Managing replication and transactions using neighbour replication on data grid

• Main Author: Noraziah, Ahmad
• Format: Thesis
• Language: English
• Published: 2007
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://umpir.ump.edu.my/id/eprint/1920/1/0004_NORAZIAH_BINTI_AHMAD.PDF; http://umpir.ump.edu.my/id/eprint/1920/

Replication is a useful technique for distributed database systems. Through this technique, a data object will be accessed (i.e., read and written) from multiple locations. Thus, it increases the data availability and accessibility to users despite site and communication failures. The all-data-to-all sites replication schemes such as Read-One- Write-All (ROW A) and Tree Quorum (TQ) are the popular techniques being used for replication and management of data in this domain. However, these techniques have its weaknesses in terms of data storage capacity and also data access times due to some number of sites must agree in common to execute certain transactions. In this study, the all-data-to-some sites scheme called Neighbour Replication on Grid (NRG) technique is proposed by considering only neighbours that have the replicated data. It is based on the logical structure of sites/servers in order to form a read or a write quorum in distributed database systems. The proposed technique considers only neighbours obtain a data copy. For simplicity, the neighbours are assigned with vote one and zero otherwise. The assignment provides a minimum communication cost with high system availability, due to the minimum number of quorum size required. In addition, it minimizes the storage capacity as well as data access time. A series of experiment was carried out by using three servers. Neighbour Replication on Grid (NRG) daemon is developed under Linux platform in the local area network (LAN) environment. It was carried out in Shell and Perl programming integrated with File Transfer Protocol (FTP) for the communications agent. The experimental results showed that the proposed model work successfully in managing replication and transaction when no failures occurred. Besides, the reconciliation and resolving conflict during system recovery are also supported when primary and neighbour replicas have failure. -Author