Elsevier

Decision Support Systems

Volume 43, Issue 1, February 2007, Pages 243-255
Decision Support Systems

Syndicating Web Services: A QoS and user-driven approach

https://doi.org/10.1016/j.dss.2006.09.011Get rights and content

Abstract

Web Services that provide a real-time e-business solution via online binding of software components must have a mechanism to win users' confidence in the quality of service (QoS) to establish a solid foot holding in the market. Inspired by the trust third-party approach in the public key system, we explored the idea of expanding the role of registrars to include (1) assessing quality of Web Services and (2) syndicating Web Services. The Analytic Hierarchy Process (AHP) and the Brown–Gibson (BG) methods were adapted to facilitate quality assessment. An optimization model was proposed for Web Service syndication. A heuristic algorithm was developed to solve the NP-hard problem, and an experiment was conducted, with two sensitivity analyses involving adjusting parameters, to compare its performance and the optimal solutions.

Introduction

Web Services have emerged as a viable e-business platform, providing a business software solution via real-time binding of software components over the Internet or corporate networks. This technology is facilitated by a set of standard protocols that allow: (1) a service provider to publish software components, (2) a registrar to serve as the repository of available services, and (3) a user to discover services suitable for the business.

To fully realize the potential, the Web Services infrastructure must have a mechanism to bestow users' confidence in the quality of the published software components. The issue is similar to the public key infrastructure, whose success hinges on the key authentication. Therefore, we propose that the role of the Web Services registrar be expanded to include Quality of Service (QoS) assessment and Web Services syndication. By syndication, we mean that a registrar with the capability of identifying a user's QoS preferences could make recommendations and/or provide a package of Web Services to the user. Such a process could turn a passive registrar into an active player. Some might suspect the impartiality of a registrar. We believe that the market self-selection process will keep the registrar in check, much the same as it does to firms providing and registering public keys.

However, QoS encompasses so many different dimensions including integrity of solution, speed, security and cost. It is difficult, if not impossible, to derive an absolute QoS measurement. This is compounded by the fact that the value of QoS is in the eyes of the beholder. Any QoS assessment effort would be to no avail if a user's QoS preferences were not taken into consideration.

In this study, we first adopt Saaty's Analytic Hierarchy Process (AHP) approach [27] for assessing quality of Web Services published to a registrar. The outcome of the AHP approach is a set of Eigenvalues representing the relative merits of available services. The AHP approach, however, involves time-consuming pair-wise comparisons. To alleviate the problem, we adapted the Brown–Gibson method (BG) [7] to further divide the QoS criteria into two categories: subjective and objective. While the subject criteria reflect a user's personal opinions or attitudes, the objective criteria are those can be measured in the monetary term. The BG approach provides an efficient way to derive the relative merits of services in terms of objective criteria. For the subjective criteria, we recommend using the AHP approach to obtain the relative merits so that evaluation biases can be avoided [31]. The BG approach also provides a way to synthesize the subjective and the objective evaluations into a composite index. We then incorporate the composite indices into a mathematical model, capturing the syndication efforts of a registrar. The resulting mathematical model is an NP-hard problem. A heuristic algorithm is developed, and an experiment is conducted to compare its performance to the optimal solutions. Optimal solutions are obtained through CPLEX because it is one of the industry leading software, and the heuristic algorithm is coded in C++ to take advantage of its object-oriented features. The experiment is also done with two sensitivity analyses that involve adjusting two parameters: variations in QoS and increases in the number of service providers.

This paper is organized in the following fashion. A brief literature review is provided in the next section, followed by the discussions of the AHP and the BG approaches. The mathematical model and the heuristic algorithm are then presented, along with the experiment and its results. Finally, the benefits and the limitations are discussed.

Section snippets

Literature review

Web Services have recently become a viable Internet-based distributed e-commerce platform to establish and maintain business relationships and product development. They support self-describing and modular software applications with open interface standards and communication protocols. They exceed previous distributed component technologies by offering a high level of interoperation between programs that are written in different languages and running on different operating systems [17]. Web

User driven Web Service syndication

We first assume that there are M providers collectively offering N services. An M by N incidence matrix, G, is used to represent services available from each provider in that gmn = 1 if provider m offers service n; otherwise, gmn = 0. We then assume that the QoS of a provider is measured along r quality dimensions including security, speed, information integrity, cost, and so on. To facilitate syndication, the registrar, using the AHP approach, assesses QoS of providers and stores the evaluation

Conclusions

Quality of Web Services has become a prominent issue when more and more services are offered by an ever-increasing number of service providers. Selecting and combining services require intimate knowledge about the offerings in the market and the expertise of evaluating the qualities of these offerings. A Web Service registrar is in the best position to provide such valuable services to its users. In this paper, we explored ways for a QoS and user-driven syndication of Web Services. The

Dr. Yi Sun is an Assistant Professor of Information Systems at the School of Business, California State University San Marcos. His degrees include BA degree from Foreign Affairs College in Beijing, China and PhD (Management Information Systems) from University of Florida. He has research interests in telecommunications, data mining and artificial intelligence, electronic commerce, and applied operations research.

