Ian Clarke - A Distributed Decentralised Information Storage and Retrieval System.pdf

文档介绍：A Distributed Decen tralised Information Storage and
Retriev al System
Ian Clark e
Sup ervisor: Dr Chris Mellish
Division of Informatics
Univ ersit yofEdin burgh
1999
Abstract
This rep ort describ es an algorithm whic h if executed b y a group of in tercon-
nected no des will pro vide a robust k ey-indexed information storage and retriev al
system with no elemen t of cen tral con trol or administration. It allo ws informa-
tion to b e made a v ailable to a large group of p eople in a similar manner to the
W orld Wide W eb. Impro v emen ts o v er this existing system include:
- No cen tral con trol or administration required
-Anon ymous information publication and retriev al
- Dynamic duplication of p opular information
-T ransfer of information lo cation dep ending up on demand
There is also poten tial for this system to be used in a mo died form as an
information publication system within a anisation whic hma y wish to
utilise un used storage space whic h is distributed across anisation.
The system's reliabilit y is not guaran teed, nor is its eciency , ho w ev er the
in ten tion is that the eciency and reliabilit y will be sucien t to mak e the
system useful, and demonstrate that suc h a system is feasible.
This rep ort also describ es sev eral exp erimen ts designed to measure the eciency
and reliabilit yofsuc w ork. These are p erformed up on a sim ulation of a
w w ork written in the Ja v a programming language. Impro v emen ts
o v er the existing W orld Wide W eb are also highligh ted in this rep ort. The
algorithm is considered a protot yp e, and areas for further study and p oten tial
impro v emen t are highligh ted throughout this rep ort.
Con ten ts
1 In tro duction 4
F reedom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
A distributed decen tralised In . . . . . . . . . . . . . . . . 5
2 Aims 6
3 Related W ork 7
The Domain Name System (DNS) . . . . . . . . . . . . .