| Team Fly | |
permitting different vendors' computers to work together providing a seemingly endless supply of shared computer resources. Oracle sees the grid as revolutionizing the way companies go about doing their business. Grid computing targets the delivery of information as a utility, similar to the way electrical and telephone services are currently delivered to the public—hence the term grid. The industry as a whole, but Oracle in particular, sees a delivery method from the grid such that consumers will only pay for what they use. Interlaced computers will allow idle capacity to be leveraged by the grid to provide for a form of parallel processing on steroids. The following are the major players that enable the Oracle grid technology:
Real Application Clusters (RAC) Involves a suite of networked computers sharing a common Oracle Database 10g and running platform independent clusterware, the glue that makes the interconnect between the clustered nodes so transparent.
Automatic Storage Management (ASM) A front-end management system that can group disks from an assortment of manufacturers together to form a suite of disks that is available to all computers on the grid. ASM encapsulates the complete life cycle of disk management and allocation into a centralized GUI interface.
Oracle Resource Manager Provides a framework within which administrators can control the computing resources of nodes on the grid.
Oracle Scheduler Allows the handing out of jobs to members of the grid to facilitate the execution of business tasks anywhere and everywhere where idle resources exist.
Oracle Streams Assists the processing requirements where copies of data need to be streamed between nodes in the grid and provide the mechanisms to keep data in sync on one database with the database from which the data originated. Oracle Stream's tight integration with the Oracle Database 10g engine facilitates this synchronization and delivers a preferred method of replication.
Figure 1-3 illustrates the primary differences between grid computing and traditional approaches to providing computer services.
The following points reinforce the details of the two scenarios depicted in Figure 1-3:
The three applications at the top each have a dedicated server, each with their own dedicated disk. If the Linux server were to go out of service, the pension application would grind to a halt. There is no built-in mechanism for pension system processing to carry on another server.
| Team Fly | |