Software architectural patterns for

International Journal of Computer Science & Information Technology (IJCSIT) Vol 6, No 4, August 2014
SOFTWARE ARCHITECTURAL PATTERNS FOR
SERVICE COMPOSITION
Ghadeer Ghazal1, Amjad. Hudaib2, Waffa Maitah3
1,2Department of Computer Information Systems, Faculty of King Abdullah II, The
University of Jordan, Jordan
3 Department of Computer Science, Faculty of Information Technology, Al al-Bayt
University, Jordan
ABSTRACT
Service-oriented computing is meant to support loose relationships between organizations; Service-oriented
architectures often have the goal to integrate various distributed services of one or more
organizations in a flexible way to be able to quickly react on business changes.
Distributed services provided a new way of distributed computing that achieve the interoperability between
heterogonous application through platform and language independent interfaces. The creation of value
added services by composition of existing ones is gaining a significant momentum. Distributed service
composition is meant to support loose relationships between implemented services in order to provide new
functions. A composite service is the one resulting from the integration, coordination and synchronization
of different service components. In this paper, we generated A Services Composition Model (SCM) that
provides a general solution for the services composition problem by realizing the requirements of a new
service using the requirements of the already existing service. We explained in details all the steps of the
composition process; services registration, services discovery, services selection, services invoking, and
services integration. Although the SCM is not bounded to one particular algorithm to compose services, we
generated an application as an example to test our Service Composition Model.
We also generated the Services Composition Language (SCL) as a simple text-based language which
allows the user to express the requirements of his request, the inserted request will then be analyzed using
our Parsing Algorithm to determine the name of the requested services, after that our Service Composition
Algorithm will execute all the steps of the composition process and return the result of the composition to
the user.
KEYWORDS
Services, Services Composition, Architectural Pattern, Services Composition Model (SCM).
1. INTRODUCTION
Services offered online are defined as a loosely coupled, reusable software component that
encapsulates discrete functionality. One essential characteristic of such services is the ability to be
described, published, discovered and invoked dynamically in a distributed computing
environment. These services can range from simple calculations and data retrievals to complex
business applications (Sommerville, 2007), (Agarwl, et al., 2008).
DOI:10.5121/ijcsit.2014.6403 31

In this research, we attempt to enhance the performance of Service Composition, which aims at
increasing reusability, flexibility and maintainability by breaking down the functionalities into
services running platform-independently and exhibiting their interface with a description of
functionality and the domain of objects employed.
Another significant point of services composition is that composite services must not be arranged
at design time. The final functionalities of services working together are defined by the service
request and must not be specified beforehand. When composing services, a composition only
needs to include a reference that identifies what service will be invoked, the dynamic binding
takes place as late as possible when the service is actually invoked.
There are situations in which a client request cannot be satisfied by any single available service,
but a composite service obtained by combining some available services might fulfill such a
request. The services used to form a composite service are referred to as component services.
1.1 SERVICES
Services provide a new way of distributed computing that achieves the interoperability between
heterogonous application through platform and language independent interfaces. The creation of a
value which has added services by composition of the already existing ones is gaining a
significant momentum. (Guoping, et al., 2008) t (Rao, et al., 2005).
1.2 ARCHITECTURAL PATTERNS
Software architectural pattern is a description of the architectural design problem and the essence
of its solution, (Hudaib and Montangero, 2002), (Cooper, 1998) and (Tut and Edmond, 2002).
Software architectural pattern has the following essential elements (Gamma, et al., 1998).
• Pattern Name: A meaningful name to refer to the pattern, typically a single word or short
32
phrase.
• A description of the problem: with a concrete example and its context, (Al-Masri,et al.,
2008). It also describes a solution to the concrete example and when to use the pattern.
• General solution: describes the elements of a design, their relationships, responsibilities
and collaborations.
• Consequences: good and bad of using the given solution. It describes the results and
trade-off of applying the pattern.
• Related patterns: naming patterns that are closely related (Buschmann, et al, 2007).
1.3 Automated Web Service Composition
Automated distributed Service Composition denotes a method to integrate services at runtime.
The complete service runs by composing components based on a service request. The service
request specifies the requested service, which does not exist before issuing the request. (Gekas,
et al., 2005).

