The Asia Pacific public sector faces challenges as it attempts to manage an ever-expanding list of services to increasingly informed and demanding citizens. IT plays an important role in assisting governments and other institutions to improve citizen service delivery. This article clarifies the value that Grid computing delivers to an organisation as a whole, underscores the considerations an organisation needs to make when it moves to the Grid, and highlights how Oracle, as the leader in Grid computing, can deliver substantial benefits to the public sector today.
What is the Grid computing advantage over previous computing models? The modern traffic light is a useful illustration of Grid computing’s benefits. Today’s traffic lights now use lots of small LEDs which act together in unison to provide a service to the end user, in this case direction for traffic flow to drivers. LEDs have replaced the traditional incandescent bulb traffic light in most cities around the world, because the group - or grid - of LEDs provides a superior alternative to the single large bulb.
LEDs are a modern innovation which provide light more efficiently and last longer than previous generations of lighting technology. This system also provides better quality of service: if one of the LEDs goes out, the traffic direction service continues without disruption to the end user. Lastly, the flexibility and extensibility of using LEDs comes into play over the long term as they can easily configured or redeployed to provide new services, eg. rearranging the LEDs to represent alternative traffic symbols.
The compelling benefits from this simple example – costs savings through greater efficiency, improved quality of service, increased availability, and modular scalability – are exactly what organisations can expect from their IT systems when they move to a Grid computing architecture.
Reality check: Southern Cross
To get a better feel for what Grid computing and its rewards look like in the public sector look no further than Southern Cross Healthcare of New Zealand. Southern Cross recently deployed a combination of off-the-shelf ERP applications on infrastructure running Oracle Real Application Clusters, one of our main Grid-enabled solutions.
“Oracle Real Application Clusters on Linux is the ultimate just-in-time solution,” says April Walker, General Manager, Information Systems, Southern Cross Healthcare. “It saves us from the need to buy 'big iron' up-front and allows us to add capacity only when we need it. Low initial costs, easy scalability, high availability, and economical support costs make for a very compelling proposition."
So if it is that good, what are the key technologies that enable a Grid computing architecture? It is important to note that Grid computing is not a single technology, but rather an architecture for hosting and running multi-user applications made possible through a combination of a number of modern technologies. With Grid we are talking about the ability to make many individual, loosely coupled systems act together as a single, virtual computer which delivers greater quality of service more efficiently than a single large dedicated system.
It is this virtualisation that is the first of two important areas of technological innovation that makes the Grid possible: virtualisation and Grid-enabled management tools.
Virtualisation
Virtualisation is the common, leading edge focus for R&D in the industry. There are many different virtualisation solutions that exist but the most important technology for deploying servers and applications in a Grid today is clustering software, or ‘clusterware’. Clusterware is the glue of the Grid and it serves as the mechanism to orchestrate the different components in the Grid to act as a single entity.
Clusterware is also the software layer that enables nodes in the Grid to failover from one system to the next without disruption of the application to the end user. When selecting clusterware for a Grid it’s critical that complementary technologies such as shared file systems and storage are compatible with the clusterware so the components can share access to data and resources.
Grid-enabled management tools
The second crucial technology required for successful Grid computing is a comprehensive set of Grid-enabled management tools. ‘Grid-enabled’ means that the tools must provide high levels of automation in system administration and sophisticated provisioning functionality.
Automation is essential because it simplifies and improves the administration of the many different components that make up the Grid. Automated Grid management tools also reduce installation time and potential for error when handling the set up and configuration of servers, shared storage and network systems. The flip side of the coin in Grid-enabled management tools is provisioning functionality.
Provisioning and automated administration go hand-in-hand, and without sophisticated provisioning capability the immediate benefits of improved efficiency and the long term benefits of extensibility cannot be fully realised. Therefore it is important to ensure that the management tools are able to allocate, redeploy, and share resources on the fly within the Grid.
In this way, as an example, an administrator could reallocate the majority of servers within a Grid handling online patient inquiries during the day, to run payroll processing during the night.
Now that we have identified the essential technologies that underpin an enterprise Grid we need to take a look at what are some of the clear steps that an organisation can take to get on the Grid.
Getting on the Grid: Step 1, Standardise
The first step along the Grid path is standardisation. During this stage it is important to pick out what are the modern, strategic standards and technologies that an organisation will rely on to get the most ROI out of its IT investment. By ‘modern’ we mean two things: open standards-based and extensible. Interoperability is the hallmark of any Grid as the many resources within the Grid need to communicate and integrate seamlessly with one another.
