While replicating data over a decentralized Peer-to- Peer (P2P) network, transactions broadcasting updates arising from different peers run simultaneously so that a destination peer replica can be updated concurrently, that always causes transaction and data conflicts. Moreover, during data migration, connectivity interruption and network overload corrupt running transactions so that destination peers can experience duplicated data or improper data or missing data, hence replicas remain inconsistent. Different methodological approaches have been combined to solve these problems: the audit log technique to capture the changes made to data; the algorithmic method to design and analyse algorithms and the statistical method to analyse the performance of new algorithms and to design prediction models of the execution time based on other parameters. A Graphical User Interface software as prototype, have been designed with C #, to implement these new algorithms to obtain a database synchronizer-mediator. A stream of experiments, showed that the new algorithms were effective. So, the hypothesis according to which “The execution time of replication and reconciliation transactions totally depends on independent factors.” has been confirmed.
Eager replication of distributed databases over a decentralized Peer-to-Peer (P2P) network is often likely to generate unreliability because participants can be or cannot be available. Moreover, the conflict between transactions initiated by different peers to modify the same data is probable. These problems are responsible of perpetual transaction abortion. Thus, a new Four-Phase-Commit (4PC) protocol that allows transaction commitment with available peers and recovering unavailable peers when they become available again has been designed using the nested transactions and the distributed voting technique. After implementing the new algorithm with C#, experiments made it possible to analyse the performance which revealed that the new algorithm is efficient because in one second it can replicate a considerable number of records, such as when an important volume of data can be queued for subsequent recovery of the concerned slave peers when they become available again.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.