Data Replication with 2D Mesh Protocol for Data Grid

Data replication is one of the widely approach to achieve high data availability and fault tolerant of a system. Data replication in a large scale distributed and dynamic network such as grid has effects the efficiency of data accessing and data consistency. Therefore a mechanism that can maintain the consistency of the data and provide high data availability is needed. This thesis discusses protocols and strategies of replicating data in distributed database and grid environment where network and users are dynamic. There are few protocols that have been implemented in distributed database and grid computing which is discussed such as Read One-Write All (ROWA), Voting (VT), Tree Quorum (TQ), Grid Configuration (GC), Three Dimensional Grid Structure (TDGS), Diagonal Replication in Grid (DRG) and Neighbor Replication in Grid (NRG). In this thesis, we introduce an enhanced replica control protocol, named Enhance Diagonal Replication 2D Mesh (EDR2M) protocol for grid environment and compares its result of availability, and communication cost with the latest protocol TDGS (2001) and NRG (2007). EDR2M proves data consistency by fulfilling the Quorum Intersection Properties. Evaluations that is suitable and applicability for EDR2M protocol solutions via analytical models and simulations. A simulation of EDR2M protocol is developed and the performance metrics evaluated are data availability, and communication cost. By getting the sufficient number of quorum, number of nodes in each quorum, and selecting the middle node of the diagonal sites to have the copy of the data file have improved the availability and communication cost for read and write operation compared to the latest protocol, TDGS (2001) and NRG (2007). Thus, the experiment has showed scientifically that EDR2M is the adequate protocol to achieve high data availability in a low communication cost by providing replica control protocol for a dynamic network such as grid environmen