Collaborative Web Hosting: Challenges and Research Directions (SpringerBriefs in Computer Science)
Reaz Ahmed, Raouf Boutaba
Format: PDF / Kindle (mobi) / ePub
This brief presents a peer-to-peer (P2P) web-hosting infrastructure (named pWeb) that can transform networked, home-entertainment devices into lightweight collaborating Web servers for persistently storing and serving multimedia and web content. The issues addressed include ensuring content availability, Plexus routing and indexing, naming schemes, web ID, collaborative web search, network architecture and content indexing. In pWeb, user-generated voluminous multimedia content is proactively uploaded to a nearby network location (preferably within the same LAN or at least, within the same ISP) and a structured P2P mechanism ensures Internet accessibility by tracking the original content and its replicas. This new paradigm of information management strives to provide low or no-cost cloud storage and entices the end users to upload voluminous multimedia content to the cloud data centers. However, it leads to difficulties in privacy, network architecture and content availability. Concise and practical, this brief examines the benefits and pitfalls of the pWeb web-hosting infrastructure. It is designed for professionals and practitioners working on P2P and web management and is also a useful resource for advanced-level students studying networks or multimedia.
need to rank the search results in a distributed manner. This will reduce network and storage overheads by transmitting only the most important and relevant results. • Distributed incremental retrieval: Gradual retrieval of search results in chunks of 10 or 20 records is supported by the contemporary web search engines for performance reasons. For achieving incremental retrieval in a P2P network, the routing mechanism must be able to track previously returned results, and already queried peers.
with another peer, it updates the knowledge about its external world. Wang et al. used two types of ranks for overall ranking: Local PageRank is computed in each peer based on the standard Pagerank algorithm, and ServerRank is computed as the highest local PageRank or the sum of all the PageRanks of a web server . SiteRank  computes the rank at the granulaity level of websites instead of web page level using PageRank. Wu et al. proposed a layered Markov model for distributed ranking where
Collaborative Web Search Fig. 4.2 Indexing process in DEWS website Base URL hash codeword Plexus Routing Website index node Keywords DMP, n-gram Bloom-filter Pattern List decoding codewords Plexus Routing Inverted index nodes Vector Space Model (VSM) is used for computing the relevance between keyword kirep j and URL ui . In VSM, each URL ui is represented as a vector νi = (ri1 , . . ., rig ), where ri j represents the relevance of the term or keyword kirep j in ui , and g is the number of
Ahmed, M. Rahman, R. Boutaba, and B. Mathieu. Diurnal availability for peer-to-peer systems. In Proc. CCNC, Las Vegas, Nevada, USA, Jan 2012. 18. S. Shi, G. Yang, J. Yu, Y. Wu, and D. Wang. Improving availability of p2p storage systems. In Proc. APPT, 2003. 19. E. Sit, A. Haeberlen, F. Dabek, B. Chun, H. Weatherspoon, R. Morris, M. F. Kaashoek, and J. Kubiatowicz. Proactive replication for data durability. In Proc. IPTPS, 2006. 20. D. Stutzbach and R. Rejaie. Understanding churn in peer-to-peer
Web search services are governed by centrally controlled search engines, which is not healthy for our online freedom due to the following reasons. A Web search service provider can be compromised to evict certain websites from the search results, which can reduce the websites’ visibility. Relative ranking of websites in search results can be biased according to the service providers’ preference. Moreover, a service provider can record its users’ search history for targeted advertisements or