References (37)

  • J.S. Dyer

    Remarks on the analytic hierarchy process

    Management Science

    (1990)
  • Expert Choice, http://www.expertchoice.com,...
  • X. Gao et al.

    The Capability Matching of Web Services

  • M.R. Garey et al.

    Computers and Intractability: A Guide to the Theory of NP—Completeness

    (1979)
  • M. Gilpin

    Who Has How Many Web Services?

    (2004)
  • K. Gottschalk et al.

    Introduction to Web Services Architecture

    IBM Systems Journal

    (2002)
  • P.T. Harker et al.

    Global effective questioning in the analytic hierarchy process

    European Journal of Operational Research

    (1990)
  • S. Kalepu et al.

    Verity: A QoS Metric for Selecting Web Services and Providers

  • Cited by (49)

    • Determining strategy of pricing for a web service with different QoS levels and reservation level constraint

      2015, Applied Mathematical Modelling
      Citation Excerpt :

      A QoS for the web service is usually evaluated using various criteria, such as reliability, availability, response time, and etc. When users intend to purchase, they consider not only the price but also the quality of that web service [1–4]. Furthermore, since the providers may sell the web service prior to the consumption period, the customers get this opportunity to cancel their orders.

    • A novel non-functional matchmaking approach between fuzzy user queries and real world web services based on rough sets

      2014, Future Generation Computer Systems
      Citation Excerpt :

      We review in what follows research initiatives related to non-functional matchmaking. Sun et al. [11] suggested to expand the role of the Web Services registrar to include Quality of Service (QoS) assessment and Web Services syndication.

    • Reliable Web service selection in choreographed environments

      2013, Decision Support Systems
      Citation Excerpt :

      Kulnarattana and Rongviriyapanish [18] instead propose a QoS model that uses uncertain attributes and a nonparametric test technique to test historical QoS information from the client side, then uses the QoS values as the Web service selection criteria. Sun et al. [32] assess subjective QoS value from pair-comparisons between candidate Web services and then combine users' preference, subjective and objective QoS values, with weights by using adapted Brown–Gibson method to derive aggregated QoS value of each candidate Web service. Given the QoS values of each individual Web service, most recent works focus on the selection of Web services as a means to satisfy the QoS requirements of a process (i.e., composite Web service).

    • Dynamic Web service selection group decision-making based on heterogeneous QoS models

      2012, Journal of China Universities of Posts and Telecommunications
    View all citing articles on Scopus

    1. Download : Download full-size image

    Dr. Yi Sun is an Assistant Professor of Information Systems at the School of Business, California State University San Marcos. His degrees include BA degree from Foreign Affairs College in Beijing, China and PhD (Management Information Systems) from University of Florida. He has research interests in telecommunications, data mining and artificial intelligence, electronic commerce, and applied operations research.

    1. Download : Download full-size image

    Dr. Shaoyi He is currently an Associate Professor of Information Systems in the School of Business Administration at California State University San Marcos in the United States. He received his PhD from the University of North Carolina at Chapel Hill in 1998. Since then, he has worked as a faculty member in Long Island University, Pennsylvania State University, and California State University San Bernardino. His current research interests include interplay of technology, culture and language in global e-business, multilingual information access and retrieval on the Web, multilingual issues in e-commerce website glocalization, and language barriers in marketing across cultures. He has published papers in such academic journals as Journal of the American Society for Information Science, Information Processing and Management, Journal of Information Communication and Library Science, Electronic Library, and Journal of China Society for Scientific and Technical Information.

    1. Download : Download full-size image

    Dr. Jack Leu is a Professor of Operations and Information Systems at California State University San Marcos. He has served as the department chair and the graduate program director. His teaching interests include Database, Networking, Programming, and Operations Management. His current research focuses on Web Services and Internet Gaming Strategy. He has published articles in journals such as Decision Sciences, IIE Transactions, International Journal of Production Research, and Annals of Operations Research.

    View full text