User Requirement and Behavioral Aspects in Web Service Discovery

Wala Ben Messaoud

, Khaled Ghedira

and Youssef Ben Halima

Institute ENSI University of Manouba/SOIE, 41 Liberty Street Bouchoucha, Bardo, Tunisia

ENSI University of Manouba/RIADI Labs, Manouba University Campus, Manouba, Tunisia

Keywords: Behavioral Aspect, State Chart Scheduling, Web Service, Web Service Discovery, WordNet.

Abstract: In web service (WS) discovery, behavioral aspect has been defined as the sequence of WS operations. The

motivation to introduce the behavioral aspect is to offer to the consumer the possibility to choose his WS

according to his requirements. The aim is to include the execution manner of WS operations as a new

criterion and to apply a selection method if more than one WS candidate is filtered. In this paper, we

envision to implement WS discovery approach based on behavioral aspects to fulfill the selection of the

precise execution order. This approach ensures an execution order of operations in accordance to consumer

needs. The execution manner criterion is defended by state chart as a scheduling method and WordNnet as a

lexical database. Moreover, semantic equivalences have to be considered in order to solve equivalence

between many WS candidates which satisfy consumer needs.

1 INTRODUCTION

The literature on web service discovery is almost

common in recognizing the existence of a major

problem in the WS consumer's requirements.

Current work on service discovery focuses on

discovery types not on the analysis of consumer

intervention.

WS discovery is the process of satisfaction of a

user request according to his requirements. It refers

to the process of finding WS that implements the

technique of search desired, interviewing service

books, to know what WS is available for binding.

Our approach unlike the other discovery approaches,

allows WS consumer to involve his exigency by

entering some sentences as a WS query. The aim is

to satisfy WS consumer by analyzing his inputs.

The WS consumer requires a new aspect

allowing functional phase (organized operations,

free operations) and non functional phase (cost,

time, availability). As a solution, we define the

behavioral aspect as the execution manner of WS

operations. Indeed, a consumer who requires looking

for a WS with a tidy list of exigencies may not be

satisfied by ordinary WS discovery. Actually, the

motivation to introduce the behavioral aspect is to

offer to the consumer the possibility to choose his

WS according to his needs. The aim is to include

requirement criterion as a new aspect and to apply a

selection method if more than one WS candidate are

filtered. The behavior aspect should guarantee

quality of service (cost, reliability, time ...).

Our approach consists on using Statechart as a

scheduling method to highlight the execution order

of WS operations and WordNet as a lexical database

to prove the semantic part of consumer query. We

used BPEL4WS (Business Process Execution

Language for Web Services) to specify and execute

business process for WS composition and

orchestration. More precisely, to explain how WS

operations are invoked and executed.

The remainder of the paper is organized as

follows: Section 2 presents some concepts used in

our approach. Our solution is reported in section 3.

Section 4 explains more our approach by an

example. In section 5, we feature the related work.

Conclusions are presented in section 6.

2 CONCEPTS

Our approach brings answers to many posed issues.

It is based on some concepts facilitating the

implementation of the different phases of the

approach like WordNet, intending to find synonyms

to all consumer inputs.

In the goal to order these inputs, we use state

charts where the WS operations are the transitions

allowing to move from one state to another. Once

we have prepared the state chart part, we launch the

199

Ben messaoud W., Ghédira K. and Ben Halima Y..

User Requirement and Behavioral Aspects in Web Service Discovery.

DOI: 10.5220/0005487601990205

In Proceedings of the 5th International Conference on Cloud Computing and Services Science (CLOSER-2015), pages 199-205

ISBN: 978-989-758-104-5

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

semantic aspect of WS discovery, we extract the

BPEL sequencing of each WS found and then we

select the relevant one.

2.1 WordNet

The goal of WordNet was to develop a system that

would be consistent with the knowledge acquired

over the years about how human beings process

language.

In (Miller, 1995), WordNet is defined as a large

lexical database of English. Nouns, verbs, adjectives

and adverbs are grouped into sets of cognitive

synonyms (synsets), each expressing a distinct

concept. Synsets are interlinked by means of

conceptual-semantic and lexical relations.

WordNet is a large semantic network interlinking

words and groups of words by means of lexical and

conceptual relations represented by labeled arcs.

