I am sure that by now we have all heard of the cloud and that many managers are being approached by their IT teams about the possibility of migrating the business applications to the cloud, and hopefully with a stronger argument than ‘because it’s the next thing’. To assist in the decision, let’s take a look at some of the benefits (and pitfalls) of the cloud, specifically Windows Azure as a Platform-as-a-Service, and some scenarios where it makes sense to migrate and some where it might not.
For most managers, of most businesses, the primary determinant of whether or not to make any IT Infrastructure change is going to be cost. Now, this could go beyond a simple ‘how much would it cost to closely mimic my current hosting services in the cloud’. It should include the question of ‘How much would it cost to do on-premise what I could do in the cloud’. Which, in turn, brings up the point of ‘Does the business need the capabilities that the cloud offers’. Let’s look at a few scenarios:
The primary benefit of the cloud is scalability. In an on-premise or Infrastructure-as-a-Service model, this would require purchasing, or leasing, additional servers, installing all required software, installing the application, setting the server up in the network domain and load balancer. This could easily take a week or more and require significant planning for the future. In Windows Azure, additional load-balanced instances of an application take only a few seconds of effort, are up in running in minutes, and you only pay for what you use. Because of this, the usefulness of Windows Azure will depend on the type of usage of the system. If the application is more highly utilized during certain times of the day, week, month, year – then it is likely a good candidate for the scalability functionality of Windows Azure, especially if it is predictable.
For example, if your application is not accessed globally then its usage might look something like this:
The red line represents the capacity your business would have to maintain if leasing or owning servers, and the green line represents the capacity that could be achieved using Windows Azure and scaling the application up and down based on the time of day. The yellow is the savings achieved.
Similarly, the same approach could be used for seasonal increases in system usage. How about the holiday season? To handle the increased capacity, make a configuration change in Windows Azure and it is done.
Back to the task of answering the question of ‘How much would it cost to do on-premise what I could do in the cloud’, Microsoft offers a few calculators to assist with this. Use these to compare to the current cost of hosting for the business.
The simple “ballpark” estimate calculator: http://www.windowsazure.com/en-us/pricing/calculator/
The more complex ROI calculator: http://azureroi.cloudapp.net/
The ROI calculator takes into account both migration and upkeep costs, this is an important distinction because there will certainly be some effort required to migrate an application from on-premise to Windows Azure. These migration costs will be highly dependent on the complexity and architecture of the applications being migrated. A simple website could take less than a week, while an intricate system comprised of several integrated applications could take 3 to 6 months or more. A cost assessment to determine which end of the spectrum you are on should take only a quick architecture assessment.
This blog is the first in a series aimed at discussing the strategy and development of Windows Azure applications.