While cloud storage technology is still a popular strategy for IT leaders, its track record increasingly begs the question: Is it still the easy decision it once was?
Traditional reasons to move to cloud services include cost and space savings, storage capacity flexibility and emergency recoverability. Organizations are reexamining their original strategy in light of their experiences, however. Cloud data repatriation back on premises often makes sense.
1. Cost savings
Going to the cloud has meant moving data, databases, applications and other resources off-site to a managed facility. IT managers expected to save money by shifting the management of these resources elsewhere and reducing staff and other overhead expenses in the process.
A structured migration phased in over a few years would have been the best approach, starting with some small Linux workloads and working up to larger applications. In this way, ROI develops over time and before the organization takes additional steps. Yet, some organizations get caught up in a “thrill of victory” mindset and pursue major cloud storage migrations, running data-intensive workloads with terabytes of memory. In many cases, that’s where increased cloud storage costs end up higher than anticipated, and surprise or hidden costs crop up that weren’t included in the budget.
The need to have a cloud workload strategy is an important aspect of cloud expenditures. It addresses how to adapt systems to run in cloud environments. Moving to cloud services is not necessarily a simple lift-and-shift process.
Cloud service providers have complicated a la carte menus of their various costs — some charge per use, and others charge on an hourly or monthly basis. It can be difficult to calculate upfront exactly what cloud services will cost. Even for storage, there may be additional applications that organizations must pay for beyond the core service, such as security, network connectivity, and management and monitoring tools.
While cost isn’t the only reason to use or not use cloud storage, it is an important one. An analysis of the true cost of cloud storage — the cost per petabyte — in some cases has shown the anticipated savings didn’t happen. Cloud data repatriation then becomes a consideration.
In addition, analyze ROI of cloud systems. While ROI in the early stages may be favorable, over time, it may be more economical to repatriate.
2. Space savings
Moving storage to another location means disconnecting on-site storage resources, such as SANs, NAS devices, RAID equipment, optical storage and other technologies. But how likely is it that an IT department making a push to cloud storage clears out the storage section of its data center and makes constructive use of the newly empty space? Not always likely, and the organization is still paying for every square foot of floor space in that data center.
Assuming IT managers performed a careful, phased migration from on site to the cloud, they probably would have analyzed the use of space made available from the migration. If the company owns the displaced storage assets, managers must consider what happens to them after a department or application moves out of the data center. From a business perspective, it may make sense to retain these assets and have them ready for use in an emergency. This approach also ensures that storage resources are available if cloud data repatriation occurs, but it doesn’t save space — or money.
Continual advances in computing power can mean that repatriation may not require as much physical space for the same or greater processing speeds and storage capacity. One result is that organizations can reuse the available space for more storage devices and other critical systems.
3. Cloud adaptability to business requirements
When an IT organization needs more storage capacity for unexpected upticks in demand, turning to the cloud makes sense. Public clouds, with their vast available resources, are especially useful for responding to emergency needs. But factor the cost of using a cloud-bursting approach into any storage analysis and subsequent contracts.
Also, consider the time it takes the cloud provider to accommodate demand spikes. If those turnaround times are too long, continued use of cloud storage might be up for debate. If it turns out the organization doesn’t need emergency data capacity or that local storage resources are sufficient for emergencies, the continued use of cloud storage may be in question.
While it is true that cloud service vendors offer significant processing power and storage capabilities, the user must examine how they may need to refactor or rework current systems and their workloads to deliver comparable performance in a cloud environment. This is an important consideration based on the potential time and investments needed to refactor workloads.
In addition, prices of HDDs, hardware, power supplies and other components continue to decrease. The potential for cloud data repatriation must be as much a business decision as a technology decision.
4. Emergency recovery of IT resources
One big reason cloud storage is popular is it can be used for data backup and DR, especially recovery of critical data, applications, VMs and other resources. Carefully evaluate the costs associated with data backup and storage for DR purposes.
In certain situations, a cloud backup strategy may not be necessary. For example, organizations with multiple locations — even small, two- or three-person field offices — can configure a comprehensive data backup and emergency storage environment with enough network bandwidth and a rack of storage servers with power supplies. Cloud resources may not be needed, and organizations could avoid the cost through repatriation of cloud data.
Examine the business value of these actions. For example, if critical systems and data are locally backed up, their ability to retrieve and recover critical systems and data is likely to be much faster than with cloud recovery. This is important when considering the firm’s recovery time objectives and recovery point objectives.
Do the math
Monitor the cost of using cloud resources. Periodically compare those costs to alternative arrangements, such as on-site storage. It’s never a bad decision to admit that cloud isn’t the best storage option.
Many organizations are doing the math and discovering ROI for cloud isn’t as good as with on-premises storage, particularly if they aren’t moving everything to the cloud and can’t just turn off what’s left behind in the data center. While cloud-based managed IT services can accommodate a variety of requirements, organizations must analyze all aspects of cloud storage upfront. Often, a more prudent and phased approach to storage outweighs the thrill of going cloud.
Business needs and repatriation
Organizations need an optimized technology architecture that defines how and where they should operate resources, based on business requirements.
There are several different platforms: fully on-premises data operations; a blend of on-site and a combination of public, private or hybrid cloud resources; or a fully cloud-based IT infrastructure. Depending on the organization’s existing IT architecture and the requirements of the new one, this may be another way to justify a move toward cloud data repatriation.