WordNet’s building blocks are synonym sets

(synsets), unordered sets of cognitively synonymous

words and phrases (Christiane, 2005).

The authors in (Morato, 2004), define WordNet

as one of a series of manually compiled electronic

dictionaries, is restricted to no specific domain and

covers most English nouns, adjectives, verbs and

adverbs. WordNet offers researchers a cost-free use

and well-documented open code. It is an ideal tool

for disambiguation of meaning, semantic tagging

and information retrieval.

In our work, the consumer inputs is a set of

keywords that define his requirements. If we don’t

specify clearly these inputs, we risk falling in

linguistic polysemy case. So, we seek to find

synonyms to all consumer inputs to properly filter

the service concerned. That's why we have chosen to

use WordNet as ontology.

2.2 Statecharts - Automaton

In literature, authors define statechart as visual

formalism for description of complex systems

behaviour. Digital controllers, which act as reactive

systems, can be very conveniently modelled with

statecharts and efficiently synthesized in modern

programmable devices (Łabiak, 2010).

In formal grammar, an alphabet Σ is a finite and

not empty set of symbols. Σ * is the closure of Σ. A

language on an alphabet Σ is a subset of the set Σ *.

A transition diagram allows achieving an

operational vision of the concept of language. It is a

finite collection of states and transitions.

The statechart notation was developed by David

Harel (Harel, 1987). Statechart diagrams are useful

for modelling the lifetime of an object. They are

used to describe the system behavior using a finite

automaton.

The automaton is represented as a directed graph

known as state graph which consists of a finite set of

vertices known as nodes, together with a set of

directed links between pairs of vertices called arcs.

Vertices are represented by circles and arcs by

arrows. We can also represent an automaton with a

state-transition table (Bhattacharjee, 2014).

The transition table can be associated to the

automaton which describes extensively the transition

function δ. A column is a character of the alphabet.

A line is a state of the automaton (the initial state is

preceded by an arrow "→", the final state is

preceded by a star "*").

The value δ (q, a) for q ∈ Q, a ∈ Σ corresponds to

the state at the intersection of the row q and the

column a. Note that from this table it is easy to find

all the statements and the alphabet and thus identify

exactly the automaton.

An automaton reads a finite string of symbols a

a2, a

, where a

∈ Σ, which is called an input word.

The set of all words is denoted by Σ*. Accepting

word is the word w ∈ Σ* which is accepted by the

automaton if q

∈ F. An automaton can recognize

a formal language. The language L ⊆ Σ* recognized

by an automaton is the set of all the words that are

accepted by the automaton. The Deterministic Finite

Automata has a finite internal memory available. At

each input letter the state of the internal memory is

changed depending on the letter scanned.

The previous memory state and the input letter

together determine what the next state of the

memory is. The word is accepted if the internal

memory is in an accepting state after scanning the

entire word (Kari, 2013).

In our approach, we use automaton as a

scheduling method of WS operations (where we

order the WS operations in the aim to respect the

non functional/functional properties). Each operation

represents a transition from a state to another. The

automation generates a language (ordering list of

operations) that specify the words to accept.

2.3 WSDL

A WSDL document is, at its simplest, a collection of

elements contained within a root definition element.

These elements describe a service and how an

endpoint implementing that service is accessed

(ALBRESHNE, 2009)

Each WSDL includes two parts, the abstract and

the concrete. Abstract part describes the messages

CLOSER2015-5thInternationalConferenceonCloudComputingandServicesScience

200

sent and received. The operation associates a

message exchange pattern with one or more

messages.

As for concrete part, it specifies transport and

wire format details for one or more interfaces, a port

(an endpoint) associates a network address with a

binding and a service which groups together

endpoints that implement a common interface.

An operation is similar to a function in a high

level programming language. A message exchange

is also referred to as an operation. Operations are the

focal point of interacting with the service.

In our work, we use WSDL for each WS to extract

the execution order of operations.

2.4 BPEL4WS

BPEL for WS is a programming language for the

execution of business processes. It is based on

WSFL (Web Services Flow Language) and XLANG

(XML LANGuage), is derived from XML.

According to (Milanvoic, 2004) and (OASIS,

2007), BPEL composes web services to get a

specific result. The composition result is a named

process, partners are defined as services and

activities are described by exchanged messages. In

other words, a process contains a set of activities and

it invokes external partner services using a WSDL

interface.

