Dynamic and Distributed Reconciliation in P2P-DHT Networks

摘要: Optimistic replication can provide high data availability for collaborative applications in large scale distributed systems (grid, P2P, and mobile systems). However, if reconciliation is performed by a single node, remains an important issue since the reconciler node fail. Thus, should also be replicated. We have previously proposed DSR-cluster algorithm, version of IceCube semantic engine designed cluster networks. However not suitable P2P networks, which are usually built on top Internet. In this case, network costs must considered. The main contribution paper DSR-P2P algorithm first propose P2P-DHT cost model computing communication DHT overlay network. Second, taking into account model, we each step. Third, that dynamically selects best nodes Our yields with acceptable performance limited overhead.