Manage Learn to apply best practices and optimize your operations.

Oracle and the grid

Once upon a time distributed application and databases were all the rage and Oracle was in the forefront.

Guest Commentary
Oracle and the grid
by Peter Abrahams, Bloor Research

Once upon a time distributed application and databases were all the rage and Oracle was in the forefront. They were seen to be good because the hardware was cheap and putting the processing near the users reduced the management costs and reduce the dependency on expensive unreliable networking.

Like most good things it grew like topsy so that by the beginning of the 21st. century enterprises had hundreds if not thousands of these islands of processing. Technology and requirements had moved on and in particular:

  • Networks were cheap and reliable
  • User had spread out so it was impossible to put the processing close to all of them
  • Non-functional requirements such as support of peak loads, reliability and disaster recovery became crucially important
  • Data and processes needed to be shared between the islands SAN, NAS and blade technology had become standard
  • Hardware was still cheap but tended to run at a low average utilization which raised eyebrows and questions
  • Support for large shared databases improved
  • Load balancing across multiple application servers became the norm
All of this drove management to the conclusion that distributed data and applications were expensive and difficult to manage and some form of centralization would provide a superior service at a lower cost.

In parallel to this story the scientific community were looking at all the underutilized processing power in their institutions and wishing they could use it to do the gargantuan calculations they needed. They could distribute bits of the calculations to lots of independent machines and then gather the results back and do the final calculations. The problem with this idea was that there was no tool to manage the distribution of the calculations and the gathering back of the results. The answer was to develop these tools and they named the concept Grid Computing. Grid Computing enabled an institution to make lots of small computers behave like one very big machine.

Commercial enterprises and their vendors looked at the scientific grid computing and saw and analogous solution to their desire for centralization. Enterprise grid computing was born with the intent of making lots of small computers behave like one with the aid of management software. However, as is common when comparing scientific and commercial computing, the management problems are different.

Commercial systems already knew how to make lots of small computers act as one. Oracle for example had RAC that could use large numbers of servers against very large numbers of disks, and their Application Server could load balance tasks across multiple servers all accessing RAC data. This all worked beautifully if there was a constant load.

The problem was that having a large virtual machine meant that it had to run multiple workloads against multiple databases. Meeting the non-functional requirements of these workloads meant juggling with the resources. For example adding server power to one work load that needed it at month end while reducing the power to another that peaked in the mornings; or stripping data across extra disks before they became over utilised. Theoretically most of this was possible but required considerable human intervention from highly skilled, highly paid administrators. The solution had to be the same as the scientist, develop software to automate these management tasks. Oracle's has implemented this concept in their new release 10g (where g stands for grid). This required three types of extensions across the product line:

  • Improved monitoring and the ability to generate alerts when resources were becoming overloaded
  • A management environment that can react to the alerts and based on a set of rules decide on and instigate process to resolve the issue
  • Improved tools to enable the reallocation of resources on the fly without having to take any service off-line
Oracle has made changes to the Database, Application Server, Collaboration Suite and Enterprise Manager. Just to give one example RAC stripes data across multiple volumes, with 10g a new volume can be added and RAC will redistribute the data across the extra volume on-line whilst applications continue to access and update the data.

Oracle's concept of the enterprise grid is to enable multiple small resources to act like one big server not just at the processing level but more importantly at the administrative and operations level. 10g is impressive because Oracle has managed to implement the concept consistently and comprehensively across their whole product set in one release.

Copyright 2003. Originally published by, reprinted with permission. provides IT decision makers with free daily e-mails containing news analysis, member-only discussion forums, free research, technology spotlights and free on-line consultancy. To register for a free e-mail subscription, click here.

For more information:

  • Looking for free research? Browse our comprehensive White Papers section by topic, author or keyword.
  • Are you tired of technospeak? The Web Services Advisor column uses plain talk and avoids the hype.
  • For insightful opinion and commentary from today's industry leaders, read our Guest Commentary columns.
  • Hey Codeheads! Start benefiting from these time-saving XML Developer Tips and .NET Developer Tips.

  • Visit our huge Best Web Links for Web Services collection for the freshest editor-selected resources.
  • Visit Ask the Experts for answers to your Web services, SOAP, WSDL, XML, .NET, Java and EAI questions.
  • Couldn't attend one of our Webcasts? Don't miss out. Visit our archive to watch at your own convenience.
  • Choking on the alphabet soup of industry acronyms? Visit our helpful Glossary for the latest lingo.
  • Discuss this article, voice your opinion or talk with your peers in the SearchWebServices Discussion Forums.

Dig Deeper on Topics Archive

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.