BPEL is used to describe the execution of

business process implicating WS. It consists on

business BPEL allowing communication with WS,

handling XML Data and managing exceptions.

BPEL provides several structure activities:

- <sequence>: define an ordered sequence of WS

activities.

- <flow>: define parallel activities.

- <switch>: Case-switch construct for implementing

branches.

- <while>: define loops.

- <pick>: select one of several alternative paths.

In WS discovery, there are three layers;

operational, organizational and intentional.

Intentional layer enables modeling purposes. It is a

conceptualization of strategic needs of required

modeling by an individual subject, group of

individuals, a work unit or organization that

involved in the system development process.

In some cases, the WS consumer cannot be a

domain expert, he launch his query by entering a list

of sentences. These sentences will be transformed to

keywords by an algorithm to facilitate the

satisfaction process. The consumer inputs can be a

list of WS that need to be orchestrated as operations.

3 SOLUTION: BEHAVIORAL

BASED APPROACH

The WS consumer seeks his WS with a list of well-

defined requirements (functional and non-

functional) but he is not satisfied by the ordinary

types of discovery.

Syntactic discovery aims to compare between the

syntactic query based on keywords and syntactic

descriptions of WS. WS consumer launch his query

by entering a set of keywords, the result of the

comparison between syntactic query based on

keywords and syntactic descriptions of the services

is a set of WS that don't satisfy the consumer

exigencies (the WS name is exactly the keyword

entered by the consumer but the content has not the

same meaning sought).

The semantic discovery is mainly based on

ontology, defined as structured set of terms and

concepts representing the meaning of information

field, developed to facilitate knowledge sharing and

reuse. WS consumer launch his query by entering a

set of keywords, the result is a set of WS that don't

satisfy the consumer exigencies. For this reason, we

intend to define a new WS discovery approach based

on behavioral aspect according to our definition. We

define the behavior as the execution manner of WS

operations. WS discovery approach based on

behavioral aspect ensures an execution order of

operations in accordance to consumer needs.

The purpose of this paper is to design a

discovery technique for choosing the WS with the

most relevant behavior. So, the query language to

develop should be based on system statechart. It is

used to check the compatibility of behavior required

by the customer and those of WS found.

The work requires its valuation by an

implementation that acquires to client to get his WS

with the execution order of the desired operations.

This implementation creates a WS automaton

according to the operations order. WS automaton

accepts only languages that correspond to it. These

languages will be generated from BPEL files

(Business Process Execution Language) of WS

operations searched in WS directories. Some

languages will be selected if they are accepted by the

WS automaton.

Figure 1 shows that the implementation is done

in three phases.

UserRequirementandBehavioralAspectsinWebServiceDiscovery

201

3.1 Phase 1 - Transform Consumer

Sentences to Keywords

If the WS consumer is not a domain expert, he may

enter his requirements as sentences. An algorithm is

defined to transform each sentence in a keyword

used in Phase 2.

3.2 Phase 2 - Create WordNet File for

Each Keyword - Create Automaton

WordNet is used to define synonyms file for each

keyword entered by the consumer. The analysis of

consumer inputs is done on a semantic level not on a

syntactic level.

In the same time, these keywords allow to create

the automaton corresponding to the behavior of

searched WS (defining the standard language to

accept).

The automaton is defined by an initial state, a list

of transitions and a list of final states. A number of

states are defined according to the consumer inputs.

To switch from a state to another, we must pass

through a transition. Consumer keywords define

transitions.

3.3 Phase 3 - Extract Sequences from

WSDL/BPEL Files of WS Searched

- Select the Relevant WS

Consists on searching WS semantically (Semantic

Aspect) basing on non functional properties (NFP).

For each WS selected, we extract the list of its

operations from WSDL (if it is a simple WS) or

from BPEL file (if it invoke other WS). The

extracted sequencing is transformed to word that

will be accepted or rejected by the automaton

Figure 1: Implementation Steps of Behavioral approach.

already defined.

Accepted words correspond to WS accepted. The

last step in Phase 3 is to select the most relevant WS

basing on functional properties (FP) that we called

(Behavioral Aspect).