33
2. LITERATUE REVIEW
2.1 SOA Structure
Keen, et al. (2004) interested in SOA; the basic assumption of SOA is that there are many
consumers that require services. Consumers are also referred to as clients or customers; there are
many providers that provide services on the network. These two groups have to be linked together
in a dynamic way.
Service providers register their services at the broker (Chmielowiec, et al., 2008); service
consumers request a service from the service broker, which returns a known provider for the
requested service. Consumer and provider agree on the semantics. The consumer then binds
himself to the service provider and uses the service.
2.2 Services Composition Models
Stein, et al. (2009) defined the workflow of the service model and the services that it
can provision in order to execute these workflows and describe the lifecycle Figure
(2.2) shows the Workflow Lifecycle for the service model.
(1) Workflow Selection. a general workflow is chosen to suit the consumer’s current objectives.
This is generally created either manually by domain experts or automatically by a planner.
(2) Matching. Once a general workflow has been selected, abstract tasks are mapped to
candidate service instances via a matching process. Here, the consumer searches a public
service registry or requests matching services from a broker.
(3) Provisioning. Given lists of matching services, the consumer provisions individual service
instances for each task of the workflow.
(4) Invocation. When appropriate services have been provisioned, the consumer starts to invoke
the chosen services as dictated by the ordering constraints of the workflow. If services fail to
complete their tasks, the consumer may provision other services, until the workflow is
successfully completed.
Figure (2.2): Lifecycle of Service Model (Stein, et al., 2009)
2.3 Patterns used in Service Composition
Barrett, et al. (2006) provided a subtle difference between two of the modelling aspect within the
service composition, namely workflow and distribution patterns. They considered workflows as
compositional services, whereby the internal and external messages to and from services are
modelled. In contrast, distribution patterns are considered composition where only the external
messages between services are modelled. This approach considers the basic four distributed
patterns: Centralized, Decentralized or peer-to-peer, Ring, Hierarchical.

Zirpins, et al. (2004) addressed using the patterns through the interaction composition. This
approach aimed to gather generic interaction patterns in a catalogue, this catalogue contains the
concrete generic interaction patterns, coordination policies and coordination idioms and their
relationships with the services to refines the service based on analysis of its interaction patterns.
3. Model and design
3.1 Services Composition Model Overview
Services Composition Model (SCM) aims to provide a general solution for the services
composition problem by realizing the requirements of a new service using the requirements of the
existing service and to automate the steps of the composition process; services discovery, services
selection, services invoking and services integration.
Although the SCM is not bounded to one particular algorithm, we generate an example to test the
SCM.
Figure (3.1) shows a brief description of the stages of the services composition model
Users: A user may compose some distributed services that may belong to different
administrative domains. Each user inserts the request of service using the service
composition language as an input language, the requirements of the requested service
may be satisfied by one of the following cases:
1. Matched service: there is one service that satisfies the user request, or
2. Composed services: the request could be satisfied by composing more than one service.
In this case we need the service composition algorithm to discover, select and compose
these services.
•Requirements Analysis: We introduce a new algorithm (Parsing Algorithm) to analyze the
inserted request from the user to specify the name of the services that satisfy the user’s request.
•Services Descriptions: The description of each service generated and stored in the database; this
database will be used to discover and select the desired service.
•Service Composition Algorithm (SCA): Services Composition Algorithm discovers and
selects the most suitable services to be integrated; this selection will be according to the inserted
requirements and composition criterion from the user, the SCA work in three steps:
34

