On optimal decision for QoS-aware composite service selection

https://doi.org/10.1016/j.eswa.2009.05.070Get rights and content

Abstract

The increasing popularity of employing web services for distributed systems contributes to the significance of service discovery. However, duplicated and similar functional features existing among services require service consumers to include additional aspects to evaluate the services. Generally, the service consumers would have different view on the quality of service (QoS) of service attributes. How to select the best composite service in theory among available service (WS) candidates for consumers is an interesting practical issue. This work proposes a QoS-aware service selection model based on fuzzy linear programming (FLP) technologies, in order to identify their dissimilarity on service alternatives, assist service consumers in selecting most suitable services with consideration of their expectations and preferences. This approach can obtain the optimal solution of consensual weight of QoS attribute and fuzzy positive ideal solution (FPIS) by extending LINMAP method, developed by Srinivasan and Shocker. Finally, two numerical examples are given to demonstrate the process of QoS-aware web service selection. The experimental results demonstrated that it is a feasible and supplementary manner in selecting the of web services.

Introduction

Web services enable business applications running on distinct platforms and exchanging data over the Internet, to be applied in business and daily life regardless of the platforms or locations. It has created unprecedented opportunities for organizations to shorten software development time by composing existing services across Internet. Effective mechanisms for supporting service discovery have considerable contribution to the success of web service composition. An efficient web service can bring a serious competitive advantage to the service providers as well as carry social welfare to the consumers. An application assisting in service selection based on certified QoS, cost and trust can bring essential benefits to the service consumers. Practically, the service providers are supposed to guarantee QoS of WS, which are advertised on the Internet for service consumers. When service providers announce their available services, current advertising approaches of web services create a WSDL or OWL-S document to subscribe the web service profile and service grounding, then promote it through UDDI registration, or other web services registries such as ebXML.

For emerging e-commerce business, the selected services are aggregated to form composite services. The composite service is a service produced by a composition of other services to complete the desired service activities (Anane, Chao, & Li, 2005). For example, Google research application are accepted as a web service and integrated with other services, such as Gmail, AdWords, You Tube and Google Maps service, to provide an integrated environment for service consumers. Microsoft and Yahoo also provide the services analogous to that of Google for business competition. The other example, consumer likes to discover the composite service, such as flight booking, restaurant reservation, and rent a car at a time, as illustrated in Fig. 1. What is the optimal approach of linking each service request to an approximate service? This problem may be nontrivial if the user requests multiple services at one time.

A number of works on composite service discovery and selection have been carried out to locate the required services and compose them to meet requirements using ontology (Zhou et al., 2004, Zhou et al., 2005) or service matchmaking techniques (Chao et al., 2005, Huang et al., 2005a, Huang et al., 2005b). Ontology technology is developed to answer the semantic confusion problem which could be effectively solved by semantic registration and discovery, by defining the appropriate meaning of the service’s functionality. Part of researches (Ankolenkar et al., 2002, Borenstein and Fox, 2003, Jorge and Amit, 2006, Zhou et al., 2005) on semantic service discovery were investigated via Semantics Web Service (SWS) technologies to locate the required services and compose them to meet requirements, as illustrated in Fig. 2.

The other approaches focus on fuzzy matchmaking technique that remains fuzzy semantics on terms and handle this problem via fuzzy theory. For example, moderated fuzzy discovery method (Huang et al., 2005a, Huang et al., 2005b), measures the similarity between services in terms of capability, syntax and semantics through a moderator initiates to minimize the differences among service consumers and providers.

For the consumer consensus of WS selection issue, service consumers and providers may have different expectations, experiences, and preferences about the services. Furthermore, consumer preferences often remain imprecise, uncertain or ambiguous on service QoS terms; the preferences over the QoS attributes are hard to be quantified especially in distinguishing the importance among these service attributes. Therefore, the adoption of fuzzy terms such as reasonable price, reliable service, and comfortable feeling in the requests becomes inevitable. Moreover, consumers usually have distinct view with providers for service terms, such as “cheap flight ticket”, “comfortable leg-room” or “delicious food”, simply because they have divergent perception of these terms.

