Dee Baker, a consulting services architect with CDW, once worked with an organization that wanted to put its entire environment — including 7,000 thick-provisioned virtual machines — into the public cloud.

“The answer was absolutely not,” Baker recalls. “The bills that they would have received would have been astronomical. We shifted the conversation toward, let’s look at the workloads that make sense, keeping in mind factors like performance requirements and security.”

Because that organization relied on CDW’s cloud expertise, it avoided a major misstep. But the scenario illustrates just how quickly the cloud can go from boon to boondoggle for businesses that fail to strategically build out and appropriately manage their cloud environments. Ideally, public and private cloud resources will provide organizations with seamless scalability, giving them the ability to rapidly ramp up IT resources to meet evolving needs. But as the consumption of cloud resources grows in an organization, small inefficiencies can scale up quickly, too — until what was once a minor issue grows into a major problem.

Cloud complexity is poised to hamper businesses even more as organizations increasingly look to a larger mix of cloud environments to host their workloads. “It’s a moving target,” says Patrick LaPorte, senior director of cloud and software solutions marketing at Aruba Networks. “When you think about when cloud started, people picked a single source so they could standardize their IT environments. But a lot of organizations are starting to go to multicloud now. They don’t want to put all their eggs in one basket, and they can leverage competition to control costs. But that creates complexity.”

By investing in the right mix of solutions, services and internal practices, IT shops can keep cloud complexity under control and better support their businesses.

To get a better handle on their cloud environments, many organizations have turned toward cloud management platform providers such as CloudBolt. CloudBolt supports more than 20 cloud infrastructures — including private and public cloud offerings from vendors such as VMware, Nutanix, AWS, Oracle Cloud, Google and Microsoft Azure — by connecting through application programming interfaces (APIs) to these cloud environments.

“The problem we’re trying to solve is the balance between control and agility,” says Grant Ho, chief marketing officer for CloudBolt. “In enterprises today, there’s an increasing amount of shadow IT. Lines of business can get access to cloud resources much more quickly. That’s really great for innovation, but on the flip side, it creates challenges around cost management and governance.”

Using an organization’s cloud credentials, CloudBolt can discover the existing resources within target environments and place them under the control of its management platform. CloudBolt has a number of “blueprints” that inject cost and compliance controls into various scenarios, while retaining the agility and scalability benefits of the cloud. “The goal is to help those developers get those resources through the power of self-service IT, but to also ensure that access to those resources is controlled,” says Ho. One enterprise, Ho says, realized $3 million in annual savings after using CloudBolt to get a clearer picture of its public cloud environment and shut down unnecessary resources.

In addition to preventing cloud sprawl and shadow IT, management platforms such as CloudBolt’s can help organizations to enable quick provisioning of private and hybrid cloud resources. Rather than filing service tickets and going through manual approvals for each new instance, organizations can use a management platform to automate and streamline the provisioning and delivery of virtual machines and container-based workloads for developers. It is common, Ho says, for organizations to reduce the time it takes to provision private cloud resources from two or three weeks down to a matter of minutes after adopting a cloud management platform.

As more organizations rely on the cloud for data-intensive applications such as machine learning and artificial intelligence, the importance of planning cloud solutions for optimal performance is more important than ever, says LaPorte. “You have to understand your limitations, in terms of latency,” he says. “Cloud providers recognize this, and they’re moving more and more resources to the point of need. It’s all about reducing latency so organizations can improve speed to take advantage of this data.”

Baker points out that organizations should carefully consider how resources should be moved to the cloud — rather than undertaking wholesale “lift-and-shift” migrations that essentially just replicate the existing on-premises environment elsewhere. For instance, organizations will often discover that they can save money by placing data in different storage tiers, depending on how often and quickly that information will be accessed.

Still, Baker says, the largest cloud inefficiencies are usually the result of organizations simply leaving resources running unused in the public cloud. “You have to have the governance in place to be in a position to not always be reactive, where you can proactively monitor for that kind of thing,” she says. “As much as everybody preaches about IT governance, a lot of times it’s an afterthought.”

Baker notes that organizations can keep management and efficiency at the forefront of their cloud efforts by establishing a “cloud center of excellence” — a cross-functional team of people responsible for developing a cloud strategy, creating governance rules and adopting best practices.

“If you keep doing the same things you’ve always done, you’re going to get same result you’ve always gotten,” Baker says. “You can put everything in the cloud and architect the perfect solution. But then if you manage it like you used to manage your on-premises environment, you’re going to end up right back where you were. All those inefficiencies are going to creep right back in.”