Exploiting Recommender Service to Enhance Efficiency of Replication

Modern, distributed systems are expected nowadays to provide high data availability, increased fault-tolerance, low bandwidth consumption, and improved scalability. To cope with the aforementioned problems, the replication mechanisms are often used. However, with the constantly growing scale of modern systems, providing the efficient replication becomes increasingly difficult. One of the ways to increase the replication efficiency is associated with the use of knowledge from social networks. This paper proposes a recommender service that exploits social relationships to provide a list of nodes that will hold replicated data. Such nodes are selected based on the value of the proposed matching degree metric. Our focus is also on evaluating the proposed solution.