Grid Computing on Database

Grid Computing/Database Technology

'Grid computing isn't a product which replaces or supercedes others, but that it is simply a concept which, to make a reality, requires the use of a lot of existing and new technologies in combination." (What is the grid?). For example, in Oracle's new grid database named 10g there isn't an option which you can license called a grid. Instead, grid computing in databases as explained below is an idea of how storage and processing power should appear to users. Essentially, grid computing coordinates the use of clusters of machines to create a single logical entity. Work can then be distributed across many servers to derive benefits such as availability, scalability, and performance using low-cost components (Oracle 10g: Infrastructure for grid computing,2003).

The most important attribute of a grid database is virtualization. Virtualization is the abstraction into a service of every physical and logical entity in a grid. A database that is a part of the grid must make the physical storage of data files, the physical location of rows in a table, and the CPU power and memory allocations that are needed to access those rows, transparent to the user (What is the grid?). Provisioning, resource pooling, self-adaptive software and unified management are also key attributes of grid databases (Oracle 10g: Infrastructure for grid computing, 2003). Provisioning of things such as server requests, data, and computations requires a grid service broker than understands resource requirements and resource availability across the grid. Resource pooling is also required to achieve better resource sharing. The pooling of individual disks into storage arrays and individual servers into blade farms gives the grid runtime processes the ability to optimize the associations between service consumers and service providers. A critical grid infrastructure requirement is self-adaptive software that automates maintenance and tuning tasks traditionally performed by IT. Further, IT professionals need unified management so that they can treat groups of systems as a single logical entity so that tasks can be performed once and executed on multiple machines.