A configurable matchmaking framework for electronic marketplaces

https://doi.org/10.1016/j.elerap.2003.09.003Get rights and content

Abstract

E-marketplaces constitute a major enabler of B2B and B2C e-commerce activities. This paper proposes a framework for one of the central activities of e-marketplaces: matchmaking of trading intentions lodged by market participants. The framework identifies a core set of concepts and functions that are common to all types of marketplaces and can serve as the basis for describing the distinct styles of matchmaking employed within various market mechanisms. A prototype implementation of the framework based on Web services technology is presented, illustrating its ability to be dynamically configured to meet specific market needs and its potential to serve as a foundation for more fully fledged e-marketplace frameworks.

Introduction

The vast connectivity provided by the Internet is constantly being exploited by entities (organisations and individuals alike) in order to locate other entities with complementary intentions. This is especially the case in the e-commerce arena, where buyers search for sellers (and vice versa), service providers for service consumers, lenders for borrowers, barterers for other barterers, etc. In particular, electronic marketplaces deployed over the Internet provide mechanisms for traders to directly or indirectly encounter other traders with complementary intentions [16].

Collecting and matching complementary intentions is a central function of any e-marketplace. Accordingly, matchmaking servers or matchmaking modules within servers offering other functionality, are at the core of existing e-marketplaces. For instance, business-to-business e-marketplaces such as ChemConnect.com provide matchmaking servers where sellers’ offers are matched against buyers’ criteria. Similar services are provided by Electronic Crossing Networks (ECN) such as Island and Instinet, by bartering sites such as IntelliBarter.com, and even beyond the realm of e-marketplaces, by career sites such as Monster.com.

There are many ways in which a trader can discover or can be discovered by other traders with complementary intentions by means of a matchmaking server. For example, a trader can simply advertise an intention, and wait until other traders with complementary intentions retrieve it and contact her directly in order to make an offer or start a negotiation. Conversely, a trader can query an existing repository of intentions in order to retrieve complementary intentions, and then contact the traders having lodged these intentions. Yet another way is for a trader to subscribe to a matchmaking server in order to receive intentions broadcasted by other traders, or conversely, a trader can broadcast his/her intention so that it is received by all the traders which have previously registered a potentially complementary intention. In some cases, trading intentions lodged in a matchmaking server are binding, meaning that if the matchmaking server finds a match with another trading intention, a trade can occur immediately without the need for the involved traders to contact each other. This is the case, for example, in a continuous double auction (CDA) [5].

This paper presents a framework designed to facilitate the development and maintenance of matchmaking servers for e-marketplaces. The approach adopted to design this framework has been to identify concepts and functionality common to various types of matchmaking servers, and to define a reference model capturing these commonalities. This reference model consists of a set of entity types and relationships that a developer can specialize and instantiate to fit the matchmaking modalities of a given e-marketplace. The reference model is complemented by a processing model which defines the sequence of steps that a matchmaking server follows when it receives a trading intention.

The rest of the paper is organised as follows. Section 2 describes the proposed reference model for configurable e-marketplaces. Next, Section 3 discusses how trading intentions are represented and processed. Section 4 discusses an implementation of the proposed framework using Web services technologies. Section 5 compares the framework with respect to similar efforts, and Section 6 concludes.

Section snippets

Types of trading intentions

The proposed configurable matchmaking framework is based on a classification of trading intentions according to three characteristics:

  • Notification mode. Whenever an incoming intention X is received by the matchmaking server and matched with another existing trading intention Y, who should be notified that a match has occurred? The lodger of the more recent intention (i.e. the incoming intention X)? The lodger of the existing intention (i.e. intention Y)? or both? The answer to this question can

Schema and syntax of trading intentions

In addition to the generic properties of a trading intention discussed above, a trading intention has domain-specific properties. These domain-specific properties vary from one trading domain to another. Specifically, the structure of the intentions that can be lodged in a given trading domain is determined by the trading intention schema associated with this domain. A trading intention schema defines the properties characterising a trading intention. Trading intentions schemas are expressed in

Implementation

We have implemented a prototype of the proposed matchmaking framework. The prototype is coded in Java and uses Web services technologies as the communication infrastructure between the actors and the matchmaking server. E-marketplace participants (i.e. actors) are implemented as software agents with a form-based user interface. The matchmaking server and the trading agents run as Web services on top of an SOAP-enabled application server, namely Apache AXIS.3 The

E-marketplace software frameworks

There are several existing frameworks for e-marketplace development. For example, the Michigan AuctionBot [22] implements several types of auction-based market mechanisms in a single parameterised platform. A given market mechanism is enabled by providing parameter values to the platform, and choosing one among a predefined set of auction types. Unlike our proposal, the AuctionBot does not support non-binding trading intentions, i.e. it does not support marketplaces in which agents place

Conclusion

The proposed matchmaking framework provides a promising alternative to developing matchmaking servers for e-marketplaces from scratch. Its major strength is that it is based on a classification of trading intentions according to three orthogonal characteristics, leading to 18 different types of intentions. These types of intentions can be associated with various roles within a trading domain, thereby allowing a wide spectrum of trading modalities to be implemented just by identifying the

Acknowledgements

This work was supported by the Australian Research Council SPIRT Grant “Self-describing transactions operating in a large, open, heterogeneous and distributed environment” involving QUT, UNSW, and GBST Holdings Pty Ltd. The work of the fourth author was conducted while at the Queensland University of Technology. The authors wish to thank the reviewers of the ECRA journal for their highly valuable comments.

References (22)

  • J.K. Lee et al.

    The eXtensible Rule Markup Language

    Communications of the ACM

    (2003)
  • B. Rachlevsky Reich et al.

    GEM: A Global Electronic Market System

    Information Systems

    (1999)
  • P.R. Wurman et al.

    A parametrization of the auction design space

    Games and Economic Behavior

    (2001)
  • Rule Markup Languages (RuleML)....
  • A. Chavez et al.

    A real life experiment in creating an agent marketplace

  • R. Cover, Cover Pages: Extensible Markup Languages,...
  • T. Di Noia et al.

    A system for principled matchmaking in an electronic marketplace

  • E. Gamma et al.

    Design Patterns: Elements of Reusable Object Oriented Software

    (1995)
  • T. Halpin

    Information Modeling and Relational Databases

    (2001)
  • J.K. Lee, J.Y. Kang, M.M. Sohn. XRML Web Page....
  • Cited by (16)

    • Semantic service matchmaking for Digital Health Ecosystems

      2011, Knowledge-Based Systems
      Citation Excerpt :

      Hence, the performance of service discovery and service matchmaking can be enhanced by semanticizing the web service descriptions. Service matchmaking refers to the process of matching service requests with similar service advertisements in the web service environment [15,26]. Semantic web technologies are used because the aforementioned web service description languages can specify the operations and data structure; however, they cannot specify the meaning and semantic constraints of data involved in web services, which may generate problems during the matchmaking process [30].

    • Multi-unit differential auction-barter model for electronic marketplaces

      2011, Electronic Commerce Research and Applications
    • Process-context aware matchmaking for web service composition

      2008, Journal of Network and Computer Applications
    • WS-contract establishment with QOS:

      2008, International Journal of Cooperative Information Systems
    • Game Theoretic Approach for Selection of Retailers in B2B E-Commerce

      2018, 2nd International Conference on Computational Systems and Information Technology for Sustainable Solutions, CSITSS 2017
    View all citing articles on Scopus
    View full text