| dc.description.abstract | To be able to access our files at any time and any where, we need a system or service which is free, has enough storage space and is secure. A centralized system can handle these challenges today, but does not have transparency, openness and scalability like a peer to peer network has.
A hybrid system with characteristics from both distributed and centralized topologies is the ideal choice. In this paper I have gone through the basic theory of network topology, protocols and security and explained “search engine”, “Middleware”, “Distributed Hash Table” and the JXTA protocol. I then have briefly examined three existing peer to peer architectures which are “Efficient and Secure Information Sharing in Distributed, collaborative Environments” based on Sandbox and transitive delegation from 1999, pStore: A Secure Peer–to-Peer backup System” based on versioning and file blocks from 2001 and iDIBS from 2006, which is an improved versions of the SourceForge project Distributed Internet Backup System (DIBS) using Luby Transform codes instead of Reed-Solomon codes for error correction when reconstructing data.
I have also looked into the security aspects related to using distributed systems for resource discovery and I have suggested a design of a resource discovery architecture which will use JXTA for backup of personal data using Super-peer nodes in a peer to peer architecture. | en |