1. Matchmaking: Once the requirements analyzed; the services composition algorithm discovers
the services by matchmaking between inserted requirements and the description of the available
services, this will generate abstract services.
2. Prepare for composition: Contains the entire abstract services that are chosen to suit the
consumer’s requirements; there are many services which may satisfy one requirement. This is
generally created automatically allowing the consumer to retrieve services from a repository then
according to the inserted composition’s criterion from the user, the composition algorithm will
select the actual services from the abstract services.
3. Make composition: According to the inserted composition’s criterion from the user, the
composition algorithm will select the actual services from the abstract services to be invoked and
integrated based on WSDL. This step will invoke and integrate the selected services from the
previous ste..
• Registered Services: Each provider can register his service by implementing the DLL and each
service contains the Web Services Description Language file (WSDL) as an indicator for the
services methods and functionalities.
In The following sections, we will discuss the stages of the Services Composition Model in
details.
3.2 Services Registry
Services registry is a searchable directory that contains a specification about the available
services from the provider. We suggest classifying these service’s specifications to be stored in
different categories in the centralized repository; each category contains all the services that
provide the same function.
35
3.2.1 Classified Repository pattern
In the following section we propose a new pattern to classify the registered services.
The structure of the classified repository pattern includes four participants as depicted in Figure
(3.2):
Figure (3.2): components of Classified Repository Pattern

36
3.2.2 DYNAMIC LINK LANGUAGE
We also suggest the idea of using the Dynamic Link Language (DLL) as an abstract interface. It
defines a unique service name, function name parameters and return value as follows:
One DLL will be implemented by all the services which provide the same function and stored in
the same category, for example the {S11, S12, S13 … S1n} are the services that stored in
(category 1) and provides the same function, these services were implemented from the same
DLL (DLL1) and stored in one layer as depicted in Figure (3.3).
Figure (3.3): DLL and related services.
3.3 Service Composition Language (SCL)
We propose a new Service Composition language (SCL) for the composition purpose, which
allows the user to express his service request in a simple way using the SCL notations.
Service Composition language is a simple text-based language. It absorbs most of the syntax from
VB.net and deals with the services that should be requested to be composed. The SCL is
considered a simpler composition language than the available composition BPEL language.
The main aim of the SCL is to automate the processes of the services selection and invoking.
BPEL is a powerful XML used in services composition, but the selection and the invoking of the
services in this language is static, so we generate the SCL which is able to select and invoke the
services dynamically.
3.4 Requirements Analysis
The inserted requirements (service composition program) will be analyzed by the parsing
algorithm. This algorithm is able to analyze the inserted program to generate the name of the
requested services and determine how the services will be composed. The pseudo code of the
parsing algorithm is depicted in Figure (3.4)

37
Figure (3.4): Parsing Algorithm Pseudo code
The parsing algorithm contains the following two major functions:
1.Parse Function parses each line of the inserted composition language and determines if the
instruction contains a variable, a return message, a service calling, or an if statement.
2. Resolve Variable trims all the external notations such that ($, $$, $$$) and invokes the Services
Composition algorithm.
3.5 Services Description
The services description is a description for registered service. Such a description would be
inserted in the database to allow the user to discover and select the desired services.
The description for each service is stated in Figure (3.5).
Figure (3.5): Attributes to describe the services.
This description of the registered web services will be as the follow:
• Information about services: Services ID and service URL.
• Information about functions (methods) in the service: method ID, method name, Criteria of the
method (budget cost, time cost).
• Information about the arguments in the method: argument ID, argument type.

3.6 Services Composition Algorithm
Services Composition Algorithm is the actual algorithm performing composition upon the user’s
request. The SCA composes the services thorough many steps as shown in Figure (3.6); (1)
matchmaking stage: the requirements are matched to the service’s description, (2) preparation
stage: the most suitable services from the matched services are selected relying on the
composition criterion and then prepared for the next stage known as composition, (3) composition
stage: the matched services are finally integrated and composed.
The Services Composition Algorithm depends on runtime discovery and selection for the
registered services
38
Figure (3.6): Composition process
3.6.1Matchmaking
Matchmaking is the process of services discovery of the registered services (Al-Masri, et al.,
2008); the description of the desired service and the description of the available services which
have to be compared. If the available services match, the required service is proved to be
successfully discovered.
The matchmaking step generates all the services that match the inserted requirements as abstract
services to prepare for composition; for example ,the abstract services for two requested services
are the following services {(S11, S12…S1n) and (S91, S92…S9n)}, these services generated
from the available services.
3.6.2 Prepare for Composition
This phase enables the selection of the services from the abstract services according to the
preferred composition criterion; the non functional requirements which inserted by the user (e.g.
cost) compared with the values of the abstract services to generate the concrete services to make
the composition, these services are really invoked and integrated e.g. {S11, S49, Snm}.