4 EXAMPLE: STAY

RESERVATION

Let's consider the example named «Stay

reservation» in Figure 2, where the consumer wishes

to book a plane ticket, rent a car, buy a concert ticket

and book a hotel stay (or rent a house) with a heated

pool. All these operations should guarantee

minimum transfer cost and reduced transfer time.

Figure 2: Behavioral WS discovery.

The consumer launches his query by entering a list

of sentences such as «Plane ticket», «Heated pool»,

«Rent a car», «Hotel stay», «Minimum execution

cost», «Reduced execution time», «Rent a house»,

«Buy concert ticket», «Minimum transfer cost» and

«Reduced transfer time».

The first step is to specify NFP from FP. NFP are

used in semantic aspect as selection properties. In

this example, «Minimum execution cost» and

«Reduced execution time» are NFP. As a result, we

find WS1 named «Stay», WS2 named «Reservation»

and WS3 named «Journey». All of this WS satisfy

the consumer needs semantically.

For the behavioral aspect, «Minimum transfer

cost» and «Reduced transfer time» are chosen as FP

to select the most relevant WS.

To explain in more detail, we follow the steps

mentioned previously:

 Transform consumer sentences to

keywords:

Phase 1 presented in Figure 3 consists on defining an

algorithm to transform consumer inputs to keywords

CLOSER2015-5thInternationalConferenceonCloudComputingandServicesScience

202

usable in phase 2.

Figure 3: Stay Reservation: Phase 1.

 Create WordNet file for each keyword:

The system launches WordNet. As a result, we get

for each keyword a file with a list of its synonyms.

Each row of WordNet file is a keyword synonym.

Figure 4 presents two examples respectively bonded

to keywords «Plane ticket» and «Car rented».

Figure 4: WordNet result.

 Create automaton:

At the same time, to respect consumer's

requirements, we should consider it to create

automaton.

The automaton in Figure 5 shows all states that

the WS «Stay reservation» can be in during the

course of its life. Furthermore, it shows the possible

transitions between the states and the events that

initiate these transitions.

We follow the automaton to check the

compatibility of behavior required by the customer

and those of WS found. The following abbreviations

are used to develop the automaton.

Firstly, we define the states:

0: Stay not reserved

1: Plane Ticket reserved

2: Car rented

3: Hotel Stay reserved

4: House rented

5: Concert Ticket bought

Secondly, we define the transitions by the

following abbreviations:

 Reserve_ticket : P

 Rent_ car : C

 Reserve_Hotel : S

 Rent_ house : H

 Buy_Concert_ticket : B

The WS automaton (Q, Σ, δ, q

, F) is defined by:

Q = {0, 1, 2, 3, 4, 5}: the states number depends on

operations number. (It is equals to the real

operations number +1). The real operations number

is calculated by neglecting free operations and for

the parallel operations we just count one operation.

In our case, we have in general five operations. If we

neglect the free operation (Buy_Concert_ticket) and

we count just one operation for the two parallel

operations (Reserve_Hotel and Rent_House), we

will have a number of four real operations. So, the

states number equals to five (the real operations

number=4 +1).

Σ = {P, C,S,H,B }

δ(0, P)=1, δ(0, B)=5, δ(1, C)=2, δ(1, B)=5,

δ(2, S)=3, δ(2, H)=4, δ(2, B)=5, δ(3, B)=5,

δ(4, B)=5, δ(5, P)=1, δ(5, C)=2, δ(5, C)=3,

δ(5, H)=4

The transitions table is represented by a matrix

where the rows are the states and the columns are

the operations.

= 0 : the initial state is always the 0.

F = {3, 4, 5}: final state is defined if we look over

all the transitions (for the parallel operations, we

count just one). In this example, to achieve 3, we

pass by PBCS. To achieve 5, we pass by PCSB. To

achieve 4, we pass by PBCH.

Figure 5: Automaton of WS «Stay reservation».

Then we deduce the transition table to facilitate the

extraction of language to accept:

Table 1: Transition table.

P C S H B

→0 {1} Ø Ø Ø {5}

Ø {2} Ø Ø {5}

Ø Ø {3} {4} {5}

Ø Ø Ø Ø {5}

{1} {2} {3} {4} Ø

According to the transitions table, the accepted

expressions are:

UserRequirementandBehavioralAspectsinWebServiceDiscovery

203

 {PBCS}: 1- Reserve_ticket 2-

Buy_Concert_ticket 3- Rent_ car 4-

Reserve_Hotel.

