If one public cloud is good, are multiple public clouds better?

Back to Blog

Virtana recently published the results of a new State of Hybrid Cloud survey. One of the findings is that 81% of companies in the study who have started their migration to the public cloud have engaged multiple providers. This result tallies with a recent Gartner survey of public cloud users, in which 81% of those respondents said they are working with two or more providers.

Why multi-cloud?

We often talk about “the cloud” as if it’s one giant, amorphous thing, which of course, it’s not. There are many public clouds from different providers, and each has its own particular characteristics and strengths. You could decide, for example, to use Azure as it integrates seamlessly with your on-premise systems, GCP for simpler pricing models and richer set of AI/ML tools, or AWS for its vast set of services and geographical reach. You could select multiple cloud provides to meet regulatory and data sovereignty requirements. Or you could make the decision based on geography to bring your workloads closest to where they’re needed. And, of course, working with multiple providers helps avoid vendor lock-in. There are many benefits to adopting a multi-cloud strategy, but are there challenges? How do you know which cloud is best for which workloads? And what steps can organizations take to optimize all of their cloud deployments?

The challenges of multi-cloud

It makes sense to use the best cloud for any given workload, but if each cloud is a silo, you run into the classic challenges. Your tools will likely be siloed, operating solely within each individual cloud. Further, the features available will differ from provider to provider giving you inconsistent management capabilities across your cloud-deployed infrastructure. You could develop an application so that it spans multiple clouds, but it will only be able to leverage the “lowest common denominator” set of features across all the clouds. Even if you had perfect command and control within each of your individual clouds, you won’t have visibility across your entire cloud estate. This creates blind spots and management headaches, resulting in wasted time, suboptimal deployments and configurations, and spiraling costs.

Take control of all your clouds

The first step in breaking down public cloud silos (or avoiding them in the first place) is to take a workload-centric approach. Instead of organizing and managing around the specific clouds, the focus needs to be on the workloads running in those infrastructures. After all, it’s the workloads that are doing the work the business needs; the clouds are just enablers.

You also need to be able to match your workloads to the best cloud provider—and then select the optimal configuration available from that provider. It’s critical to remember that the ideal deployment might not be in the public cloud at all. Knowing which workloads should stay in the data center can save you from expensive and disruptive repatriation.

Finally, you must be able to manage it all over time. Your business will evolve and your workloads will drift, so you need cloud cost reporting that delivers meaningful details and flexible analysis capabilities—and that does it all in a consolidated fashion across your public clouds. Because optimization sometimes means moving workloads from one cloud to another, or even back to the data center, you also need to be able to operationalize migration as part of your standard IT operations.

Private cloud is an important part of the equation

Unless your organization runs 100% in the public cloud—which would put you in a tiny minority—you still have applications and workloads running in the data center and/or in a private cloud. You don’t want these infrastructures to be siloed either; they must be factored in as part of a comprehensive workload-centric approach. You need consolidated visibility and tools that cover your public clouds, private cloud, and data center so you can manage, optimize, and migrate your workloads to best support business requirements and risk profiles at the lowest cost.

So, are multiple clouds better?

The short answer is: It depends. If working with multiple cloud vendors makes sense given the specific needs of your organization and your workflows, AND you can match workloads to the best cloud, AND you can easily manage and optimize all your workloads irrespective of where they’re running, then you’re well positioned to benefit from a multi-cloud approach. Virtana can help. With the Virtana unified observability platform, you can migrate and optimize across hybrid, public, and private cloud environments.