Abstract

Nowadays, we are witnessing an increasing growth of Web 2.0 content such as micronews, blogs and RSS feeds. This trend exemplified by applications like Twitter and LiveJournal is starting to slow down not only by the limitations of existing services – proprietary and centralized, but also by the cumbersome process of discovering and tracking interesting content. This content is generally ephemeral and thus difficult to index by conventional Web search technology. This problem is exacerbated by the passive role adopted by Web content providers: it is surprising that Web servers publish information and expect that thousands of other servers (search engines, Web-based aggregators like GoogleNews, etc.) advertise their content to the world. In this work we propose p2pWeb, an open, decentralized infrastructure to enable Web servers to use their spare capacity to filter out, aggregate and disseminate Web content in a scalable and timely manner. p2pWeb is flexible enough to support a broad variety of services. The main property of p2pWeb is that all communication abstractions, including aggregation and multicast, are implemented hierarchically and using the HTTP protocol. Simulation results certify the viability of our approach.

Links and resources

Tags