Oracle Database Tips by Donald Burleson

Oracle and Grid Computing Initiatives  

Oracle is a leading member of the Enterprise Grid Alliance(EGA), a consortium of leading vendors and customers for enterprise grid computing. The focus of the EGA is on promoting tools and standards for developers and users, and fostering understanding and adoption of grid computing within the enterprise. Oracle's focus with the EGA is to promote grid computing in enterprises and provide support of enterprise computing via grid technologies and database architectures.

Oracle has provided grid features at no additional cost to the customers. It has also provided the Grid SDK (Software Development Kit) free of charge to support this initiative. Oracle provides distributed SQL features to transparently query or update data in other Oracle databases or non-Oracle databases using transparent gateways, and makes it appear local. Oracle Transportable tablespaces allow Oracle datafiles to be copied from one server to a different one, even across different operating systems.

Oracle Streams can share data between databases or database nodes automatically, enabling efficient communications between applications on a grid. The Grid Infrastructure will be addressed in the next few paragraphs.

How Does Grid Infrastructure Work?

In the last decade, availability and affordability of super fast computers, storage devices and networking devices has improved dramatically. By using the increased efficiencies of processes and systems, grid computing can be achieved as in any utility grid. To explain this further, please refer to Figure 1.1. By putting all the hardware and applications together, islands of under utilized resources are eliminated.

Figure 1.1 Grid Infrastructure

Hardware vendors are selling cheaper and faster machines, server blades, networking equipment, etc., which was unheard of ten years ago. A server blade is a circuit board with memory, CPU and hard disk to be mounted on a rack. They offer the lowest cost of computing that is scalable, efficient and reliable. Blade farms or groups of fast computers are the basis of clustering technology and grid computing.

On the software side of computing, Linux continues to attract more companies by providing an inexpensive and reliable operating system. Linux runs very well on blades and smaller machines up to four CPUs. This CPU constraint is assumed to be overcome in the near future. When considering a blade farm or server farm running on Linux, the cost advantage is much higher than in a Symmetric MultiProcessorSMP) environment.

When a grid environment similar to the one in Figure 1.1 is deployed, users cannot determine where the server is located and where the data is being stored. The experience to the users is a faster, seamless operation of their applications. 

 Oracle started leveraging grid technology with the introduction of Real Application Clusters(RAC) in Oracle 9i Release 1. Oracle 9i has proven its scalability on Linux clusters with up to 32 nodes, though theoretically there is no architectural limit. RAC enables the utilization of blade farms. The RAC technology enables applications to choose the blades depending on the need for resources. This capability has matured in Oracle Database 10g to the Automatic Storage Managementsolution. In addition, the introduction of BigFile tablespaces, etc., has just extended the storage limits, which was unimaginable in earlier Oracle releases and with competing database vendors. In fact, BigFile tablespaces can support up to 8 exabytes, which is 8 million terabytes.

Companies with multiple databases on heterogeneous, dedicated hardware can use grid computing to consolidate the hardware resources, improve utilization and efficiency of resources, and manage a larger environment with less administrative overhead. 

Oracle's portability across multiple operating systems and hardware platforms can be used as an advantage for supporting heterogeneous grids. These heterogeneous grids will allow use of all existing hardware and major operating systems. Applications developed for one operating system can be run on another without any changes. Over the last few years, flexible storage accessible over storage area networks has become very popular, enabling multiple servers to access the same set of disks in a distributed environment.

Oracle Database 10g has introduced the grid control functionality to have centralized management of databases and application servers running on multiple grids. This feature is useful in small as well as large enterprises, when managing tens to hundreds of database servers. All these databases can be linked on the grid to ensure automated database administration of various tasks and data movements among them.

The Automatic Storage Management solution will be reviewed in this book along with the many other new database features. All these features enable the database administrator to expand the existing databases without any storage limits. This book is more focused on easy database administration of the database and not on installing a grid solution using multiple computers. The reader is advised to consult Oracle Enterprise Manager Grid Control Installation and Basic Configuration manual- 10g Release 1 (10.1) (Part No. B12012-02) for more details.

Oracle provides distributed SQL features that make the data seem local regardless of the actual location of the database. Also, Oracle allows access to non-Oracle databases and any ODBC compliant databases by means of transparent gateways and makes the data appear as it does in Oracle databases. This topic will be explored during a review of transportable tablespaces in a later chapter.

Streams is an Oracle feature that adds to the abilities of grid computing by providing a uniform framework for information sharing, message queuing, replication, data movement, data warehouse loading, etc. Oracle Streams helps the grid to keep one or more databases in sync, even when updates are made only on one database.  Streams automatically captures database changes and propagates them to subscribing nodes, applies the changes and resolves any conflicts.  Oracle 10g extends Streams support by adding the streams pool in the SGA   

In a nutshell, grid computing allows flexible management and deployment of the computing resources of the organization that easily adapts to organizational changes.

This is an excerpt from the book Easy Oracle Automation.  You can get it for more than 30% by buying it directly from the publisher and get instant access to scripts from the code depot: