For many years we have favored our own Data Center hosted private cloud operation as we were seeking absolute control over our hardware and data, and had significant concerns about the security of publicly exposed clouds. However, we are rapidly coming to a point where the benefits of migrating to a public cloud could match or even outweigh the advantages of maintaining our private cloud, at least to a certain extent. In this document I try to assess the right moment to transition from our own data center to a public cloud, and how to find the optimal balance between the two for the longer term.

Assess our business needs – before embarking on this transition, it is crucial to conduct a comprehensive assessment of our business needs. This evaluation should involve considering factors such as scalability requirements over the foreseeable future, data type sensitivity, compliance and regulations we abide by, cost considerations (not straight forward at all), and performance expectations.

Scalability and elasticity – our primary driver behind transitioning to public cloud is probably its more financially reasonable scalability and elasticity. A public cloud will provide on-demand resources, so we can easily and relatively cheaply scale our infrastructure when demand fluctuates. A comprehensive analysis is required to assess the value of this elasticity in our case, to what extent it is required and whether, at that expected extent – it is still worth the hassle of the move and the initial investment.

Data sensitivity and security – data security is vital as we deal with sensitive customer data and subject to strict security and compliance regulations. As long as we maintain our critical data on a private cloud, we have complete control and visibility. However, less sensitive data that does not have strong compliance requirements can still be migrated to a public cloud while ensuring adequate security measures are in place.

Cost – cost efficiency is often a leading factor influencing the decision to transition to public cloud infrastructure, thanks to the ‘pay-as-you-go’ pricing model. In our case however we have to carefully assess how much CAPEX are we actually going to save. There certainly will be opportunities for saving when compared to keeping our data centers at current scale, but this requires careful analysis in advance to determine which resources are viable for migration and which resources we will have to continue hosting ourselves.

Performance – it seems that the performance requirements for some of our critical applications will not be met by public clouds. These applications will have to continue running on our own dedicated infrastructure to meet customer expectations.

Summary:

We have to carefully consider the optimal balance between our private and public cloud deployments. We also have to keep in mind that there’s not going to be a single definite answer, but more of an ongoing assessment and reassessment, to determine which workloads and data are best suited to each environment. We will most certainly end up with a hybrid cloud strategy, which will allow us to leverage the benefits of both. Mission-critical and highly sensitive data will remain on our private cloud for the foreseeable future, while non-sensitive data with fluctuating demands can be migrated. We should also continuously evaluate the economic benefits of maintaining our private cloud versus migrating workloads to public cloud, regularly reassess resource utilization and availability to be able to optimize cost-effectiveness.

Our next step should be an extensive hardware and application profiling. We should conduct a thorough analysis of each application’s requirements and dependencies, in terms of other applications, data and hardware, consider factors such as data locality, network performance, and integration needs, and then determine the most suitable deployment model.