A MOF-based Social Web Services Description Metamodel

Amel Benna

1,3

, Zakaria Maamar

and Mohamed Ahmed Nacer

DSISM, CERIST, Algiers, Algeria

College of Technological Innovation, Zayed University, Dubai, U.A.E.

LSI Laboratory, USTHB, Algiers, Algeria

Keywords:

MDA, MOF, Social Web Service, Metamodeling.

Abstract:

To promote and support the development and use of social Web services by the IT community on the Web,

both social Web service-based applications and their support platforms should evolve independently from

each other while sharing a common model that represents the characteristics of these social Web services. To

achieve this duality, this paper proposes a model-driven approach. First, the approach identiﬁes a social Web

service’s properties. Then a Meta-Object-Facility (MOF)-based social Web services description metamodel is

developed. Finally, a prototype illustrates how the MOF-based metamodel is used.

1 INTRODUCTION

Web services (WS)s offer a standardized way for de-

ploying interoperableWeb-based applications (Chung

et al., 2003). However, several issues like discovery,

composition, and monitoring continue to undermine

their beneﬁts. To address some of these issues, there

is a research trend that looks into WSs from a social

perspective (Xie et al., 2008; Chen et al., 2015).

A Social Web Service (SWS) is the result of

blending social computing with service-orientedcom-

puting. Compared to regular WSs, SWSs ”establish

and maintain networks of contacts; count on their

(privileged) contacts when needed; form with other

peers strong and long lasting collaborative groups;

and, know with whom to partner so that ontology rec-

onciliation is minimized” (Maamar et al., 2011a). For

instance, prepareJob SWS “likes” to collaborate with

postJob SWS because of previous successful compo-

sitions and also “recommends”, for similar reasons,

planJob SWS as a replacement in the case of failure.

postJob SWS and planJob SWS are part of prepare-

Job SWS’s ”collaboration” and ”substitution” net-

works, respectively.

The existing solutions for discovering and com-

posing WSs from a social perspective (Maamar et al.,

2011b; Chen et al., 2015; Maaradji et al., 2011),

rely on a speciﬁc SWS’s networks (e.g., substitution

network) and speciﬁc platform (e.g., Java EE plat-

form, and .Net). Therefore, the evolution and update

(e.g., new social networks and platform update) of

these existing solutions are difﬁcult. There is a seri-

ous lack of a common model that permits to describe

SWSs’ properties (e.g., who are a SWS’s contacts)

across different stakeholders (e.g., designers, devel-

opers, providers, and users) and regardless of the plat-

form used. Moreover, Web Service Description Lan-

guage (WSDL) does not capture the “social” dimen-

sion of a WS in terms of who are the collaborators

and substitutes, for example. What if we think of S-

WSDL for Social-WSDL?

Our objective is to deﬁne S-WSDL; a common

representation of SWS to be made available for all

stakeholders regardless of the implementation plat-

forms. S-WSDL enhances WSDL with a “social”

dimension without altering WSDL’s original con-

tent. This can be done using Model Driven Ar-

chitecture (MDA). MDA represents everything from

requirements to business modeling and to technology

implementation (OMG, a). MDA requires models to

be expressed in a Meta-Object-Facility (MOF)-based

language (OMG, b). MOF is a standard for specify-

ing, constructing, and managing technology neutral

metamodels, i.e., models that describe other models.

When a system model refers to a speciﬁc platform it is

called Platform Speciﬁc Model (PSM). Contrarily, the

model is called Platform Independent Model (PIM)

(OMG, a). A model in PIM can be transformed into

another PIM when it needs to be enhanced, ﬁltered, or

specialized or into one or more PSMs. PIM into PIM

transformation represents model reﬁnement.

Our approach reﬁnes WSDL metamodel, inde-

Benna, A., Maamar, Z. and Nacer, M.

A MOF-based Social Web Services Description Metamodel.

