This paper proposes the use of a wide-area file system for storing and retrieving documents. We demonstrate that most of the functionality of the World-Wide Web (WWW) information service can be provided by storing documents in AFS. The approach addresses several performance problems experienced by WWW servers and clients, such as increased server and network load, network latency and inadequate security. In addition, the mechanism demonstrates the value of a global, general purpose file sharing system and the advantage of layering existing technologies.
We would like to thank Bob Sidebotham and Bill Camargo for contributing ideas and code. We are grateful to John Ockerbloom of Carnegie Mellon University for providing data from their World-Wide Web server.
This research was funded by the Advanced Research Project Agency, under contract number F19628-93-C-1076. The views and conclusions expressed in this paper are those of the authors and do not represent the official position of ARPA or Transarc Corporation.
Source code is available for the URL translation library and repository server code.