 {PBCH}: 1- Reserve_ticket 2-

Buy_Concert_ticket 3- Rent_ car 4- Rent_

house.

 {PCBS}: 1- Reserve_ticket 2- Rent_ car 3-

Buy_Concert_ticket 4- Reserve_Hotel.

 {PCBH}: 1- Reserve_ticket 2- Rent_ car 3-

Buy_Concert_ticket 4-Rent_ house.

 {PCSB}: 1- Reserve_ticket 2- Rent_ car 3-

Reserve_Hotel 4- Buy_Concert_ticket.

 {PCHB}: 1- Reserve_ticket 2- Rent_ car 3-

Rent_ house 4- Buy_Concert_ticket.

 {BPCS}: 1- Buy_Concert_ticket 2-

Reserve_ticket 3- Rent_ car 4-

Reserve_Hotel.

 {BPCH}: 1- Buy_Concert_ticket 2-

Reserve_ticket 3- Rent_ car 4- Rent_

house.

So, all other execution enchainment of WS

operations will be rejected.

 Extract sequences from WSDL or BPEL

files of WS searched:

The first step is to launch WS discovery

semantically basing on NFP «Minimum execution

cost» and «Reduced execution time». As a result, we

find WS1« Stay», WS2«Reservation» and

WS3«Journey». This step is an ordinary WS

discovery based on semantic aspect that we can find

in many works like (Kritikos, 2007).

The second step is to extract the sequencing of

WS operations from the concerning file of each WS

filtered. More precisely, to extract words that will be

accepted or rejected by the automaton defined

previously.

Every WSDL file contains the tag <operation>

that defines the list of WS operations. We extract the

execution order and transform it to words. Hence, all

the WS words are ready to be filtered.

If we have composed WS, every BPEL file

contains the tag <sequence> that defines an ordered

sequence of WS activities. We extract this WS

operations sequencing and transform it to word.

Hence, all the WS words are ready to be filtered.

 Select the relevant WS:

Basing on FP «Minimum transfer cost» and

«Reduced transfer time», we launch the selection

phase of the most relevant WS.

The services candidate namely WS1«Stay», WS2

«Reservation» and WS3 «Journey» satisfy the

consumer semantically but only WS2 is selected as

the most relevant WS. Calculations are made in

order to reduce transfer time and transfer cost.

5 RELATED WORK

The WS behaviour in literature is described by

sequences of messages, data types, data constraints

and properties that specify time limits within\where

messages are exchanged (Elabd, 2011).

Previously, Maamar et al. propose an approach

for modelling and specifying behaviours of WS in

(Maamar, 2009). This approach sheds the light on

two types of behaviours: control (that demonstrate

the business logic that supports the functioning of a

WS) and operational (that regulates the execution

progress of this control behaviour by stating actions

to carry out and the constraints to put on this

progress). The idea is to coordinate both behaviours

at run-time by developing conversational messages

and transmit details between these two behaviours.

Unlike (Maamar, 2009), our approach treats

behaviour as a sequence of ordered operations and

abstracts away from considering behaviours

conversation.

Another alternative is to first propose a service

system by describing the overall behaviour of each

consumer, and then to instantiate such consumers

retrieving services exposing a behavioural contract

which is adequate to the matching given behaviour

(Bravetti, 2009).

The work in (Sriharee, 2003) presents an

approach based on ontology to improve descriptions

of WS that are defined in WSDL with ontology-

based behavioural information, the query for

services are based on behavioural constraints and

have a service ontology linked with each WS. It can

benefit from inferring semantics of the service from

the service ontology. This work neglects the

consumer needs, it doesn't depend on the execution

order of WS operations.

The aim of WS discovery based on behavioural

aspect prove that WS consumer is the most

important part in discovery process, whose role is to

specify the WS description as well as its behaviour.

The authors in (Ramollari, 2008) propose an

approach where the service provider enhances the

WSDL document by means of a formal model of the

WS behaviour, expressed in the stream X-machine

formalism (SXM). This model is used by the service

broker during publication and by the service

consumer during discovery.

All these works treat behavioural aspects as

conversational messages exchanged between WS

and abandon the internal structure of WS execution.

