Relocating an application and data set running on a traditional platform to a cloud-based platform might seem like a humdrum shift. However, you must take into account many points before, during, and after the move. In many instances, you can boil down these considerations to simple do's and don'ts.
Do consider the changes to the application architecture that should be made to take better advantage of the cloud platforms. In many instances, this means the application should undergo a change in architecture before moving to the cloud. Such changes might include the decoupling of data to allow for better distribution intracloud or intercloud. Consider moving to a services-oriented approach, which works and plays well with the abstraction of cloud services through APIs.
Do consider altering the application to use native cloud features. This means changing the application so that it's optimized for the use of native cloud services, such as direct access to elastic storage, management interfaces, and autoprovisioning.
There are many tools on the market that provide a quick "lift and shift" of applications during the move from traditional platforms to the cloud. Although the shifted applications may indeed run, they typically don't take full advantage of the native cloud features, making them costlier to operate and less efficient.
Don't follow the hype. Follow your own requirements. Although many public and private clouds may feel like the right place to put your application and data, you need to do your own requirements planning and proof-of-concept testing before you select the right path to the cloud. In many cases, your solution will be more complex than you originally thought, with multiple cloud providers and technologies in the mix.
Don't forget about performance, security, and governance. I know this advice may feel like a bit of a broken record in this blog, but these concepts should be systemic to everything you do in the migration planning and implementation process.
Sign up for Computerworld eNewsletters.