Förderverein Informationstechnik und Gesellschaft
Ian Clarke - May 23rd 1999, 20:44 EST
A pre-release of a report describing "FreeNet" has been made available to obtain comments from the Internet public. The report is entitled "A Distributed Decentralised Information Storage and Retrieval System". The system is designed to avoid having any form of centralised control or administration, allowing participating computers to act together in a form of organised anarchy. It will allow anybody to store information, and is designed to make censorship of any form difficult or impossible. The report is available here in downloadable postscript, and browsable HTML.
A distributed, decentralised information storage and retrieval system
Supervisor: Dr Chris Mellish
School of Artificial Intelligence, School of Computer Science, Division of Informatics, University of Edinburgh.
This paper describes an algorithm which if executed by a group of interconnected nodes will provide a robust key-indexed information storage and retrieval system with no element of central control or administration. It allows information to be made available to a large group of people in a similar manner to the ``World Wide Web''. The system's reliability is not guaranteed, nor is its efficiency, however the intention is that the efficiency and reliability will be sufficient to make the system useful, and demonstrate that such a system is feasible. The system is considered as a potential replacement for the ``World Wide Web''. Improvements over this existing system include:
- No central control or administration required
- Anonymous information publication and retrieval
- Dynamic duplication of popular information
- Transfer of information location depending upon demand
There is also potential for this system to be used in a modified form as an information publication system within a large organisation which may wish to utilise unused storage space which is distributed across the organisation.
This paper also describes several experiments designed to measure the efficiency and reliability of such a network. These are performed upon a simulation of a working network written in the Java programming language. Improvements over the existing ``World Wide Web'' are also highlighted in this paper. The algorithm is considered a prototype, and areas for further study and potential improvement are highlighted throughout this paper.
Contents 1 Introduction 2 Aims 3 Project structure 4 Related Work 5 Initial Ideas - A redundant hierarchy 6 Final Design - An adaptive network 7 A simple simulation 8 A full simulation 9 Experiments 10 Areas for further investigation 11 Conclusion 12 Acknowledgements References About this document ...
Ian Clarke Sun May 23 23:39:59 BST 1999