A Publish/Subscribe Framework: Push Technology in E-Commerce

  • Authors:
  • E. C. Cheng;George Loizou

  • Affiliations:
  • -;-

  • Venue:
  • BNCOD 17 Proceedings of the 17th British National Conferenc on Databases: Advances in Databases
  • Year:
  • 2000

Quantified Score

Hi-index 0.00

Visualization

Abstract

Publish/Subscribe (P/S) technology conceptually divides resource managers (RM's) and applications within a transaction tree into two categories: resource producers and resource consumers. Publishers enqueue information and the P/S system pushes it to the subscribers. Applying the P/S paradigm to Internet-based E-Commerce (sometimes also known as Webcasting), the RM's that produce and consume information can be real people. In this case, customers subscribe voluntarily or unknowingly to provide merchants their own information. Based on these customer profiles which may include information about their household or buying patterns, companies make decisions to push products and marketing messages to the relevant prospects electronically. Up until now, the Internet P/S mechanism was managed in a rather ad-hoc manner. Although in most cases the customer profiles are stored in RDBMS's, other objects such as products, promotion plans, web pages, news items, marketing messages, and service contracts are not modelled formally. As a result, it requires excessive human intervention to match the subscriber-base; it thus fails to support very high volume and flexible push of relevant and just-in-time information. In this paper, an object-oriented organizational model, called OMM, is presented as an underlying model to support two aspects of P/S in E-Commerce. Firstly, it allows companies to flexibly model both the consumers and the materials to be consumed. Secondly, it allows users to specify business policies in SQL-like queries to match the consumers with the materials. Every time a new material is published, OMM automatically pushes it to the right audiences. The paper describes the OMM reference model and shows how it can be applied flexibly to capture the different classes of resources involved in an E-Commerce application, and to maintain the complex and dynamic matching relationships between consumers and resource materials. The paper also discusses our experience of applying the research prototype, OMM/PS, to support the E-Commerce strategy of a commercial insurance firm.