DOI: 10.5220/0005687302170224

In Proceedings of the 4th International Conference on Model-Driven Engineering and Software Development (MODELSWARD 2016), pages 217-224

ISBN: 978-989-758-168-7

217

pendently of any platform. We inject some so-

cial elements into the description of this metamodel.

The outcome called S-WSDL metamodel is MOF

based. S-WSDL metamodel can serve different pur-

poses such as specifying social elements of a WS

(e.g., know with whom to partner and collaborate).

Automatic transformation from WSDL PIM into S-

WSDL PIM can be carried out using model-to-model

transformation languages such as Queries, Views, and

Transformations (QVT) (OMG, b). S-WSDL PIM

generated can in turn be transformed into PSM.

Our main contributions in this paper are as fol-

lows: identiﬁcation of SWSs’ properties; adoption of

MDA to develop S-WSDL metamodel (i.e., a way to

abstract these properties), and demonstration of how

S-WSDL metamodel based on MOF and QVT Rela-

tion supports automatic transformation of models.

The rest of this paper is organized as follows. Sec-

tion 2 provides an overview of SWSs approaches and

WSs models. Section 3 deﬁnes SWSs’ properties and

then illustrates SWS modeling. Section 4 discusses

how the proposed model is instantiated. Finally Sec-

tion 5 draws some conclusions.

2 RELATED WORK

We provide ﬁrst a brief overview of SWSs and then a

discussion on MDA use for WSs development.

2.1 Social Web Service Overview

A signiﬁcant amount of research looks into blending

social computing with service-oriented computing.

To perform WSs discovery, Maamar et al. develop

LinkedWS, a social network discovery model that

captures interactions between WSs (Maamar et al.,

2011b). This network identiﬁes collaborators and

substitutes of WS. For a better quality of WS discov-

ery, Chen et al. (Chen et al., 2015) construct a global

social network of WSs. They transform WSs’ WSDL

and OWL-S descriptions on the Web into RDF de-

scriptions. In this global social network, social links

between WSs that can work together during compo-

sition are deﬁned. Chen et al.’s network allows users

to identify a WS as an entry point in the network and

then navigate along social links to discover the WSs

deemed necessary for composition. El-Goarany et al.

propose a WS recommendation system within a so-

cial network of WSs that makes use of generated on-

tologies to discover similar and complementary WSs

(El-Goarany et al., 2008). Maaradji et al. use in-

formation collected from a user’s social networks

so that WS composition schemas are recommended

(Maaradji et al., 2011). In (Maamar et al., 2012)

Maamar et al. interleave social networks of users

and social networks of WSs to support WSs composi-

tion, execution, and monitoring. To provide semantic

WS composition, Xie et al. (Xie et al., 2008) deﬁne

social networks based on trust relationship between

users of WSs, providersof WSs, and WSs themselves.

Xie et al’s networks help select WSs or composi-

tion plans and make the search results meet users’ re-

quests. Last but not least, Bansal et al. (Bansal et al.,

2010) present a framework for trust-based WS com-

position. This trust rating is used to ﬁlter composition

results and then produce solutions that comprise WSs

provided by trusted providers.

The aforementioned approaches use different rep-

resentation of SWSs and depend on their platform

support and technologies(e.g., Java, .Net, and Prolog)

making it difﬁcult to identify all WSs’ social elements

(e.g., trust and collaboration).

2.2 MDA for Web Services Development

MDA addresses speciﬁc issues related to WSs such as

automating the generation of platform-speciﬁc imple-

mentations using WS models (B´ezivin et al., 2004),

extending WSDL for describing the qualities of WSs

(D’Ambrogio, 2006), and promoting interoperability

between WS

* protocols (Simon et al., 2013).

The adoption of MDA for describing WSs inter-

actions is quite new and very few studies are avail-

able. Bouchakour et al. use MDA to develop

SWSs (Bouchakour and Benslimane, 2013) in terms

of building social networks from a log ﬁle that deﬁnes

speciﬁc WSs interactions, assessing and modeling the

social qualities of each WS, and then adding these

qualities to a WSDL ﬁle using an XSLT template.

