An analytical framework for the performance evaluation of proximity-aware structured overlays

In this paper, we present an analytical study of proximity-aware structured peer-to-peer networks under churn. We use a master-equation-based approach, which is used traditionally in non-equilibrium statistical mechanics to describe steady-state or transient phenomena. In earlier work we have demonstrated that this methodology is in fact also well suited to describing structured overlay networks under churn, by showing how we can accurately predict the average number of hops taken by a lookup, for any value of churn, for the Chord system. In this paper, we extend the analysis so as to also be able to predict lookup latency, given an average latency for the links in the network. Our results show that there exists a region in the parameter space of the model, depending on churn, the number of nodes, the maintenance rates and the delays in the network, when the network cannot function as a small world graph anymore, due to the farthest connections of a node always being wrong or dead. We also demonstrate how it is possible to analyse proximity neighbour selection or proximity route selection within this formalism