Insight and analysis on the data center space from industry thought leaders.

Understanding the Cloud Lifecycle: When to Go, When to Return

If your enterprise adheres to the core principles of the cloud operating model, cloud repatriation can be a seamless process.

Ugur Tigli, MinIO

December 12, 2023

4 Min Read
Cloud repatriation and data centers: When to Go, When to Return

The subject of repatriating from the cloud is surprisingly controversial these days. It shouldn’t be – after all, the cloud is an operating model, not a location. Nonetheless, the cloud industrial complex has deemed it a threat, with reports insisting that companies are not repatriating or that there’s “nothing to see here.”

Repatriation is a natural part of the cloud lifecycle. If your enterprise adheres to the principles of the cloud operating model, it will be a fairly seamless transition. Seamless doesn’t mean inconsequential; on the contrary, where to run those workloads has a material impact on operational efficiency, cost structure, and strategic agility. It is just about knowing when to go and when to return.

When to Start in the Public Cloud 

Let’s start with when you should go or start with the cloud. The team at Andressen Horowitz wrote a post entitled The Cost of Cloud, a Trillion Dollar Paradox. They start with, “You’re crazy if you don’t start in the cloud…”

The cloud has the ability to provide infrastructure immediately and at the exact scale needed, fostering operational and economic efficiencies and driving innovation by freeing up resources to focus on new products and growth. Further, a rich portfolio of applications available as-a-service can be jet-fuel for developer teams.

Related:How Data Center Operators Can Capitalize on Cloud Repatriation Trend

The cloud also has elasticity for emerging workloads. Paying for what you need, when you need it (coupled with the vast array of services and APIs it offers) is ideal for workloads that are growing rapidly or whose demand attributes are not well understood. 

Finally, for developing cloud-native skills, the best place to do that is in the public cloud. It is an excellent platform for companies to develop cloud-native muscle, learning skills like containerization, orchestration, APIs, and microservices.

This environment allows for rapid assembly and deployment of infrastructure and applications to solve business problems. This is the cloud operating model, and it will be needed in the next phase. 

When to Repatriate and Move Your Operations On-Prem  

The second part of A16Z’s sentence is  “…you’re crazy if you stay on it.” Here is the argument to return.

Emerging workloads are unpredictable. Mature workloads are the opposite. This is the time to start considering what to optimize for next. As you grow in the public cloud, the benefits of the cloud change – particularly economically.

The pressure on margins from high cloud costs can outweigh its early advantages, especially as growth slows down. For enterprises handling vast amounts of data, like petabytes, the economics tend to favor private cloud environments. This is due to the high costs associated with data transfer and service costs in public clouds, which can make large-scale operations unsustainable.

Related:Container Security in the Cloud: Understanding Concepts and Requirements

Repatriation cost savings run around 60%. This is what we see consistently from the teams that have published their numbers like 37Signals,, Prerender, and Ahrefs. Many companies we know continue to run workloads in AWS, Azure, and GCP so they don’t publish their numbers, but according to their spreadsheets that is where they land too. 

Finally, repatriation increases control. AI is accelerating this trend. You cannot outsource core competencies and expect to survive. Data and data infrastructure is a core competency for every company because every company is now a technology company. If you are letting others learn from your data, you aren’t going to be around very long. That means bringing it together, and on-prem. 

Implementing the Cloud Playbook

A successful cloud repatriation strategy involves several key steps:

  1. Start with an understanding and appreciation of the cloud as an operating model, not a place. Adopting the core principles of the cloud like orchestration, containerization, automation, and software-defined everything will provide options for all scenarios. 

  2. Prioritize infrastructure spend as a key performance indicator and optimize cloud usage early in the lifecycle. This involves tracking cloud spend closely and incentivizing engineers to manage cloud costs effectively.

  3. Plan for potential repatriation early in the cloud journey. By architecting systems with future repatriation in mind, the transition back to private clouds or hybrid models can be more manageable and less disruptive.

  4. Be open to a hybrid approach, involving a mix of public and private clouds. This allows companies to retain the flexibility and scalability of the cloud while managing costs more effectively. In some cases, even companies that aggressively repatriate workloads retain a significant percentage in the cloud.

The journey to and from the cloud is not a one-size-fits-all scenario. It requires careful consideration of a company’s stage of growth, the nature of its workloads, and its strategic priorities. The cloud offers significant advantages for early-stage innovation and development, but as companies scale, the economic benefits can diminish, prompting a shift towards private cloud environments or hybrid models. 

The key is to maintain flexibility in the infrastructure strategy, optimizing for both innovation and cost efficiency through the cloud operating model. 

Ugur Tigli is chief technology officer at MinIO, overseeing enterprise strategy and working with its enterprise client base.

Subscribe to the Data Center Knowledge Newsletter
Get analysis and expert insight on the latest in data center business and technology delivered to your inbox daily.

You May Also Like