COMMUNITY-BASED SERVICE LOCATION.(Industry Trend or Event)

HOW VIRTUAL COMMUNITIES AND SOFTWARE AGENTS CAN BE USED TO PROVIDE AND LOCATE TRUSTWORTHY SERVICES ON THE INTERNET.

Although both computing and communications capabilities are improving exponentially, there has been a more rapid increase in available bandwidth than in computing power over recent years [6]. Naturally enough, while the network-related improvements create new possibilities for applications, the improvements and especially their differential, render obsolete traditional ways of thinking about telecommunications and computing.

There is an ongoing debate occurring among two main factions within the telecommunications industry [1, 4]. On one side are traditional telephone companies, who advocate intelligent networks with increased intelligence embedded within the network and controlled by the telecommunications providers. The intelligence would be reflected in features such as for caller identification or call forwarding. On the other side are the proponents of (what may be called) stupid networks, who take their inspiration from the Internet. In this view, as bandwidth becomes plentiful, intelligence will propagate to the edges of the network and the network itself will provide no more than bit transport. The overwhelming advantage of stupid networks is that, like the Internet, they naturally support heterogeneity and extensibility. End users can choose whichever applications they like and invoke whichever services they prefer without requiring consistent changes throughout a large network.

Consider telephone directories. White pages and yellow pages are essential for making a telecommunications system practical, but are classical centralized functionalities. Internet portals are a close analogue of telephone directories, and provide a one-size-fits-all solution to users' information needs. However, just as advances in computing are driving manufacturing from mass production to mass customization, advances in communication are driving information access from large portals toward personal contacts. Large portals won't be eliminated, but will increasingly be supplemented and superseded by personalized sources of information as people increasingly want to receive information and advice from those whom they know and trust.

Accordingly, in this article we consider the problem of service location. We study an approach that places the intelligence on the endpoints, enabling the users to locate desirable services based on trustworthy, personalized recommendations of their peers. The task is not only to locate a particular service, but also to locate a service that is rated highly by one's friends and associates, and their acquaintances.

Open environments are characterized by having components that are autonomous (acting independently) and heterogeneous (designed independently), of dynamic membership (joining, changing, and leaving arbitrarily), and of large scale (numerous). These properties are most compatible with the stupid network doctrine described previously, and are best supported by an approach such as ours that is based on a peer-to-peer model of interaction among intelligent entities existing at the edge of the network.

Communities and Agents

Let's review some key concepts and challenges. We define an online community as a set of interacting members or principals. The principals could be people, businesses, or other organizations. The members of a community provide services as well as referrals for services to each other. Our notion of services is general in that they need not be business services provided for a fee, but may be volunteer services, or not even services in the traditional sense, for example, just companionship or lively discussion about some topic. By the same token, quality of service includes not only the quality of the basic service but also of ancillary features, such as privacy. Thus a dentist who does a good job on your root canal, but tells everyone about it may be treated as offering a …