However, social qualities update (e.g., new interac-

tions between WSs) leads to rebuilding social net-

works, which is time consuming and error prone for

SWSs developers, providers, and users. Zheng et al.

(Zheng et al., 2012) extend OWL-S metamodel to

support the description of WS providers social con-

text.

It should be noted that (Bouchakour and Bensli-

mane, 2013) and (Zheng et al., 2012) approaches do

not embrace MDA principles (e.g., PIM and PSM)

and standards (e.g., MOF and QVT). They do not con-

sider the overall SWSs’ properties, and do not allow

incremental model transformation.

Contrarily, we propose to reﬁne WSDL PIM in or-

der to deﬁne a MOF-based S-WSDL metamodel, a

model abstracting a WS’s social elements. The use of

QVT Relation language provides an automatic trans-

formation of WSDL model into S-WSDL model, and

support target incremental model transformation.

MODELSWARD 2016 - 4th International Conference on Model-Driven Engineering and Software Development

218

3 MODELING SOCIAL WEB

SERVICE

This section identiﬁes properties that should deﬁne

WSDL description of SWS and describes the MDA

approach for enriching WSDL with these properties.

3.1 Properties of Social Web Service

To identify SWSs’ properties, we looked at the steps

that Beydoun et al. peformed in order to deﬁne the

concepts of a generic multiagent metamodel (Bey-

doun et al., 2009). These steps are summarized in

Table 1 and are looked at from both bottom-up and

top-down perspectives. The bottom-up perspective

(Steps 1-4) identiﬁes common SWSs’ properties us-

ing the existing literature on SWSs. The top-down

perspective (Step 5) validates the identiﬁed SWSs’

properties with regard to the existing SWSs literature.

In Step 1, we investigate SWSs literature using

academic library databases (e.g., Web of Science)

and academic search engines (e.g., Google Scholar).

SWSs approaches published in journals with high im-

pact factors and in conferences with high ranking

are the measures taken into account in selecting the

top n (n=10) SWSs approaches.

In Step 2, we decide on the general concepts rel-

evant for SWSs. These concepts are selected on the

basis of tasks that are necessary for building a concept

dictionary in Methontology

method (G´omez-P´erez

et al., 2004). The tasks are as follows:

• Task 1. Build a glossary of terms that includes all

relevant terms for SWS like concepts, instances of

concepts, attributes of concepts, and relations be-

tween concepts, their synonyms, and descriptions.

To build this glossary, we capture and collect

terms related to SWSs’ social networks and social

element from approaches identiﬁed in Step 1.

• Task 2. Build concept taxonomies to structure

concepts. We select terms that are of type con-

cept from the glossary of terms. Concepts are

examined based on middle-out strategy used in

Uschold and King’s method for building ontology

(G´omez-P´erez et al., 2004). This strategy recom-

mends identifying ﬁrst, the core of basic domain

terms and then specifying and generalising them

as required. In our case, we start by identifying

the main concepts that deﬁne a SWS’s social net-

work. As social relations mean connecting par-

ties together, graphs capture them (King et al.,

2009) using node and edge between nodes. We

http://www.core.edu.au/

Methontology enables building ontology from scratch

refer to edge as relationship. We have then gen-

erated top and bottom concepts for node and re-

lationship. However, the existing SWS literature

presents several terms that are similar and differ-

ent, and sometimes overlap. To reach a consensus

over SWS network community, we proceed as fol-

lows:

– Two concepts with overlapping or disjoint

meanings have the same name. As in

ONIONS

method (G´omez-P´erez et al., 2004),

we create a concept for each meaning. The new

concepts are linked to others concepts through

specialization or generalization (e.g., trust as

rating of WS specializes node degree concept

and is named trust degree, while trust between

user and WS provider specializes relationship

type concept and is named trust relationship).

