Distributed Indexing of Autonomous Internet Services.
Peter B. Danzig, Shih-Hao Li, Katia Obraczka · 1992
ABSTRACT: This paper describes the architecture and the design decisions behind a resource discovery tool that we prototyped to knit together the Internet's resource discovery fabric. \\ù/e call the architecture distributed indexing or Indie for short. Indie consists of a directory of services and an unlimited number of broker databases that index their own data, data stored in other brokers, and data available from other resource discovery services. The indexing mechanism doubles as alazily consistent data replication mechanism that can replicate the directory of services or any other broker at will. An Indie broker automatically clusters references to related objects stored in other autonomous discovery and database services. Since Indie brokers cluster related information skimmed from thousands of scattered services, efficient exhaustive search is possible. This centralization led to the success of the qrchie file location service. In a way, Indie is a generalized archie that locates autonomously maintained data stored in different discovery services. We believe that other discovery tools can benefit from the architectural principles that Indie illustrates and the @ Computing Systems, Vol. 5. No. 4. Fall t992 433ability to cluster related information that Indie provides. 1.