3.6.3 Services Composition Criteria
Multiple distributed services may provide similar functionalities with different non-functional
property values (e.g., different prices). To select the services that will be invoked, we suppose
many criteria for the composition.
In our example, we consider three generic quality criteria for composition process (1) execution
price, (2) execution duration and (3) reliability.
•Execution price (cost): the execution price is the amount of money that a service requester has to
pay for executing the operation.
•Execution duration: the execution duration measures the expected delay in seconds between the
moment when a request is sent and the moment when the results are received.
•Reliability is the probability that a request is correctly responded within a maximum expected
time frame (which is published in the service description) (Zeng, et al. 2003).
39
3.6.4 Make Composition
This step invokes and integrates the selected services from the previous step. The composition
algorithm will invoke the selected services based on the WSDL.
4. IMPLEMENTATION
We develop (6) services in our Application; Banking service, Police service, Average Calculation
Service, Grade Point Average (GPA) Service, File Search Service and Get File Attributes
Service. The first step is to develop these Web Services and deploy them. The actual code is
written in VB.Net classes.
The description and deployment of the Web Services are accomplished with WSDL documents.
1. Banking Service allows the client of the bank who has an account to interact with the service
to carry out certain transactions online, such a withdraw transaction is given in Figure (4.1).
Figure(4.1): Banking Service
2. Police service: this service examines if the credit card of the user is blocked or not. The
Police Service has one main function called IsBlackListed
IsBlackListed as shown in Figure (4.2).

40
Figure(4.2): Police Service
3. Average Calculate Service: calculate the average of the student marks using the following
formula:
Mean = sum of marks / number of marks
The Average Calculate Service has one function named with Calculate as shown in Figure (4.3)
Figure(4.3): Average calculate service
4.GPA Service calculates the Grade Point Average according to the shown standard
provided in Figure (4.4).
Figure (4.4): GPA Standard.
The GPA service has one function named with GetGPA as show n in Figure (4.5).

41
Figure (4.5): GPA Service
5. File Search service: search for file at specific location on the server based on name
and extension of the file (e.g. c:com.doc), the File Search service has one function
named with Search File as shown in Figure (4.6).
Figure (4.6): File Search Service
6. Get file attributes service retrieves the information about a specific file; this service retrieves
the file name, file extension, creation time and size of the file. The Get file attributes service has
one function known as GetFileAttributes as shown in Figure (4.7).
Figure (4.7): Get File Attributes service
4.1 Case Study1 (Payment Scenario)
The following payment example illustrates a possible application for Web Service Composition.
A customer consults a web page of online payment service.
The payment service has many business processes; such as (1) check the information of the
Credit Card then (2) check if the credit card is black (3) withdraw the inserted amount of the
money if the information is correct (4) or deny the process if there is any error in the inserted
information.