– Same concepts having different names. As in

ONIONS method (G´omez-P´erez et al., 2004),

we adopt a name based on what is widely

agreed upon in the domain (i.e., in social net-

working community). For instance, we adopt

relationship concept name for relationship be-

tween nodes which is referred as explicit rela-

tionship (Maaradji et al., 2011).

– Some speciﬁc concepts are omitted, so they

specialize more general concepts of the con-

cept taxonomies. For instance, formulas for as-

sessing relationship weight are omitted. These

formulas can specialize get relationship weight

concept which in turn specializes a relationship

weight concept.

– Other speciﬁc concepts can be obtained from

the concept taxonomies by specializing more

general concepts. For instance, collabora-

tion and substitution between WSs in (Maamar

et al., 2011b) can be obtained by specializing

relationship type concept.

– When a concept is closely related to a speciﬁc

approach and cannot be obtained by specializ-

ing more concepts from concept taxonomies,

we create this concept. For instance, relation-

ship dependency concept is created to spec-

ify trust relationship between two WSs in (Xie

et al., 2008).

– Implicit concepts are included in the concept

taxonomies. For instance, relationship dura-

tion concept is inferred from (Chen et al., 2015)

and (Maamar et al., 2011b) approaches that re-

fer to update or management of relationship.

ONIONS method deﬁne how to generate a unique on-

tology concepts from original ontologies concepts.

A MOF-based Social Web Services Description Metamodel

219

Table 1: Steps for identifying SWS properties.

Step 1 Identify SWSs approaches from a literature search

Step 2 Extract general concepts related to SWS from approaches identiﬁed in Step 1

Step 3 Short-list candidate deﬁnitions of selected SWS concepts in Step 2

Step 4 Reﬁne list of SWS concepts, their corresponding deﬁnitions and relationships: output of this

step is the initial SWS properties

Step 5 Validate SWS properties by their instantiation on SWS approaches selected in Step 1

Table 2: SWSs’ properties deﬁned and their use in the literature.

SWS

properties

Approaches

Bansal

et al.

(Bansal

and

Bansal,

2011)

Bansal

et al.

(Bansal

et al.,

2010)

Maamar

et al.

(Maa-

mar

et al.,

2011a),

(Maa-

mar

et al.,

2011b)

Maaradji

et al.

(Maaradji

et al.,

2011)

Maamar

et al.

(Maa-

mar

et al.,

2012)

Xie et

al. (Xie

et al.,

2008)

Chen

et al.

(Chen

et al.,

2015)

El-

Goarany

et al.

(El-

Goarany

et al.,

2008)

Li et al.

(Li and

Chen,

2010)

Node WS

Provider

WS User and

WS’s

user and

User,

WS, and

Provider

WS WS WS,

User

and

Provider

Relationship

Type

Trust

between

WSs’s

pro-

viders

Trust

between

WSs’s

pro-

viders

Collabo-

ration,

substitu-

tion, and

compe-

tition

between

WSs

Reco-

mme-

ndation

conﬁ-

dence of

WS to a

user

Reco-

mme-

ndation

of WS

to a

user,

compe-

tition,

collab-

oration,

and sub-

stitution

between

WSs

Trust Pattern

social

link

(e.g.,

Se-

quen-

tial,

and

choice

Incom-

ing/

Out-

going

social

link)

Similar

and

comple-

mentary

WSs

similar

func-

tional-

ities or

object

relation-

ships in

the real

world

Node degree Reputa-

tion per

Trust

rating

per WS

× × × Trust

degree

of WS,

of WS

user and

of WS

Provider

degree

of node

× Repu-

tation

of WS

provider

Relationship

weight

× ×

√ √ √ √ √ √ √

Relationship

duration

× ×

√ √

× ×

√

× ×

Relationship

transitivity

√ √ √ √ √ √

× × ×

Relationship

dependency

× × ×

√

× ×

√

legend:

√

means available property in the approach and × means the opposite.

