Using Monte Carlo simulation for improving Data Availability in P2P network

In this paper we present a replication strategy to improve data availability in P2P Networks. The focus of the paper is to replicate data to nodes which are highly available and complement one another in terms of uptimes. This would decrease the replication management overhead until the number of replicas falls to a certain threshold. Replication to reliable node would improve the cost of replication by avoiding irregular nodes of the network. We run Monte Carlo simulation based on past traces of Kad, OverNet, Bittorrent and Planet Lab, to present how our replication keeps data sustainable in the network. In our evaluation we show that a life pattern along with the availability of nodes improves overall data availability. We perform our evaluation on a Kademlia network, and show that our approach reduces over head compared to existing approaches to data availability.