Semantic Web: A Study on Web Service Composition Approaches

A Web service is a collection of open protocols and standards used for exchanging data between applications and systems. Web Service discovery, selection and composition are the important tasks of the any automated business processes. In a web service composition set of web services are collectively executed to achieve the objectives. This study reveals the existing approaches used for web service composition in both syntactic and semantic environments. We have reviewed more than 40 articles in this domain and concluded with merits and demerits of the methodologies applied for the implementation of the web service composition.


I. INTRODUCTION
Web service is a component can be remotely identified by the URL and they are heterogeneous in nature [1]. These components are described and accessed by using the various standards XML, SOAP, WSDL and UDDI [2]. Web services are constructed based on the Service Oriented Architecture (SOA), which is depicted in the figure 1 and are distributed across the web to service numerous business applications. Implementation of business applications with the help of web services includes various operations such as service request, request preprocessing, service matching, service discovery, service selection, service composition.
These operations can be done in many ways and various methodologies and approaches have been developed and employed. Considering and selecting those approaches for the specific environment is a tedious task because numbers of services are keep on increasing over the year. This article aims to simplify the searching process for the appropriate methodologies for the designated web service composition. We have surveyed more than 40 research articles and are listed with merits and demerits of it.
A web service composition problem can have six dimensions such as composition language, knowledge reuse, automation, tool support, execution platform, target users. It is also determining the work flow for an automated process [3].The selection of the approaches was a three-step process that involved preliminary selection, further selection, refinement.
The general framework of a web service composition has different components, one of them is process generator based on the external and internal specification of a service of a service consumer and service provider respectively. There are 12 platforms: eFlow, FormSys, Intalio BPMS, Self-Serv, SHOP2, Sword, Taverna, XL, Yahoo!Pipes, YAWL, jBPM, JOpera [8].  [20]. The criteria based web service selection methods are mostly used to achieve the objectives of domain, which can be classified as follows [14]  service-ranking approach  linear programming method  fuzzy model  QoS-based service selection method  hybrid models  ANP The entire article is organized as follows Section-I contains the introductory concepts. Section-II give the overview of the various approaches, merits and demerits of them and Section-III concludes article with the observations of the review.

II. OVERVIEW
Web service integration is an important task of any automation of business process. Web Service Discovery is the process of finding a suitable web service for a given task. The life cycle for semantic web services composition includes technologies used for composition approaches, modeling and service management [32]. Web Service Composition is the process of integrating related web services, is to provide set of composite services. Web service searching, selection and composition can be done in two ways: a) syntactic b) semantic [31] The various approaches for dynamic web service composition have been discussed. To meet the user's requirements regarding on-demand delivery of customized services, dynamic web service composition approaches have emerged. In a dynamic environment, realizing dynamic web service composition is not so easy. The evaluation shows that transactional support is still missing. The verification of the compositional correctness is also missing Self- This literature survey about web service composition, discovery and selection shows evaluation in the service oriented architecture domain. Proposed an approach to compose the web service based on the trustworthiness of it. From the basic observation a graph is formed which denotes the web service participated in the service composition. The basic notions used for representing the graph is such as Petri nets, oWF-nets, Labeled Transition Systems and Opacity of a Labeled Transition System [12]. Criteria based service selection methods were introduced and approach considers various parameters such as Quality of service, User Preference and Scalability [14].
Semantic web service discovery has so many challenges in-terms of technologies to be used such as OWL-S, WSMO,WSML,WSMX,SAWSDL and the tasks needs to be carried out by the discovery process such as publishing, mediation, storage, request, matchmaking, negotiation and selection [17] The privacy data are distributed across the web for example medical data. These data are to be shared by means of web services and its composition. This type of composition is disused in [15]. The level of privacy is determined by the rules defined in the composition architecture. Clustering Web services is very much important to facilitate service discovery. It can be done in many ways, one of them is based on the parameters used in the web service invocation [16].QOS-based web service composition based on Genetic Algorithm (GA) is introduced in [18] [21]. This GA based approach is considering QoS parameters. Business Process Execution Language (BPEL) based composition uses behavioral approach by verifying the concurrent properties [19].
Web service composition process can employ either manual or semi-automatic or automatic methodologies. Automatic web service composition is used when dynamic composition is required [22]. Firefly approach is the which provides optimal solutions for dynamic service composition [23] and another web service composition framework provides the solution for the composition problems on the fly that is dynamically called User-Centric WS-Mediator framework for on-the-fly Web Service Composition [24] [26].
An Optimization Method of Workflow-based Web Service Composition Model is a computational model, calculates QoS and then formulates the composition [25].The researchers, software developers, advanced practitioners, documentation writers, and users involved in Web services domain for QOS predictions exploring development opportunities especially in web databases.
Log-based mining techniques are used to identify the patterns of the web service access and composition. It's a reengineering approach for service composition and another easy web service discovery approach is a query-by-example approach [28] [29]. Well defined business process can employ the goal-based approach for Web service composition [34]. A linear approach for web service composition uses QoS parameters and provides transactional aware results [36] and context aware solution is useful for domain specific problem and yields better results [34].
A theorem proving framework for the formal verification of web services composition has explained and the methodologies can be used in post composition [38] [39].Optimization and Ranking in web service composition using Performance Index (PI) [42]. Huge number of services in the repository increases the processing time for selection, discovery and composition. The performance of the composition framework can be reduced by means of parallel processing [44].
A number of web services have been carried out on cloud computing, including performance analysis, market-oriented graph Semantic Web Services, management tool, workload balance, dynamic selection, etc. semantic I/O information of the services. presenting the nonfunctional characteristics of the software systems and services.
QoS of cloud services can be measured from either the client side (e.g., response time, throughput, etc.) or at the server side (e.g., price, availability, etc.).

A. QOS Perspective
Quality-of-service can be measured either at the server side or at the client side. Client-side QoS properties provide more accurate measurements of the user usage experience. The commonly used clientside QoS properties include response time, throughput, failure probability, etc. This paper mainly focuses on ranking prediction of client-side QoS properties, which likely have different values for different users (or user applications) of the same cloud service.

Definition and Execution of Composite Web Services
The SELF-SERV Project [40] The SELF-SERV project aims at providing tool support and middleware infrastructure for the definition and execution of composite Web services. SELF_SERV system: key aspects, Declarative service composition the number of services to be composed may be large and continually evolving highly distributed nature of services. In SELF-SERV, the process model is specified as a state chart, states and invocations to Web services transitions events, conditions, and variable assignment operations.
SELF-SERV exploits the concept of service community. Prototype system in which Web services are declaratively composed and the resulting composite services can be orchestrated either in a

11.
QoS approach.GA based algorithm has 2 constraints: 1) only one web service among candidate web services should be chosen for a task 2) the service composition must satisfy user constraints [18].

III. CONCLUSION
This review has an objective to give an overview about the recent trends in the development of various web service composition approaches. Initially we identified the existing approaches used for the composition. These approaches are classified according to the processing of the service descriptions, which can be either syntactic or semantic based service process. The review also comprises the methodologies used for composition and the parameters used for selection and matching processes. Every composition approaches have its own merits and demerits. Most of the syntactic approaches have used QoS parameters for composition and semantic approaches are domain specific and the implementation needs basic knowledge about the domain, semantic description, ontologies and composition engine tasks.