In P2P (peer-to-peer) systems, file replication technology is widely used to reduce hot spots and achieve high availability. One approach is to create replicas on the content-provider side. However, this type of methods rely its implementation heavily on the low-layer DHTs, they also generate a heavy query load on the master replica nodes. Furthermore, some of them do not perform well when dealing with churn due to a high updating overhead. This paper proposes a General Replication (GenRe) scheme for structured P2P systems. GenRe is built on the abstraction of DHTs and consequently can be applied to any implementations of DHTs. GenRe chooses replica nodes by performing several bitwise exclusive operations and all replica nodes can directly reply to queries, achieving load-balancing in both data placing and data queries. For any given data object, all the replicas are updated through a virtual binary tree, achieving a high efficiency and scalability. Moreover, GenRe has low overhead when dealing with churn since the replicas that a node stores can be moved in batches when the node joins or leaves the system. Our theoretical analysis shows the high performance of GenRe and simulation results demonstrate the efficiency and effectiveness of GenRe. The latency of updating is a logarithm function of the number of replicas. Our experiments also show a high data availability achieved under GenRe.
展开▼
