In the ever-evolving landscape of IT infrastructure, the pendulum often swings between embracing the cloud and reconsidering the benefits of on-premise solutions. This phenomenon, known as cloud repatriation, continues to spark discussion among organizations seeking the optimal balance between cost-effectiveness, security, and performance.

While high-profile cases like 37signals have documented substantial savings by moving back on-premise, the question remains: Is cloud repatriation an isolated trend or is there a legitimate business case for it?

In this article, we delve into the concept of cloud repatriation and explore the factors driving organizations to rebalance their infrastructure. By examining the potential benefits and risks, we aim to provide valuable insights for IT professionals making critical decisions about their company’s IT ecosystem.

Join us as we uncover the great rebalancing act and shed light on whether cloud repatriation holds the key to unlocking a more cost-efficient and secure future for your organization.

Cloud repatriation, moving workloads back on-premise after being in the cloud, has garnered attention in recent years.

While some high-profile examples like Dropbox and 37 signals have successfully migrated off AWS and showcased significant cost savings, examining whether this trend is isolated or indicative of a broader business case is essential. So let’s explore the various facets of cloud repatriation and its implications for organizations.

The buzz around cloud repatriation goes all the way back to early 2018. At that time, Dropbox made headlines when it disclosed its cloud data migration and subsequent savings of nearly $75 million in infrastructure costs. While Dropbox’s success story was and is notable, their situation was somewhat atypical (unless you too have $53 million to spend on custom architectures).

Later that same year, IDC released a report titled, “Cloud Repatriation Accelerates in a Multi-Cloud World.” This report found that data and application interoperability is more important than ever, but many organizations simply don’t have access to hybrid cloud systems. Instead, the private cloud is a more popular on-ramp to public cloud interoperability.

More recently, 37signals drummed up a lot of attention as well when it moved off AWS. After spending $3.2 million on cloud services in 2022, 37signals decided to work within a sovereign cloud. This ended up costing them about $840,000 per year. While that figure is out of reach for many organizations, it’s clearly much less than what 37signals was paying for AWS.

But Platformonomics Repatriation Index suggests that the overall repatriation phenomenon might not be as widespread as initially speculated. This index analyzes the growth rates of data center providers like Digital Realty Trust and Equinix compared to hyperclouds like AWS, revealing a decline in hypercloud growth without a corresponding data center uptick.

There is an argument that adopting Site Reliability Engineering (SRE) practices may be correlated with public cloud repatriation. SRE principles, focused on operational excellence and reliability, can influence organizations to reevaluate their infrastructure strategy and consider repatriating workloads.

Instead of complete repatriation, many organizations are opting for a rebalancing approach. This involves leveraging more mature hybrid architectures, combining on-premise infrastructure with the cloud. As a result, organizations can optimize their infrastructure strategy and achieve the desired outcomes by striking a balance between the two environments.

As we further explore the cloud repatriation argument, we’ll explore why some companies choose repatriation or rebalancing. Factors such as higher costs, security concerns, and degraded performance contribute to the decision-making. Organizations can make informed decisions and align their infrastructure strategy with their business needs by understanding these factors.

1. Higher costs: Cloud costs often end up being higher than initially projected. Surveys by 451 Research and IDC have ranked cost as the top reason for repatriating workloads. According to Statista, cloud waste accounts for around 30% of total cloud spending.

   Gartner estimated total cloud spending to be about 500 billion dollars in 2022. If these values are to be trusted, cloud waste is higher than the annual GDP of nearly two-thirds of the world’s countries! This substantial waste contributes to the financial burden and prompts organizations to consider repatriation to optimize costs.

2. Security concerns: IDC’s 2018 survey highlighted security as a primary driver for repatriation. IBM’s 2022 Transformation Index revealed that 54% of respondents believe public cloud still is not secure enough.

   Recurring vulnerabilities in cloud providers, such as Azure, AWS, and GCP, further contribute to security concerns. Notable cloud failures, like the AWS outage, have had a profound impact.

3. Degraded performance: Organizations may experience performance issues in the cloud, prompting them to consider repatriation. Businesses may seek alternative solutions when applications, systems, and infrastructure must operate at the desired level.

As the cloud repatriation phenomenon unfolds, it’s crucial to consider why companies decide to repatriate workloads or rebalance their infrastructure. Factors such as higher costs, security concerns, and degraded performance significantly shape these decisions.

To determine whether the repatriation of workloads makes sense for your organization, it is crucial to consider several key factors and build a solid business case. By evaluating costs, benefits, and risks, architects can make informed decisions about optimizing platforms for cost-effectiveness and value. Here are some essential concepts to consider.

Assessing each platform’s monthly costs and values is crucial in determining whether to stay on the cloud or move back to the enterprise data center. This evaluation requires a comprehensive understanding of the costs associated with the cloud, including the talent needed and the intangible benefits such as agility and scalability. Unfortunately, calculating these costs can be complex, as certain benefits might be overlooked, leading to costly mistakes.

Consider all the costs and benefits of being on-premises. This includes the expenses associated with hardware and software maintenance, data center space, depreciation, insurance, power, physical security, compliance, backup and recovery, and more. Also evaluate the value of agility and scalability that may be compromised if they repatriate workloads to the data center. Finally, weigh the potential loss of native cloud services, such as AI and data analytics, which is challenging to replicate on-premises.

Factor in the costs of maintaining cloud-based platforms. Again, this includes expenses related to human resources, security systems, backup and recovery systems, and data movement costs (ingress or egress). Also, consider potential price increases or decreases, both currently and in the future. Finally, evaluate whether the workload leverages native cloud services that are not easily replicable on-premises. Many investments and innovations now occur primarily on cloud platforms.

Repatriation comes with costs and risks. For example, if the workload was refactored to leverage cloud-native features during migration, those changes must be addressed during repatriation. The workload must also be retested and tuned on the on-premises platform. Navigating office politics and justifying the repatriation decision can be challenging.