There's no doubt cloud computing is hot, and the hybrid model allows the best of both worlds. Here's what you also need to consider.
Hybrid cloud adoption discussions typically start with Gartner’s prediction that half of large enterprises will utilize the approach by 2017.
That proclamation was made in October 2013, and much of what has happened since then has followed that trajectory nicely.
Enterprise IT organizations, the argument goes, have enough workload diversification that the odds of a single cloud meeting the specific demands of each workload in that spectrum are low.
Simply put, a hybrid cloud approach gives a large corporation choice.
In order to use these choices responsibly, companies should closely examine their specific needs, deciding how to create a setup to maximize their unique hybrid cloud arrangement. Here are five aspects of hybrid cloud usage to consider.
1. What Runs Where?
We’ve already established that hybrid cloud provides choice. Some workloads run best in your data center on a private cloud, while others do better on the public cloud. The tricky part is establishing guidelines to help decide what runs where.
The two axes to think about here involve the variability of the demand each workload sees and the sensitivity of the data running within the workload. For example, your marketing website contains public data, and its workload is highly variable, subject to different demand loads based on different marketing campaigns that may be active—it’s a perfect match for public cloud. By contrast, the batch workloads that gather financial data on a daily basis for a CFO to analyze has more static demand and highly sensitive data, making that workload a better candidate for private cloud.
Each company will have a different attitude toward data sensitivity and different metrics on workload variability, but that is the structure a guideline should be based on.
2. Multiple Clouds, Single Pane of Glass
While the first consideration is a matter of policy, the rest of the items on this list involve functionally making a hybrid strategy a reality. The ability to run workloads on multiple clouds should not create multiple interfaces for system administrators who have to keep the entire set of workloads running properly, nor should their expertise be stretched by asking them to know everything about every cloud in play.
Instead, a classification of tools that Gartner calls Cloud Management Platforms (CMPs) can help here by sitting on top of multiple cloud providers and offering a single pane of glass from which to manage workloads running on them. Such tools typically insulate administrators from the details of the underlying clouds and instead offer a view of the workloads running on each.
3. Who Is Allowed to Do What, Where?
A key feature of CMPs is the ability to dictate the capabilities of individual users through role-based access controls. Governance such as this allows administrators to, for example, let a team of developers deploy the dev/test version of a workload on a public cloud to match the elasticity of that platform with the variable nature of dev/test while giving operations personnel the ability to run the production version of that same workload on a private cloud.
Developing guidelines, as recommended in the first consideration, as to what runs where requires enforcement to be useful, and that is accomplished through governance aspects of a CMP.
4. Metering and Internal Billing
Allowing a constituency to run workload in particular circumstances on specific clouds is one thing, but what about keeping track of the resources they consume once applications are deployed? By extension, the next item needed is the ability to meter that usage so that it can be analyzed and potentially have costing applied to it for showback or even chargeback.
Enterprise IT is traditionally a cost center whose budget is based on some percentage of company revenue. Unfortunately a funding model like this encourages IT to reduce expenses above all else, even providing innovative technologies that can transform businesses.
The use of hybrid cloud metering and billing, however, enables an Enterprise IT department to bill usage directly to a line-of-business P&L while providing on-demand provisioning of resources. That substantially changes the funding approach to IT in a way that ties it more directly to individual businesses and provides the finances required to react innovatively to the wide range of challenges that hybrid cloud can solve.
5. Changing Your Mind About What Runs Where
The days of five-year Enterprise IT roadmaps may be fading, but the concept that things change over time has not. Public and private cloud capabilities continue to advance, and corporate loosening of data sensitivity issues means that whatever decision is made about what should run where is likely to change over time.
Any CMP solution should be able to simply move a workload from one cloud to another if guidelines change either for individual applications or more seismically across a larger land mass of an IT asset portfolio. In addition, the ability to conduct price/performance tests across different cloud platforms is another CMP feature that can prove useful when determining things like whether or not the latest public cloud price drop is worth a move off a current platform.
Related Article: Head in the Clouds: Can Your Business Fully Operate in the Cloud?
With these considerations as a guide, any Enterprise IT organization should be able to benefit from hybrid cloud. The unprecedented choice provided by mixing cloud technologies is a powerful economic lever that, in some cases, frees IT cost models and enables real innovation. Change along the way in this journey is inevitable, so flexibility provided by a well-featured CMP is an important enabler when harnessing hybrid cloud capability.