Over the last decade, cloud computing has become one of the biggest trends in the IT industry and this trend is projected to grow for the foreseeable future as Gartner estimates more than 85% of organizations will embrace a cloud-first principle by 2025, and will not be able to fully execute on their digital strategies without the use of cloud-native architectures and technologies.
Over the last few years, a reverse trend has also begun to develop. Organizations have started to move some of their workloads from the public cloud back to on-premises environments. This phenomenon is known as unclouding, declouding or cloud repatriation. A recent survey of 350 companies by Fortinet revealed that 74% of companies had moved at least one application to the cloud and then moved it back to on-premises infrastructure, after failing to achieve the expected benefits. This trend is supported by a study by IDC which states that 53% of enterprises are considering bringing back some workloads on-premises.
This article will explore the reasons for unclouding, and the considerations organisations need to make to ensure they make the best decisions on this subject.
Why are organizations choosing to uncloud?
Some of the biggest reasons for unclouding based on user surveys and discussions with enterprises are explored below.
- Failure to achieve the expected cost benefits of the cloud
One of the biggest anticipated benefits of the cloud is cost savings. Using the cloud allows organizations to manage rapidly growing and unpredictable workloads without having to make large capex outlays on IT infrastructure. The cloud therefore provides these organizations with the flexibility to manage costs by replacing large capex outlays with smaller opex expenditure based on metered consumption, with improved discounts on longer term reserved compute power. However, several organizations have experienced bill shock with their cloud deployments due to much higher-than-expected operational expenditure, caused by higher compute CPU and memory allocations than required (no rightsizing).
Some of these failures to achieve anticipated benefits, point to issues during the cloud migration exercise, which make it difficult to remediate. These issues include inadequate discovery and dependency analysis prior to migration, lack of application refactoring prior to migration and the absence of a coherent migration approach. For example, moving two application servers that are communicating between two on-premises data servers, to different regions in the cloud would incur network bandwidth costs between the servers (per K/bs) whereas prior to migration it would have been a fixed network cost shared by all services.
- Changing business case due to changes in workload predictability
Another reason for unclouding is a change in the predictability of workloads. Organizations often take advantage of the cloud to manage unpredictable workloads as they grow rapidly. At this stage, the flexibility and agility gained from the cloud provides sufficient advantages especially since the demand patterns are often not yet known. As their businesses stabilise and demand is more predictable, the business case may favour on-premises or hybrid deployments over the public cloud. However it is worth noting that with the correct optimization and dynamic resource scaling it is possible to optimise cloud services, but this takes expert knowledge and a good understanding of existing workload patterns.
An example of this is Dropbox. The company reportedly saved over $74 million in operational expenditure over two years, when it unclouded about 90% of its workloads from AWS into its own data centres. The company has been able to streamline its costs based on an infrastructure-optimised hybrid cloud environment.
- Security concerns
According to a report by IBM Security researchers in 2022, almost half of all data breaches occur in the cloud. According to this report, 83% of companies experience data breaches and the average cost of each data breach to each company is over $4.2million. The report also noted that the cost of these data breaches was less for companies with hybrid cloud models than organizations with public cloud models. The findings of this report help to underscore the reasons why many companies are unclouding some of their workloads, as a default connected public cloud service is more open to standard default policies (such as leaving AWS S3 buckets open to the public). However, with correctly configured services and strong security governance, the cloud can be made as secure as on-premises deployments.
- Limited opportunities for customization
The cloud offers several major competitive advantages to organizations however, to reap these benefits companies must ensure that workloads are adequately optimized for the cloud. This optimization requires skills and resources which are scarce and expensive. Several companies are therefore choosing to keep some workloads on-premises to avoid the need to conform to the services and solutions offered by cloud providers. In a recent survey done by Intel, 24% of respondents cited customisation as a key reason for unclouding. Furthermore, the survey noted that certain applications are designed in a way that makes optimization for the cloud difficult.
- Regulatory compliance
The responsibility for regulatory compliance remains the responsibility of a company, regardless of the cloud model it chooses. As a result organizations must consider these regulatory requirements in the choice of a cloud strategy. Whilst cloud service providers offer solutions to meet regulatory requirements such as the AWS DoD cloud, some organizations are choosing to maintain hybrid environments or private environments to take direct control of their compliance requirements.
Is unclouding the way to go?
Despite some of the perceived advantages of the on-premises model to overcome the challenges highlighted above, unclouding is not necessarily the best or the only solution. There are several considerations every organization must grapple with, before deciding to uncloud.
Firstly, organizations considering a move to the cloud must ensure that this move is underpinned by a coherent well-developed strategy. Furthermore, care must be taken to avoid the known pitfalls of cloud migration programmes.
Secondly, many challenges that are encountered once an organization has moved to the cloud can usually be resolved by knowledgeable and skilled professionals. Unclouding should not be an immediate solution to resolve these issues.
Thirdly, it is important to remember that on-premises environments also pose challenges such as large upfront capex requirements, slower organizational agility, and reliability challenges. According to the enterprise strategy group, 43% of organizations that return workloads on-premises incurred costs related to downtime. This underscores the challenges that pushed organizations to the cloud in the first place. A move back to on-premises therefore must be a carefully measured decision backed by a coherent strategy.
Finally, the move to the cloud represents a departure from the comfort zone for many companies. This is because it requires a paradigm shift in IT divisions including the deployment of new operating models, new skills, and new resources. Businesses must ensure that any challenges encountered in the cloud do not result in a knee-jerk decision to uncloud. Unclouding is a decision that must be carefully made and implemented with the benefit of a well-developed vision and strategy.
How can Infosys help?
At Infosys, we have successfully managed some of the largest public cloud migrations globally. We have also helped several clients to setup private cloud and hybrid cloud models. We are a global leader in next-generation digital services and have been recognized by Gartner as a leader in cloud professional and manged services. We have over 100,000 Cloud Certified professionals, in addition to specialist vertical teams who bring domain expertise to help in your journey.
Managing Partner, CIO Advisory
Chris is a technology and sales leader with over 25 years’ experience and a proven track record in designing, developing and leading multidisciplinary teams to implement complex large-scale solutions for many different business sectors, globally.
CTO, CIO Advisory
Ross has over 25 years of Enterprise IT experience. He started with roles in Engineering and Infrastructure, then moved into Strategic Cloud Consultancy on Azure and AWS. More recently, Ross has focused on complex cloud transformations and Alliances Partner Networks.
Tolu has over 22 years’ experience in management and technology consulting, as well as in complex program delivery. He combines his expertise in delivering strategic and business-critical initiatives with strong business acumen, delivering several major technology transformation programs with large matrix multi-geographic teams.