From the consumers’ point of view WS providers usually advertise on the Internet exaggerating the features of web services for appealing to customers, which might lead to misunderstanding or confusing about some service terms for WS consumers. In addition, the providers prefer to advertise their services to customers in subjective terms, which might be short of considering the consumers’ expectations and preferences.

Hence, it is imperative to reach the consensus for service consumers on the specific specification terms (i.e., QoS), where they find and search WSDL document in the service discovery process. Based on these requirements, W3C working group has defined various QoS attributes for WS. That document comprises a number of generic and specific items for cross-referencing between the possible needs of service consumers and the functions supported by web services. Although regular QoS attributes have been listed, some unclear problems are yet to be clarified on selection of WS processes. For example, QoS attributes perception of importance is generally different from consumers and providers preferences. It is widely accepted that the consumers have been taking an active role in the expansion of e-commerce.

In this paper we consider optimal service selection based on a given set of service requests interacted with a set of service candidates using fuzzy linear programming (FLP) model (Li & Yang, 2004). This investigation leads to a need of developing a group consensus-centric approach to investigate QoS attribute preferences and determine the ranking order of service alternatives according to the distance from the positive ideal solution under group consensus. Consequently, service consumer is able to reduce redundancy in search, and service provider can improve the quality of services.

The remainder of the paper is organized as follows. Section 2 describes the existing QoS-aware selection of web service methods. Section 3 describes proposed method. Section 4 reports on two illustrational examples of selection of service alternatives. Finally, Section 5 illustrates the conclusion and the future work.

Section snippets

Available solutions for web service selection

A number of studies for web service selection have been carried out. One of the most well known techniques is “matchmaking”. It is employed in the situation where services with semantic descriptions for their functional attributes are available in the Internet search system. How to ensure the QoS of web services for service consumers. Ran (2003) proposed a new QoS-based service registration and discovery model to explore the possibility of QoS involving into UDDI registry information. In this

An QoS-aware services selection model based on LINMAP

In this section, we introduce a new fuzzy group consensus-aware service selection model, which extends LINMAP (LINear programming techniques for Multidimensional Analysis of Preference) method, developed by Srinivasan and Shocker (1973). In the LINMAP, decision maker gives the performance ratings matrix of alternatives with a pair wise comparison form to obtain the best solution that has the shortest distance to positive ideal solution (PIS) (Hwang & Yoon, 1981). The aim of the model is to find

Cases study

In this section, two illustrational examples for selecting an appropriate web service are used as the application of the proposed model. To examine the process of solutions respectively, the former example is a case which emphases the selection of composite service alternatives using the traditional LINMAP method, wherein QoS attributes are crisp data assessed by a single decision maker; whereas, the latter example is regarded as a group decision problem where rating format is fuzzy form given

Discussion

Even if the optimal solution is obtained, a significant issue of LINMAP is remained for debate, i.e., consistent solution. From the solution process of two cases, one could note that (i) LINMAP method need not require the complete paired comparison, it can gain a transitive ranking order for service alternatives, after computed, (ii) when the number of service alternatives (i) exceeds the number of attributes (j), then it is easy to yield a reliable solution of weight by the LIMAP, for example,

Conclusion

This paper presents a fuzzy group decision model to solve the selection of QoS-aware web services provisioning using a fuzzy linear program. The proposed model has the following features.

  • 1.

    This proposed approach not only deals with the decision maker’s imprecise perceptions under incomplete information, but also objectively determines the importance weights of QoS criteria. The weightings are based on group preferences for a group of participants and realistically attain a QoS-based ranking of a

Acknowledgements

This research is partially supported by Grant No. NSC 97-2410-H-168-013 and TWISC@NCKU Project: NSC 98-2219-E-006-001 from the National Science Council of the Republic of China.

References (25)

  • Huang, C.-L., Lo, C.-C., et al. (2005a). Service discovery through multi-agent consensus. In Proceedings of ieee...
  • C.-L. Huang et al.

    A moderated fuzzy matchmaking for web services

  • Cited by (0)

    View full text