Payment service is basically composed of two services; the Banking Service and
IsBlackListedCreditCard Service as shown in Figure (4.8).
42
Figure (4.8): Payment Service.
1. The user writes a services request using the composition language, The user uses the (if)
statement to compose withdraws function with the is blacklisted service.
2. Parse and analyze the inserted program line by line to specify the requested services using
the parsing algorithm.
3. Discover all the services that matched the inserted one by matchmaking between the
requested services in composition program and the registered services, the discovery will
generate an abstract work flow contains all the payments and isblacklisted services (e.g.
payment1, payment2, islacklisted1).
4. Select the most suitable services from the generated workflow according the selected
criterion (e.g. payment1 and isblacklisted1). This step will generate the concrete workflow
using the GetSuitableClass function.
5. Invoke the selected services based on the WSDL using the WebClient Object; these services
will be composed.
6. Integrate the invoked services and return the result for the user.
4.2 Case Study (2)
Another scenario works with two similar Web Services representing composition process use the
while composition style.
The file discovery service Figure (4.9) has many services; (1) file search, this service will not stop
until it finds the requested file, (2) return the information about the found file.

43
Figure (4.9): File Discovery Service.
The user writes a services request using the composition language. The user used the (while)
statement to compose search file service with get file attributes service.
The file search service takes the name, location and extension of a file as input, then search for
the file at the specified location. This service will search for the file (while loop) the file is found.
After that it invokes the get file attributes to return the information about the file.
5. COMPARISON BETWEEN SCMVIA ANOTHER COMPOSITIONMODELS.
This section aims at providing a comparison between many of the presented models in the
related work chapter. This comparison is achieved by creating a table, listing many of the Models
and listing some of the most important features that were identified to categorize the composition
Models.
The comparison between the presented models will be according to the following categories:
• Composition strategy: Static and dynamic composition strategies concern the time when
web services are composed; the Static composition takes place during design-time, while
the dynamic composition takes place during run-time. Another composition strategy is
the declarative way. The declarative approach consists of two phases: the first phase takes
an initial situation and the desired goal as starting point and constructs generic plans to
reach the goal; the latter one chooses one generic plan, discovers appropriate services and
builds a workflow out of them.
• Execution monitor: The central scheduler controls the execution of the components of the
composite service.
• Dynamic Conversation selection: allows the user to select the conversation within a
service node at run-time, which may be useful when services are also discovered at run-time.
• QoS Modeling: service discovery in UDDI is not limited to the functional requirements
only, but it also considers the nonfunctional requirements QoS to discover and select the
services.
• WSDL Language: uses the Web Services Description Language to describe the services
functionalities.
• Graph support: Composite services are modeled by a graph defining the flow of service
invocations similar to UML activity diagrams.
• Service Composition Language is used by the user to express his requirements for
services composition.

44
Table (5.1): Comparison between Web Services Composition Models
Composition
Strategy
Execution
Monitor
Dynamic
conversation
selection
QoS
Modeling
WSDL
Language
Graph
Suppor
t
Service
Composition
Language
Categories
Models
SCM Dynamic
Composition
Yes Yes Yes Yes No Yes
PAMCS
Dynamic
Composition
Yes Yes No No No No
SELF-SERV
Declarative
Composition
No No No No Yes No
E-Flow
Dynamic
Composition
Yes Yes No No Yes No
FAWSS Dynamic
Composition
No No Yes Yes No No
6.1 Summary and Conclusion
Service composition enables flexible process of new application from reuse available services by
composing and integrating services.
This paper presents many results:
1. Develop a new model for the service composition problem.
2. Generate a composition language to express the user request and an algorithm (Parsing
Algorithm) which is able to analyze the composition program.
3. Generate a new composition algorithm; this algorithm able to discover, select and invoke the
services based on WSDL.
4. Generate a new architectural pattern (classified repository pattern) as suggestion to store the
registered services.
5. Applying and executing services composition process by providing many case studies as
examples; these cases based on the provided services composition approach.
6. Providing solutions for many of the difficulties in the services composition process.
7. Provide a formal description for the services registration process and for the composition
process using the Z-Language.
8. Explain the Services Composition Architecture with the Composition Process.
9. Comparison between the Business Process Execution Language (BPEL) and the provided
approach.
This research also discussed in details the stages of the services composition model:
1. Services registration: the providers implements the DLL for services registration process, the
registered services will be stored in the classified repository pattern.
2. Services description; Web Services Description Language (WSDL) and the description in the
database.

