Database Replication in a Service-Oriented Architecture

Principal goal: to implement, evaluate and refine a database replication mechanism for heterogeneous database systems.

Database replication systems keep multiple copies of a database in sync, so that changes to a master database are reflected in the replicas. These systems make the data available if the original database goes down; they can also improve access times for distributed systems and can be used for disaster recovery.

Existing database replication systems are vendor-specific and do not interoperate. This project will take implement, evaluate and test a design that incorporates replication systems into a web-service Grid architecture [1].

Some issues to address are,

a) Towards scalable Grid database replication
The problem need to be solved is how to offer scalability in the Grid database replication. The tasks are to design a scalable transaction-based Grid database replication model and implement the prototype. Study areas and technologies include, Grid/web service architecture and technologies, distributed system, database replication models, large-scale computing, OGSA-DAI, Grid-FTP, network transfer, etc. The evaluation should based on simulation experiments or comparison with one or few relations replication tools, also analysis of the scalability against responsiveness and fault tolerance.

b) Enable Interoperability in Grid database replication
The problem need to be solved is how to offer interoperability in the Grid database replication. The tasks are to design a transaction-based Grid database replication model providing a certain level of interoperability, and implement the prototype. Study areas and technologies include,information integration, Grid/web service architecture and technologies, distributed system, database replication models, OGSA-DAI, relational database replication, etc. The student should demo the designed replication system is able to interact with heterogeneous database/database replication systems, and evaluate the performance of it, e.g. scalability, extensibility etc.

c) Performance monitoring in the Grid database replication
The problem need to be solved is how to monitor the replication processes and measure the performance of the Grid database replication. The tasks are to design a monitoring mechanism and provide performance measurement metric for transaction-based Grid database replication, and implement the prototype. Study areas and technologies include, Grid/web service architecture and technologies, distributed system, database replication models, performance monitoring and measurement, OGSA-DAI, relational database replication, etc. The student should demo the monitoring ability of designed system, and evaluate the performance of it.