Best Practices to Move your Applications to Cloud

In this previous article, we had a quick overview on the different cloud service models and how it eases your challenges w.r.t administration. Now, let me ask you a simple question, can every software or workloads be moved to Cloud? The answer will definitely be a NO.

“No single delivery model meets all needs, but the combination of public, private, and hybrid clouds offer a range of options suitable for many business requirements”

Every application is different; and having a better understanding on what workload can be moved to a Cloud platform is a key to have the right cloud enterprise architecture. These below scenarios will give you an idea on what is a general market trend and patterns which are optimal for Cloud platforms adoption; and when should you move your applications to Cloud.

On/Off Work Pattern: In this scenario, the application only runs at certain times. Jobs which are batch processed fall under this pattern. Over Provisioned capacity is always wasted and the time to market can be cumbersome. For eg: The pdf printer job used by the HR team during the monthly payslip sheet generation is not a usual task but a batch process, e-book publication software which converts books to .PUB formats when a new book is ready for market.

Cloud: on off patterns

Growing Pattern: In this scenario, we can see how some business tend to grow too fast to be able to use regular hosting approach. As the business growth rate is high, one cannot procure and provision hardware fast enough. The successful services needs to grow or scale; keeping up with the growth is a big IT challenge. Example: A simple web application, which has seen great adoption and is being used by everyone in the community. You cannot have a downtime for this application and then scale it up; but need to manage it while the application is live. This is one best scenario when a cloud platform can be adopted.

Cloud Growing Patterns

Predictable Increase: In this scenario, the peak demands periods of a requirement are well known. Peaks due to the periodic increased demands. For eg: A product testing phase, when the utilization can go multiple and be predicted. Or even a Simple Quarterly results announcement website, where the page visits are high at specific periods. For eg: University Websites when results are announced every semester; News portal showcasing the quarterly results etc.

Cloud Predictable Bursting Patterns

Un-Predictable Increase: Scenarios where the bursting cannot be predicted or “situations driving business” fit in this category. Unexpected/unplanned peak in demand; sudden spike impacts performance. For eg: During Disaster Recovery, the number of utilizations of the machines and computer to run the backup job might spike high. Another general example could be, a sudden demand of product and vendors placing orders on the request portal. A Video posted on your website which goes Viral on Social Media etc.

Cloud Unpredictable Bursting Patterns

Summing up, you can look at using cloud Services in these following cases (not just limited to these listed below) –

  • Whenever you have an application with variable load which also has a need to scale massive with high reliability. One best example could online shopping portals during festival seasons or Thanks Giving day.
  • Web-apps with a short lifetime can also be a good candidate for cloud; take an example of a social media campaign your company is planning to do.  It could also be a Contest which you do for Christmas.
  • Sometimes large applications which does not fit well in your own data center, becomes a good option for cloud.
  • If you are planning to run many similar jobs at one single time, running the same on two or more different cloud instances will give you a same result at a much faster pace. Of-course, you also have an option to replicate the same machine configurations across any number of computers (VMs) you want.
  • If your application is going to leverage a variety of data available on the Internet (like Pictures, Tweets etc) then having a public data store option becomes more efficient. In such a scenario, running your software(as a service) from the Cloud makes it even more flexible.

Leave a comment

Your email address will not be published.