Before the popularity of cloud computing, businesses stored their data on local hardware within their premises. This approach provided them with complete control over their data and the ability to monitor costs over time.
Since about 2010, businesses worldwide have increasingly shifted their workloads to the cloud, citing flexibility, efficiency, resilience, fast access to resources, and improved data security as reasons for doing so. Today, most organizations rely heavily on the cloud for a range of activities, including transactional, batch, and real-time workloads; application development; and analytics. But perhaps the primary use of cloud computing is storage. According to Zippia, the percentage of corporate data stored in the cloud has increased from 30% in 2015 to 60% in 2022.
However, in recent years, industry observers have noticed an interesting trend: cloud storage growth is slowing down and, to some extent, even reversing. A recent Nutanix survey, for instance, revealed that nearly half of organizations plan to repatriate applications to on-premises data centers. This shift in application location also affects data storage choices.
This marks a significant departure from the mindset of just a few years ago, when most organizations had adopted a cloud-first approach.
What Is Driving the Repatriation of Cloud Storage Workloads?
The primary reason for considering the repatriation of cloud storage workloads is the cost. Cloud storage costs have increased, and companies are generating more data than ever before. According to a survey by Virtana, 69% of IT leaders say that storage accounts for over one-quarter of total cloud costs, with 23% saying it represents more than half of the costs.
Egress costs also contribute to the consideration of repatriation. Every time data is moved from a cloud repository to any other location, a fee is incurred. S&P Global Market Intelligence reports that 34% of organizations claim that cloud storage egress charges directly affect their reliance on cloud storage. Some cloud vendors also charge API gateway fees for accessing data through API calls, resulting in additional costs.
Even tech vendors are not immune. “At one point, we noticed that our cloud costs were going up much faster than expected,” said Kumar Goswami, CEO of Komprise, a data storage and management vendor. “After we looked at the bills, we realized that we had kept certain sets of data in certain regions and, as a result, were paying several costs that didn’t need to be paid. So, we simply moved things around and reduced our cloud costs by 25% to 30%.”
Ensuring regulatory compliance is another major factor influencing data storage strategies. While it’s certainly possible to maintain compliance with cloud-based storage, some companies prefer to avoid the risk and remain on-premises, noted Henry Baltazar, research director at S&P Global Market Intelligence.
Additionally, degraded performance in the cloud is a concern, often caused by network or zone latency issues. One way to address zone latency problems is to ensure that primary storage is in the same zone or region as the applications it serves. However, high availability requires using multiple availability zones, which can exacerbate the performance challenges.
The complexity of cloud storage options is yet another factor. While it used to be fairly simple to choose a cloud storage option, those options today have become complicated. For instance, Amazon offers more than a dozen different storage tiers. Just figuring out which storage type based on cost and performance considerations can take more time and effort than it did in the past.
At the same time, choosing to keep servers or storage arrays on-premises isn’t a straightforward decision. It requires maintaining additional in-house resources, including servers, disaster recovery and backup systems, and IT staff.
Considering all these factors, it’s clear that there is no one-size-fits-all approach. Each organization must develop its own data storage strategy.
How To Determine the Right Data Storage Strategy
So, how can organizations determine the right data storage strategy for their needs?
Conduct a cost-benefit analysis
Firstly, organizations should conduct a cost-benefit analysis. This involves understanding the current expenditure on various storage options and identifying the sources of those costs. Failing to know this information can lead to overspending, Goswami said.
Make informed decisions based on data
By collecting relevant data, organizations can run different scenarios and use online tools or calculators provided by vendors like Komprise, Wasabi, StorageDNA, Amazon, and Pure Storage. These tools help to optimize storage costs and ensure that data is stored in the right place at the right time.
Have a clear reason for data placement
It’s equally important to have a good reason for putting data in a specific storage vehicle. Organizations must ensure that data is moved to the cloud only when there is a specific need, such as running an algorithm that requires cloud-based data, rather than simply following an autopilot approach, Baltazar said.
There are instances, of course, where organizations have no choice but to remain in the cloud, such as when data scientists and developers need to expose data to AI algorithms that are only available in specific cloud platforms. Similarly, regulatory compliance or performance limitations may require certain data and applications to remain on-premises. Mission-critical databases and large-scale systems often fall into this category.
Consider a hybrid storage model
For many organizations, adopting a hybrid storage model, which combines both cloud and on-premises resources for different workloads, proves to be the best strategy. The ability to switch between these resources based on changing needs is ideal. A G2 survey found that more than two-thirds of companies consider cost savings as the primary benefit of the hybrid storage model.
Explore colocation facilities
Colocation facilities have also gained popularity as an alternative to on-premises storage. BlueWeave Consulting said it expects the global data center colocation market to see a 15.65% compound annual growth rate through 2029.
“Colo centers are highly connected to the cloud, making it a good middle ground for a hybrid environment,” Baltazar said. “If you need to expose data to Azure for a specific algorithm, for example, you can ingress a copy of that data into Azure, do your processing, and then send it back to the colo facility, which is connected directly to the cloud.”
Regularly evaluate and adapt
While these general guidelines are valuable, it’s important to periodically reevaluate your data storage strategy, particularly as new technologies like generative AI emerge. Regular assessments ensure that the chosen storage approach aligns with evolving business needs and technological advancements.
About the authorKaren D. Schwartz is a technology and business writer with more than 20 years of experience. She has written on a broad range of technology topics for publications including CIO, InformationWeek, GCN, FCW, FedTech, BizTech, eWeek and Government Executive.