This is one in a series of posts on when and where to use a distributed architecture design in your organization's computing needs. You can find the main post here: http://blogs.msdn.com/b/buckwoody/archive/2011/01/18/windows-azure-and-sql-azure-use-cases.aspx 

Description:

Physical hardware components take up room, use electricity, create heat and therefore need cooling, and require wiring and special storage units. all of these requirements cost money to rent at a data-center or to build out at a local facility.

In some cases, this can be a catalyst for evaluating options to remove this infrastructure requirement entirely by moving to a distributed computing environment.

Implementation:

There are three main options for moving to a distributed computing environment.

Infrastructure as a Service (IaaS)
The first option is simply to virtualize the current hardware and move the VM’s to a provider. You can do this with Microsoft’s Hyper-V product or other software, build the systems and host them locally on fewer physical machines. This is a good option for canned-applications (where you have to type setup.exe) but not as useful for custom applications, as you still have to license and patch those servers, and there are hard limits on the VM sizes.

Software as a Service (SaaS)
If there is already software available that does what you need, it may make sense to simply purchase not only the software license but the use of it on the vendor’s servers. Microsoft’s Exchange Online is an example of simply using an offering from a vendor on their servers. If you do not need a great deal of customization, have no interest in owning or extending the source code, and need to implement a solution quickly, this is a good choice.

Platform as a Service (PaaS)
If you do need to write software for your environment, your next choice is a Platform as a Service such as Windows Azure. In this case you no longer manager physical or even virtual servers. You start at the code and data level of control and responsibility, and your focus is more on the design and maintenance of the application itself. In this case you own the source code and can extend or change it as you see fit. An interesting side-benefit to using Windows Azure as a PaaS is that the Application Fabric component allows a hybrid approach, which gives you a basis to allow on-premise applications to leverage distributed computing paradigms.

No one solution fits every situation. It’s common to see organizations pick a mixture of on-premise, IaaS, SaaS and PaaS components. In fact, that’s a great advantage to this form of computing - choice.

References:

5 Enterprise steps for adopting a Platform as a Service: http://blogs.msdn.com/b/davidmcg/archive/2010/12/02/5-enterprise-steps-for-adopting-a-platform-as-a-service.aspx?wa=wsignin1.0 

Application Patterns for the Cloud: http://blogs.msdn.com/b/kashif/archive/2010/08/07/application-patterns-for-the-cloud.aspx