MODELSWARD 2016 - 4th International Conference on Model-Driven Engineering and Software Development

220

• Task 3. Build a concept dictionary which will

include all concepts, their corresponding deﬁni-

tions, their attributes, and relations.

In Step 3, we short-list the candidate deﬁnitions of

the selected concepts in Step 2. This happens by iden-

tifying what knowledge the deﬁnition speciﬁes and if

the deﬁnition deﬁnes explicitly the word. For knowl-

edge that is required in SWS social networks domain

but not explicit, it is inferred from other deﬁnitions.

Differences between deﬁnitions were reconciled

in order to ensure consistency across concept names.

We adopt a deﬁnition that is most coherent to main-

tain generality when contradictory use of concepts be-

tween two or more literatures occurred. For instance,

node refers to WS in (Chen et al., 2015) while it refers

to WS and user in (Maamar et al., 2012) and to WS

provider in (Bansal et al., 2010). Therefore, we de-

ﬁne node as an entity that could be WS provider, WS

consumer, or WS. The ﬁnal output of Step 3 is a re-

ﬁnement of the list of concepts obtained in Step 2 with

their corresponding deﬁnitions.

In Step 4, we reﬁne the list of SWS concepts ob-

tained in Step 3 and identify dependencies between

SWS concepts. The ﬁnal output of this step is the fol-

lowing deﬁned key properties of SWS.

• Node refers to an entity (physical or moral) that

could be WS provider (e.g., organization), WS

consumer (e.g., user), or WS as well. In the rest of

this paper, we use entities and nodes interchange-

ably.

• Relationship establishes a link (or association)

between pairs of nodes. More than one relation-

ship can exist between nodes (e.g., collaboration

and substitution), whether the nodes are of the

same kind or not. In the rest of this paper, we

use relationship and association interchangeably.

• Relationship Type speciﬁes the label of a re-

lationship. For instance, competition, collab-

oration, and recommendation relationships be-

tween WSs, trust relationship between WSs cus-

tomers, WSs providers, and WSs themselves, and

friendship or business partnership relationships

between WSs providers and WSs consumers.

• Relationship Weight evaluates the relationship

between two nodes in terms of how collaborative

they are, how cooperative they are, etc. This eval-

uation may depend on the number of interactions

that involve nodes.

• Relationship Duration establishes how long a re-

lationship between nodes will last. This could be

permanent like similarity between WSs’ function-

alities or temporary like upon-requestpartnership.

• Relationship Transitivity determines relation-

ship type that supports transitivity like similarity

between WSs. Transitivity cycles may be limited

by a threshold that is deﬁned by WS provider.

• Relationship Dependency captures the reliance

of a relationship on another. For instance, trust

relationship between two WSs in (Xie et al., 2008)

relies on existing relationships between their WS

users and WS providers.

• Node Degreereﬂects the social qualities of a node

as it interacts with other nodes. A social quality of

node is a mapping between a set of social parame-

ters (e.g., trustworthy, and cooperative) onto a set

of their values. For instance, a node social quality

can be used to build its reputation degree and help

selection in the case of competition.

In Step 5, we validate the SWS properties by their

instantiation on SWS approaches selected in Step 1.

Table 2 summarizes how the deﬁned SWS properties

are mapped onto concepts reported in the SWS litera-

ture.

3.2 MDA for Social Web Service

Description

In this section, we propose a MOF-based S-WSDL

metamodel for abstracting SWS’s properties. S-

WSDL metamodel consists of reﬁning WSDL meta-

model at the PIM level. In MDA terms, WSDL

model is an instance of WSDL metamodel. Like-

wise, S-WSDL model (i.e., WSDL model with social

dimension) is an instance of S-WSDL metamodel.

These metamodels are also instances of the MOF

model (Fig. 1). A S-WSDL metamodel’s concepts are

identiﬁed and grouped according to Interaction and

service views.

