Key Considerations for Moving Applications to the Public Cloud

Originally posted March 22, 2016 on VMblog.com

So, you have decided (or maybe it has been decided for you) that it is time to move one or more of your current application workloads to the public cloud. As previously addressed in "Transitioning Workloads from the Data Center to the Cloud," there are several reasons why an organization might select infrastructure as a service (IaaS) as the desired platform to host an application workload that is currently running in the data center, but there are several aspects that need to be understood before moving forward with this transition.

Sizing
Provisioning of virtualization resources in a private data center is usually more liberal than what is typically necessary. While some organizations do invest in understanding the performance requirements of each application workload and provisions only the resources necessary, many organizations still prescribe to a methodology of provisioning from a predefined set of resources, which may or may not be fully optimized for the requirements of the guest application workloads. When transitioning to IaaS, it's very important determine the resource requirements of your application workload, including CPU, memory, storage and network, and selecting the proper cloud resources to ensure optimal performance.

Cost
Properly sizing your workloads is not only important to the performance of your workload, but it also has implications on the cost of running those workloads in the cloud. When moving to a cloud solution like IaaS, organizations need to determine what their reoccurring costs will be. These reoccurring costs are based on factors like what type resources and how much will be consumed, as well as a few others vendor-dependent factors. Public cloud vendors like Amazon have online calculators that can help with better understanding what those costs can be. It's easy to see how right-sizing application workloads has a major impact on the reoccurring costs that will be incurred.

Network
Another implication when moving application workloads to the public cloud is associated with aspects of the network connectivity. There a number of scenarios that will have to be considered and accounted for. For example:

• Will public, end user access need to be configured?
• Do secure, private networks need to be configured?
• Will applications running in the public cloud need access to resources still running in the private data center?
• Which public cloud data center will the application workload be deployed to, and what impacts will that have on any of the other network implications mentioned above?

Test
Above all else, it is important to test your application workload in the cloud before engaging in a full production deployment. Beginning the move to the cloud by simply deploying the workload for testing and performance monitoring purposes only can speed up the requirement discovery process and ensure all the issues are worked out before the workload is put into production. Trial and error can sometimes be the best test mechanism when too many unknowns might exists.

Conclusion
While this is not a complete list of all the factors that must be considered when transitioning an application workload to the public cloud, it does allow many crucial aspects of the move to be better understood. Carefully planning the move and leaving no stone unturned will allow for a successful cloud deployment and finally bring your application workloads into the modern cloud era.