :

r4: Manufacturer:

r3:Shipper:

r2:W arehouse:

Figure 5: Service view-architecture diagram of SCMS.

Figure 6: Functional view-interface diagram of SCMS.

applying our method, the analysis of the SCMS

system produce a set of models related to each

modeling view:

• Business view: in this view, we represent

all the business processes of the system by

the UML activity diagram. These processes

are used in the identification of services;

• Service view: the Figure 5 illustrates a

simplified architecture diagram of our

system. The SCMS system architecture

binds the roles of participants: Client,

Retailer, Shipper, Warehouse and

Manufacturer. Also, the participants

participate to the following services:

“purchasing service”, “ship service”, “ship

status”, “requesting goods service” and

“manufacturing service”. To specify each

service we use the contract diagrams. For

example the service “Purchasing service”

links two roles “buyer” who is the

consumer and “seller” who is the supplier;

• Functional view: in this view, we construct

the interface diagrams and messages

diagrams of the system, but because of the

pages limitation, we just present a sample

of the interface diagram in Figure 6. The

SOAML stereotype «ServiceInterface»

defines the interface and the responsibilities

of a participant to produce and consume a

service. The interface with a name

beginning with ~ indicate the conjugate

interface (consumer level), using the

operations offered by the interface of the

service producer;

• Interaction view: for a better explanation of

the interactions between the various

A MODEL DRIVEN APPROACH SUPPORTING MULTI-VIEW SERVICES MODELING AND VARIABILITY

MANAGEMENT

251

participants of our system, it is interesting

to enrich the model with the interaction

diagrams which is represented by the

service choreography diagram.

5 CONCLUSIONS

Nowadays, services modeling approaches become

very important. In order to improve services

modeling, it is interesting to use MDA, which brings

automation ability, increasing reuse and productivity

and reducing cost. Among the important issues that

should be considered in modeling approach, we

mention the variability management, which permits

service reuse and service adaptability. In this regard,

we have presented in this paper a model driven

services modeling method that allows the variability

management. The design artifacts of the method are

organized into multiple modeling views for better

separation of concerns.

Since our work is still ongoing, future work

would involve detailing non functional view,

presenting method’s iterations, detailing the process

of variability resolution and defining transformation

rules in order to allow automatic generation of

services. We will then be using this to provide tool

support for our approach.

REFERENCES

Arsanjani, A. (2004). Service-oriented modeling and

architecture. IBM Corporation, available in

http://www-128.ibm.com/developerworks/webservices

/library/ws-soa-design1/.

Chakir, B. and Fredj M. (2011). Towards a modelling

method for managing variability in SOA. Proceedings

of the IADIS International Conference Information

Systems 2011, Avila, Spain.

Chang, S. H. and Kim, S. D. (2007). A variability

modeling method for adaptable services in

service-oriented computing. Software Product Line

Conference, 261-268.

Jacobson, I. Griss, M. Jonsson, P. (1997). Software

reuse: architecture process and organization for

business Success. ACM Press, New York.

Erl, T. (2005), Service-Oriented Architecture: Concepts,

Technology, and Design, Prentice Hall PTR.

Mittal, K. (2006). Service Oriented Unified Process

(SOUP), available from http://www.Kunalmittal.com/

html/soup.shtml, 2006.

Narendra, N. Ponnalagu, K. (2007). Variation-Oriented

Requirements Analysis (VORA). IEEE Congress on

Services, 159-166.

Narendra, N. Ponnalagu, K. Srivastava., B. and Banavar,

G. S. (2008). Variation-oriented engineering (VOE):

enhancing reusability of SOA based solutions.

SCC’08, Vol. 1. IEEE, 257–264.

OMG. (2011). Object Management Group. Available in:

http://www.omg.org/.

Papazoglou, M. Traverso, P. Dustdar, S. Leymann, F.

(2007). Service-Oriented Computing: State of the Art

and Research Challenge. IEEE Computer Society, Vol.

40, No. 11, 38-45, ISSN 0018-9162.

Robak, S. Franczyk, B. (2003). Modeling Web services

variability with feature diagrams. In: Revised Papers

from the NODe 2002 Web and Database-

RelatedWorkshops on Web, Web-Services, and

Database Systems, Springer,Verlag, pp. 120-128.

Segura, S. Benavides, D. Ruiz-Cort´es, A. and Trinidad, P.

(2008). A taxonomy of variability in Web service

flows. In first workshop on Service-Oriented

Architecture and Software Product Lines.

SOAML, (2009). Service oriented architecture Modeling

Language (SoaML) - Specification for the UML

Profile and Metamodel for Services (UPMS). FTF

Beta 2, Available in: http://www.omg.org/docs/ad/08-

08-04.pdf.

WS-I, (2007). Web Services Interoperability Organization

(WS-I) Web site, Available in: http://www.ws-i.org/.

Zhang, L. J. Arsanjani, A. Allam, A. Lu, D. Chee, Y. M.

(2007). Variation oriented analysis for SOA solution

design. Proceedings of the 2007 Salt Lake City, 560-

568.

Zimmermann, O. Krogdahl, P. and Gee, C. (2004).

Elements of Service-Oriented Analysis and Design.

IBM developerWorks.

ICEIS 2011 - 13th International Conference on Enterprise Information Systems

252