3.2.1 Interaction View

Interaction View models the aforementioned pro-

posed SWS properties. The basic view of the Inter-

action metamodel is illustrated in Fig. 2. A metaclass

deﬁnes the behavior of certain classes and their in-

stances. 1 or 1..* cardinalities denote the required as-

sociation, while 0..1 or 0..* cardinalities denote op-

tional associations. ‘SN” for “Social Network” is

added as a preﬁx to name the new Interaction meta-

model classes. The following outlines the role of the

key metaclasses, their metaattributes, and metafunc-

tions.

• SNNode refers to node. It is characterized by

name, state and a set of properties represented by

SNProperty metaclass.

A MOF-based Social Web Services Description Metamodel

221

Figure 1: Overview in using MDA for SWS description and transformation.

Figure 2: Interaction core metamodel View.

• SNAssociation

Type refers to relationship type

and aims to identify the nature of the social re-

lationships (e.g., collaboration and recommenda-

tion relationship as value instance of the SNAsso-

ciation

Type metaclass) and the representation of

information regarding symmetry, transitivity, de-

pendency, and temporal aspects.

• SNAssociation

Weight specializes the metaclass

SNproperty. In this case, prop

name attribute

refers to the label of association weight (e.g., col-

laboration weight as value instance of associa-

tion

weight metaclass) and Value attribute to its

weight. This value depends on the relationship

type and is calculated using the metafunction

Get

ass weight.

3.2.2 Service View

Service View extends WSDL metamodel with Inter-

action to create S-WSDL metamodel. WSDL meta-

model is generated from WSDL 2.0 XML schema

description. WSDL 2.0 XML schema allows ele-

ments representing a speciﬁc technology, called ex-

tensibilityelements, under various elements deﬁned

by WSDL. S-WSDL metamodel gives rise to social

characteristics as an optional entry for WSDL meta-

model. Thus, any service (e.g., Prepare Job) can re-

fer to services (e.g., Post

Job, and Plan Job) with

whom it has a relationship (e.g., collaboration) using

SNNode and SN

Association metaclasses.

MODELSWARD 2016 - 4th International Conference on Model-Driven Engineering and Software Development

222

4 IMPLEMENTATION AND

VALIDATION

In order to validate our proposed MOF-based

S-WSDL metamodel, we need to prove its com-

pleteness. However, incompleteness is a fundamental

problem in an open environment such as the Social

Web. In fact, we cannot prove the completeness of

a proposed MOF-based S-WSDL metamodel nor the

completeness of its deﬁnitions, but we can prove the

incompleteness of an individual SWS property or its

deﬁnition, and therefore we can deduce the incom-

pleteness of the proposed metamodel if at least one

SWS property deﬁnition is missing in the established

S-WSDL metamodel. So, S-WSDL metamodel is

complete if and only if all that is supposed to be in

the SWS is explicitly stated in it, or can be inferred

for various social WS scenarios.

We validate our S-WSDL metamodel by show-

ing its applicability and relationships to existing SWS

approaches such as (Maamar et al., 2011b) and (Xie

et al., 2008) described in Table 2.

Upon MOF compliant metamodels for sources

(e.g., WSDL, and Interaction metamodels) and target

(e.g., S-WSDL metamodel) introduced, Model Trans-

formation can be carried via transformation rules (not

included here for space reasons) between MOF com-

pliant metamodels, and implemented via a transfor-

mation engine (Fig. 1). Transformation rules are de-

scribed in QVT Relation model transformation lan-

guage. The use of QVT Relation language provides

an automatic transformation of WSDL model into S-

WSDL model, and support target incremental model

transformation. Thus, the formal semantics of MOF

and QVT Relation language reﬂects the conformance

relation between models and metamodels, and the

satisfaction of transformation rules between pairs of

models (Calegari and Szasz, 2013).