That is why the approach that we propose in this

paper, involves the importance of sequencing of WS

operations to guarantee the quality of service.

CLOSER2015-5thInternationalConferenceonCloudComputingandServicesScience

204

6 CONCLUSIONS

We hope you find the information in this template

useful in the preparation of your submission.

Service consumers have a choice between

different WS candidates that provide similar

functions. Accordingly, comparing services requires

more sophisticated patterns of discovery.

In this paper, we have proposed a WS discovery

approach based on behavioral aspects. We sought to

satisfy consumer needs by introducing new criteria

based on user requirements. This approach is based

on execution manner of WS operations. To fulfill the

aim of our approach, we arranged consumer

requirements in semantic and behavioral aspect.

Hence, we have organized our work in three phases;

phase1 consists to transform consumer sentences to

keywords if the consumer is a non expert domain.

Phase2 is a semantic WS discovery basing on

WordNet as lexical database and then create the

automaton to put in order WS operations and to

extract the language accepted. Phase3 aims to

extract sequences from WSDL/BPEL files of WS

searched to select the relevant WS.

In future work, we plan implementing our

approach by developing a query language that uses

the underlying automaton to the required WS. A

case study will be set up to explain the objectives of

the approach. Also experimentation should be made

to highlight the advantage of using the approach. A

work is underway to fulfill this objective.

As for Selection phase, we can add other criteria

to select the most appropriate service if many

services satisfy the desired behavior.

REFERENCES

G. A. Miller, 1995. WordNet: A Lexical Database for

English, Communication of the ACM, Vol. 38, No. 11.

F. Christiane, 2005. WordNet and wordnets, In: Brown,

Keith et al. (eds.), Encyclopedia of Language and

Linguistics, Second Edition. Oxford: Elsevier, pages

665-670.

J. Morato et al., 2004. WordNet Applications. GWC 2004

Global Wordnet Conference (poster session).

http://www.fi.muni.cz/gwc2004/. Brno.

G. Łabiak

and G. Borowik, 2010. Statechart-based

Controllers Synthesis in FPGA Structures with

Embedded Array Blocks. International Journal of

Electronics and Telecommunications. Volume 56,

Issue 1, Pages 13–24, ISSN (Print) 0867-

6747, DOI: 10.2478/v10177-010-0002-7.

D. Harel, 1987. STATECHARTS: a visual formalism for

complex systems*, Science of Computer

Programming 8, 1987, pages 231-274.

A. Bhattacharjee, B. S. Purkayastha, 2014. A Novel

Approach to Construct Deterministic Finite State

Automata. International Journal Of Engineering And

Computer Science ISSN:2319-7242 Volume 3 Issue 1,

Page No. 3700-3703.

J. Kari, 2013. Automata and formal languages. Fall

semester 2013 University of Turku.

A. ALBRESHNE et al., 2009. Web Services Orchestration

and Composition Case Study of Web services

Composition, WORKING PAPER.

N. Milanvoic, and M. Miroslaw, 2004. Current Solutions

for Web Service composition. s.l.: IEEE Computer

Society.

OASIS, 2007. Web Services Business Process Execution

Language (WSBPEL).

K. Kritikos and D. Plexousakis, 2007. OWL-Q for

Semantic QoS-based Web Service Description and

Discovery, Proceedings of the 6th International

Semantic Web Conference and the 2nd Asian Semantic

Web Conference,.BEXCO, Bussan KOREA. pages

123-137.

Elabd, 2011. Conformité de services Web par rapport à

des spécifications de haut niveau.

Z. Maamar et al., 2009. A New Approach to Model Web

Services Behaviors based on Synchronization.

International Conference on Advanced Information

Networking and Applications Workshops.

M. Bravetti, 2009. Foundational Aspects of Service

Discovery based on Behavioral Contracts, 3rd

International Workshop on Verification and

Evaluation of Computer and Communication Systems.

N. Sriharee and T. Senivongse, 2003. Discovering Web

Services Using Behavioral Constraints and Ontology.

Distributed Applications and Interoperable Systems -

DAIS, pages. 248-259.

E. Ramollari et al., 2008. Towards Reliable Web Service

Discovery through Behavioral Verification and

Validation. Proceedings of the 3rd European Young

Researchers Workshop on Service.

UserRequirementandBehavioralAspectsinWebServiceDiscovery

205