Supporting the Web: A Distributed Hyperlink Database System

Abstract In our last paper[16], we presented an integrated scheme for an intelligent publishing environment that included a locally maintained hyperlink database. This paper takes our previous work full cycle by extending the scope of the hyperlink database to include the entire Web. While the notion of hyperlink databases has been around since the beginnings of hypertext, the Web provides the opportunity to experiment with the largest open distributed hypertext system. The addition of hyperlink databases to the Web infrastructure positively impacts several areas including: referential integrity, link maintenance, navigation and visualization. This paper presents an architecture and migration path for the deployment of a scalable hyperlink database server called Atlas. Atlas is designed to be scalable, autonomous, and weakly consistent. After introducing the concept and utility of link databases, this paper discusses the Atlas architecture and functionality. We conclude with a discussion of subscriber and publisher policies that exploit the underlying hyperlink infrastructure and intelligent publishing environments.