We have implemented a prototype in Eclipse

Modeling Framework (EMF)

and integrates me-

diniQVT

plugin. This later supports transforma-

tions expressed in QVT Relation. The WSs models

used are built on real-world WSs extracted from WS-

DREAM

a Web Service QoS Datasets (Zhang et al.,

2010) and converted to XMI standard representation.

XMI representation of Interaction models is created

from Interaction Ecore diagram (e.g., using Eclipse

Rich Client Platform). We executed our transforma-

tions on JVM version 1.8, running on MicrosoftWin-

dows 7 Professional, in computer system of Intel(R)

http://www.eclipse.org/modeling/emf/

http://projects.ikv.de/qvt

http://www.wsdream.net/

Figure 3: An excerpt of Social SOAP engineer S-WSDL

model.

Core (TM) i5-920CPU, running at 2.67 GHz, and

4 GB of RAM.

Transformation takes as inputs SOAP

engineer

WSDL model and its Interaction model and re-

turns Social

SOAP engineer S-WSDL model (Fig. 3).

QVT evaluation created 1 new element, set 19 fea-

tures and takes 250 ms. Social

SOAP engineer S-

WSDL model is validated using EMF validation, as

instances of WSDL metamodel.

The social dimension of Social

SOAP engineer S-

WSDL is described, as optional elements in WSDL,

inside dash line shape in Fig. 3. It includesTrust prop-

erty as in (Xie et al., 2008) approach and Collabora-

tion and Robustness relationship type as in (Maamar

et al., 2011b) social networks approach:

• Collaboration with JobAlert and JobSearch-

Helper SWSs. The collaboration weight

value between SOAP

engineer and JobAlert is

labeled Job alert Coll Weight, while collabo-

ration weight value between SOAP

engineer

and JobSearchHelper is labeled JobSearch-

Helper

Coll Weight.

• Robustness with LookupSoap SWS. The latter has

similar functional and non-functional properties

with SOAP

engineer and can replace it in case of

failure. The robustness relationship weight value

is labeled LookupSoap

Rob weight.

A MOF-based Social Web Services Description Metamodel

223

5 CONCLUSION

In this paper, we identiﬁed SWS properties and pro-

posed a MOF-based S-WSDL metamodel for the de-

scription of SWS. S-WSDL metamodel is deﬁned to

ensure consistency between different SWS applica-

tion models and extensibility in terms of new inter-

actions, regardless of the implementation platform.

MOF-based S-WSDL metamodel and QVT Relation

are used to support automatic transformation from

WSDL model into S-WSDL model (WSDL with so-

cial dimension) without altering the original content

of WSDL model. We implemented a prototype to

test our approach. The prototype illustrates how the

MOF-based S-WSDL metamodel is deﬁned and how

to automate the transformation of a WSDL model

into a S-WSDL model, using EMF and QVT Rela-

tion tools. Furthermore, the proposed S-WSDL can

be applied to serve different purposes such as adding

social dimension when querying WSs registries and

mapping from WSs model (e.g., existing UML mod-

els of BPELs process) to WSs with social dimension.

As future work, we aim to validate our metamodel in

real-world use-cases.

REFERENCES

Object Management Group, MDA Guide Revision 2.0.

Object Management Group, OMG Speciﬁcations.

Bansal, S. K. and Bansal, A. (2011). Reputation-Based Web

Service Selection for Composition. In World Congress

on Services. IEEE.

Bansal, S. K., Bansal, A., and Blake, M. B. (2010). Trust-

based Dynamic Web Service Composition using So-

cial Network Analysis. In Workshop on: Business Ap-

plications of Social Network Analysis. IEEE.

Beydoun, G., Low, G. C., Henderson-Sellers, B., Moura-

tidis, H., G´omez-Sanz, J. J., Pav´on, J., and Gonzalez-

Perez, C. (2009). Faml: A generic metamodel for mas

development. IEEE Trans. Software Eng., 35(6).

