Neighbor Table Construction and Update for Resilient Hypercube Routing in P2P Networks

Several proposed peer-to-peer networks use hypercube routing for scalability. Consistency of neighbor tables in hypercube routing guarantees the existence of a path from any source node to any destination node. Such consistency, however, can be broken by node failures. To improve the robustness of hypercube routing, we first generalize the concept of consistency to K-consistency, for K ≥ 1, which is shown to provide at least K disjoint paths for any source-destination pair with a probability close to 1. Our next objective is to design and specify a new join protocol together with a proof that it generates K-consistent neighbor tables for an arbitrary number of concurrent joins. We first present a conceptual foundation, called C-set trees, for protocol design and reasoning about Kconsistency. We then present a detailed specification of a join protocol, and a rigorous proof of correctness for the join protocol. The crux of our proof is based upon induction on C-set trees. Both theoretical analysis and simulation results show that the join protocol is scalable to a large number of network nodes. Keywords—Peer-to-peer networks, consistency, K-consistency, hypercube routing, join protocol, protocol design and verification, C-set tree