3. Services discovery: matchmaking between the inserted requirement and the description of the
services.
4. Services selection, matchmaking between the inserted criterion the most suitable services.
5. Services invoking based on the WSDL and the integration of these services.
REFERENCES
1. Sommerville, I. (2007), Software Engineering, 8th edition, united State of America, Pearson
45
Education Limited.
2. Guoping, Z. Huijuan, Z. Zhibin, W (2009), A QoS-based web services selection method for dynamic
web service composition, First International Workshop on Education Technology and Computer
Science, vol. 3, Wuhan, Hubei, China, 832-835.
3. Hudaib, A. and Montanero, C. (2002), A UML Profile to Support the Formal Presentation of Software
Architecture, Proceeding of the 26th IEEE international Computer Software and Application
conference on prolonging software life : Development and Redevelopment (COMPSAC'02), 217-
223.
4. Cooper, J.W. (1998), The Design Patterns Java Companion, (1st ed.), James W. Cooper.
5. Gamma, E. Helm, R. Johnson, R. and Vlissides, J. (1998), Design Patterns: Elements of Reusable
Object-Oriented Software, (1st ed.), Addison-Wesley, Baarn, Holland.
6. Tut, M.T and Edmond, D. (2002), The use of Patterns in Service Composition, Revised Papers from
the International Workshop on Web Services, E-Business and the Semantic Web, Vol. 2512, London,
UK, 28 - 40 .
7. Agarwl, V. Chafle, G. Mittal, S. and Srivastava, B. (2008), Understanding approaches for web service
composition and execution, Proceedings of the 1st Bangalore annual Compute conference, Article
No. 1 Bangalore, India.
8. Al-Masri, E. and Mahmoud, Q.H. (2008), Investigating Web Services on the World Wide Web,
Proceeding of the 17th international conference on World Wide Web, Beijing, China, April 21–25,
2008, 795-804.
9. Rao, J. and Su, X. (2005), A Survey of Automated Web Service Composition Methods. Proceedings
of the First International Workshop on Semantic Web Services and Web Process Composition, San
Diego, California, USA, Volume: 3387, 43-54.
10. Buschmann, F. Henney, K. and C.Schidt, D. (2007), pattern – Oriented Software Architecture on
Patterns and Pattern Language, Vol.5.
11. Gekas, J. and Fasli, M. (2005), Automatic Web Service Composition Using Web Connectivity
Analysis Techniques, W3C Workshop on Frameworks for Semantics in Web Services 2005 Position
Paper.
12. Keen, M. Acharya, A. Bishop, S. Hopkins, A. Milinski, S. Nott, C. Robinson, R. Adams, J.
Verschueren, P. (2004), Patterns: Implementing an SOA Using an Enterprise Service Bus SOA Steps,
(1st edition), IBM Corp.
13. Chmielowiec, A. Pierre, G. Gordijn, J. (2008), Technical challenges in market-driven automated
service provisioning, Proceedings of the 3rd workshop on Middleware for service oriented
computing, Leuven, Belgium, 25-30.
14. Stein, S. Payne, T.R. Jennings, N.R. (2009), Flexible provisioning of web service workflows,
Proceedings of the 17th European Conference on Artificial Intelligence, ACM Transactions on
Internet Technology, Volume 9, United Kingdom, 295–299.
15. Barrett, R. Pahl, C. M.Patcas, L. and Murphy, J. (2006), Model driven distribution pattern design for
dynamic web service compositions, Proceedings of the 6th international conference on Web
engineering, Palo Alto, California, USA, July 11-14, 2006, 129 – 136.
16. Zirpins, C. Lamersdorf, W. and Baier,T. (2004), Flexible Coordination of Service Interaction
Patterns, Proceedings of the 2nd international conference on Service oriented computing, New York,
USA, 49-56.

Software architectural patterns for

More Related Content

What's hot (10)

Viewers also liked (20)

Similar to Software architectural patterns for (20)

Recently uploaded (20)

Software architectural patterns for