B´ezivin, J., Hammoudi, S., Lopes, D., and Jouault, F.

(2004). Applying MDA Approach for Web Service

Platform. In 8th International Enterprise Distributed

Object Computing Conference. IEEE.

Bouchakour, E. H. and Benslimane, S. M. (2013).

Social Web Services Development Based on

MDA:Extending WSDL to Inject Social-QoS.

In 14th Arab Conference on Information Technology,

Proceedings.

Calegari, D. and Szasz, N. (2013). Institution-Based Se-

mantics for MOF and QVT-Relations. In Formal

Methods: Foundations and Applications. Springer.

Chen, W., Paik, I., and Hung, P. C. K. (2015). Constructing

a Global Social Service Network for Better Quality of

Web Service Discovery. IEEE Transactions on Ser-

vices Computing, 8(2).

Chung, J.-Y., Lin, K.-J., and Mathieu, R. G. (2003). Guest

Editors’Introduction : Web Services Computing-

Advancing Software Interoperability. IEEE Com-

puter, 36(10).

D’Ambrogio, A. (2006). A Model-driven WSDL Extension

for Describing the QoS of Web Services. In Interna-

tional Conference on Web Services. IEEE.

El-Goarany, K., Saleh, I., and Kulczycki, G. (2008). The

Social Service Network - Web 2.0 Can Make Seman-

tic Web Services Happen. In 10th Conference on E-

Commerce Technology & Enterprise Computing, E-

Commerce and E-Services. IEEE.

G´omez-P´erez, A., Fern´andez-L´opez, M., and Corcho,

O. (2004). Ontological Engineering: With Exam-

ples from the Areas of Knowledge Management, e-

Commerce and the Semantic Web. Springer.

King, I., Li, J., and Chan, K. T. (2009). A brief survey

of computational approaches in social computing. In

International Joint Conference on Neural Networks.

IEEE.

Li, S. and Chen, Z. (2010). Social Services Computing:

Concepts, Research Challenges, and Directions. In

Int’l Conference on Green Computing and Commu-

nications,& Int’l Conference on Cyber, Physical and

Social Computing. IEEE.

Maamar, Z., Faci, N., Sheng, Q. Z., and Yao, L. (2012).

Towards a User-Centric Social Approach to Web Ser-

vices Composition, Execution, and Monitoring. In

Web Information Systems Engineering. Springer.

Maamar, Z., Hacid, H., and Huhns, M. N. (2011a). Why

Web Services Need Social Networks. IEEE Internet

Computing, 15(2).

Maamar, Z., Wives, L. K., Badr, Y., Elnaffar, S., Boukadi,

K., and Faci, N. (2011b). LinkedWS: A Novel Web

Services Discovery Model Based on the Metaphor of

”Social Networks”. Simulation Modelling Practice

and Theory, 19(1).

Maaradji, A., Hacid, H., Skraba, R., Lateef, A., Daigre-

mont, J., and Crespi, N. (2011). Social-Based Web

Services Discovery and Composition for Step-by-Step

Mashup Completion. In International Conference on

Web Services. IEEE.

Simon, B., Goldschmidt, B., and Kondorosi, K. (2013). A

Metamodel for the Web Services Standards. J. Grid

Computing, 11(4).

Xie, X., Du, B., and Zhang, Z. (2008). Semantic Service

Composition based on Social Network. In Proceed-

ings of the 17th International World Wide Web Con-

ference. Springer.

Zhang, Y., Zheng, Z., and Lyu, M. R. (2010). WSExpress:

A QoS-aware Search Engine for Web Services. In In-

ternational Conference on Web Services. IEEE.

Zheng, X., Wu, Q., Ke, D., Li, H., and Shi, Y. (2012).

Social Context Enabled Description Model for Web

Services. In Information Computing and Applications

conference. Springer.

MODELSWARD 2016 - 4th International Conference on Model-Driven Engineering and Software Development

224