Diploma Thesis, TU-Vienna, 2013

Nowadays, all popular online social networks in the WorldWideWeb are based on a client-server architecture. This leads to severe privacy issues as the provider of the online social network holds all data and could easily misuse it. Other contributions already showed how to resolve this problem by decentralizing the data with a peer-to-peer architecture. Yet these solutions have their own drawbacks, e.g. many complex technologies, bad extensibility or a lack of important privacy and security features.

 

A cause for these drawbacks is that no satisfying high-level peer-to-peer frameworks exist, allowing users to develop complex peer-to-peer applications in an easy way. Available frameworks are rather low-level, concentrating on the communication layer. Therefore, all of the peer-to-peer online social networks analysed have been created more or less from scratch.

 

To overcome this issue, we introduce the XVSM Micro-Room Framework which focuses on ensuring those aspects, not covered sufficiently by the existing peer-to-peer online social networks. This framework allows to create peer-to-peer applications based on configurable, shared data rooms. We then use the XVSM Micro-Room Framework for creating a proof of concept peer-to-peer online social network settled in an e-learning environment.

 

By comparing our solutions (both the XVSM Micro-Room Framework and the generated peer-to-peer online social network) with the analysed related work, we demonstrate the benefits of our solutions regarding functionality, simplicity, extensibility, privacy and security. Finally, several performance benchmarks show the practicability of the XVSM Micro-Room Framework even in heavy-load scenarios.