Consistency Control and Recovery with Common Object Programmer Library Access (COPLA) Architecture

Most of the companies distribute theirinformation per branch, ensuring that the data aremainly updated locally, by the branch where theywere created. Examples of such companies couldbe hypermarkets and banks. So, it seems advisableto replicate all data objects, independent of thecompany branch where they have been created. Ifdata are replicated, “remote” accesses can beaccomplished locally, improving access times aswell as availability, in case of node failures in thenetwork. However, each time a data object ismodified, updates have to be multicast to a givennumber of database replicas.Data replicationamong different sites is viewed as a way toincrease application performance and its dataavailability.This system proposes a concurrencycontrol and recovery in a middleware architecturecalled COPLA (Common Object ProgrammerLibrary Access). This architecture providespersistent object state replication. This system isbased on time-stamp, it is an adaptation of theOptimistic protocol to this architecture. Therecovery process of this system allowsapplications to continue (or start) executingtransactions at all nodes, even in the node beingrecovered