Scalable Methods for Spam Protection in Decentralized Peer-to-Peer Networks

Abstract: Decentralized peer-to-peer networks offer several benefits over the today more commonly employed centralized client-server networks. Peer-to-peer networks allow for seamless distribution of web content over all participating nodes. This makes the hosting and access of web content much more fault tolerant, secure, faster and cheaper. However, today's peer-to-peer networks suffer from the inability to effectively protect against spam. This has led to the web becoming increasingly centralized instead. Methods to protect against spam on small peer-to-peer networks already exist. However, they all fail to scale up to large networks with a large number of participating peers. In this master thesis work, a new spam protection method based on design structures known to scale efficiently has been created. In order to evaluate how well it protects against spam, it is through simulations compared to what currently is the state of the art in that regard. The proposed method is found to provide noticable spam protection and scale up with ease. Although the method still lags behind the less scalable methods in terms of effectiveness in keeping spam out, in certain cases it is the only realistic option to implement. This master thesis work takes one additional step towards enabling the web to become more decentralized - meeting the requirements of the future.