These communication protocols and interfaces should be open standards-based which in turn provides an organisation with a more portable IT solution. Once a solution is portable, it gives an organisation greater choice in which vendors it can work with. And choice means the opportunity to drive down costs. Openness in standards and portability in technologies will also factor in a solution’s extensibility.
By standardising on a reduced set of technologies that in turn are more flexible, modular and scalable an organisation can significantly lower costs because it requires less personnel to manage a variety of speciality or legacy systems. Oracle 10g is the only database that has full Grid capabilities for supporting all types applications – transaction processing (OLTP), decision support (DSS), and enterprise content management - in a single solution.
Oracle 10g-based systems are also designed to complement the following modern, industry standard technologies that provide the optimal foundation for the Grid: low-cost, high volume Intel or AMD processors to provide blazing computing power at very competitive costs, network storage technologies such as Network Attached Storage (NAS) and Storage Area Networks (SANs) which provide an extensible shared storage systems; Gigabit Ethernet and Fibre Channel that provides high speed connection between components in the Grid; and Linux which provides a cost-effective and enterprise-ready operating system.
Of course Oracle has a lot to add to standards and portability beyond grid. Since its inception Oracle’s flagship database has been based upon the SQL standard, but Oracle is also a leader in the Java, XML and Linux spaces, whilst continuing to offer broad platform support to Windows and UNIX.
Getting on the Grid: Step 2, Consolidate
Once an organisation has picked the technologies and standards that it will base its IT systems upon, the next step is to consolidate this infrastructure by starting to cluster servers. This sort of consolidation is similar to what many data centres have done recently around consolidating storage on SAN and NAS systems.
This new wave of consolidation should look to leverage technologies that further improve previous consolidation efforts such as Oracle’s Automated Storage Management (ASM) which simplifies use of shared storage. To accomplish this goal customers can leverage the proven and industry leading clustering capabilities that are found with Oracle Application Server and the Oracle Real Application Clusters option for the Oracle Database.
The best news about using Oracle’s clustering solutions is that they can run existing applications in a Grid computing environment without requiring applications to be rewritten. Furthermore, Oracle’s solution portability and broad platform support means organisations can continue to use existing solutions deployed on older architectures while harmoniously integrating Grid computing on a step by step basis.
Getting on the Grid: Step 3, Automate
The last stage in Grid deployment is automation to facilitate management. Because enterprise Grids can have potentially hundreds, or even thousands of servers, a Grid is simply too large to be managed manually server-by-server. Oracle Enterprise Manager 10g has automated the day-to-day maintenance required for an enterprise Grid and provides a centralised management console called Oracle Grid Control. System provisioning, software installation, patching, upgrading, workload balancing, security, and much more are all handled centrally from Oracle Grid Control. This means even the largest grid data centre can be managed by one or a few administrators by leveraging Oracle Grid Control.
Start small, grow big
So there’s a lot to know and learn, but the good news is that it is possible, and even advised, to start small and grow big. Grid is not about ‘rip and replace.’ It’s an evolution, not a revolution and as such should be taken up gradually.
Using Oracle10g Database Standard Edition with built-in free Real Application Clusters is the perfect starting point. Deploying this solution aboard industry standard servers running Linux not only offers a powerful, open standards based foundation for a Grid, but it provides the lowest cost option for getting familiar with more advanced Grid technologies like clustering and automated storage management.
During the early stages of trying Grid computing it is important to invest in education and training to leverage the Grid capabilities built into Oracle’s Database, Fusion Middleware, and Grid Control solutions. As an organisation’s appreciation and understanding for what the Grid can deliver grows, so will its use and investment in the architecture.
From a customer satisfaction stand point Oracle continues to witness our customers moving successfully along the path to Grid. These include the likes of Holmesglen TAFE, Singapore Land Authority, Cairns City Council, Southern Cross Healthcare, Beijing Social Security, and Islamic University of Malaysia, to name but a few.
Our customers are immediately realising lower costs of ownership and an enhanced ability to adapt to and service business needs that supports top line growth. Oracle's 10g Grid enabled solutions are the only ones in the industry that can provide this proven combination of benefits to application users in the public sector today.
Autor: Guy Cross
Quelle: Public Sector Technology & Management, 02.03.2006
