Towards Popularity Aware Hybrid Caching to Improve Search in Social Relationship Based P2p Networks

The widespread use of recent Peer-to-peer (P2P) file sharing has been mainly influenced by the scalability of their architecture and high versatile search mechanisms. However, most of the P2P networks construct loosely coupled overlay on the top of the internet based on the physical network constraints without taking user preferences or relationship into account. It leads to high inefficiency in their search algorithms, which mainly relies on the simple flooding or random walk strategies. In this paper, we present the architecture of adaptable fully decentralized social based P2P overlay as well as efficient cognitive community P2P search technique to improve the searching efficiency. This paper proposes a Dynamic Overlay Adaptation (DOA) algorithm, which creates social communities by connecting group of peers having similar interest within the shortest path based on the social relationship between them. The basic premise of the creation of social communities is that generated queries are most probably satisfied within its own community. Therefore, it significantly improves the searching efficiency with a higher success rate and less response time. In addition to, this paper proposes a Popularity aware Hybrid Caching (PaHC) which caches the file based on its size to improve cache performance. To accomplish high data availability, popularity aware strong data consistency is proposed that eliminates the duplicate content replication among the peers. Experimental evaluation reveals the effectiveness and efficiency of the proposed approach in terms of success rate, network traffic, response time and cache hit ratio.