Your Cloud Strategy Should Start With the Workload

Play Audio Version

Moving all your computing power and storage into a cloud is a big decision, but at the same time there are also many complexities and risks in maintaining your own on-premises data center. Any interruption in power, a storm, or a flood, could create a disastrous situation for your business.

The decision about what to keep in your own data center and what to locate in the cloud has often been presented as a highly complex and difficult decision. What type of cloud? Which provider? How do I connect everything together?

But what if you didn’t need to worry?

What if you could manage your own data center with multiple remote clouds and manage it all as it it were the same infrastructure?

S&P Global’s 451 Alliance reported in August this year that 47% of surveyed organizations had repatriated data from a public cloud to a company-owned data center or colocation provider during the previous year – up from 41% in 2025.

451 Alliance pointed especially to issues of cloud sovereignty and cloud egress charges – these are extra fees charged when data is moved from one cloud to another, or to the public internet.

These are live issues. Nobody wants to pay extra just for moving data around so we need to be thinking beyond questions like are we going to use AWS, Microsoft, or Google? You need to start asking where is the best place for each different part of my business – each workload – to be running and how do we manage everything together?

This leads to six major issues:

• Cloud-first is giving way to workload-first. It is time to stop asking the most basic “cloud or on-premises” question but start asking “what is the right environment for this workload?” There will be different issues around performance, latency, regulation, security, cost and availability – all can produce different answers for different applications.
• Hybrid cloud is really a management problem. It is easy to run three servers in three different locations, but making them behave like a single coherent infrastructure is the clever part. Centralized provisioning, monitoring, security, resource allocation and DevOps are where the value is created – this is how open source platforms like ICDC can be extremely valuable.
• Data sovereignty is becoming strategic. IBM has published research highlighting that the physical location of your data is no longer just an operational concern – this needs to be decided in your boardroom. This is especially true in Europe where governance and control over infrastructure is essential.
• Cost is changing the cloud argument. If you have fast-growing and unpredictable workloads then the public cloud is great, but it’s not cheap. If you have stable workloads that are continuously operating then there are better options for your business. Cloud management impacts your bottom line and can generate significant savings.
• AI makes infrastructure placement more important, not less. Once you are using AI – and sharing or analyzing sensitive or proprietary data – then you also need to understand exactly where and how your AI workloads are being processed. In 2026, Broadcom found that 83% of respondents were considering moving data from public to private clouds for this reason.
• Avoiding lock-in has become part of resilience. Nobody wants to be trapped by their cloud provider or the management platform. Using open source hybrid architecture allows you to change your hardware, provider, or workloads without a complete redesign of your solution.

IBA Group has data center facilities in Prague that support all this flexibility and use the ICDC open cloud platform to avoid the lock-in issue. Let’s consider what this means using a practical business-focused example rather than just sharing more cloud jargon:

Imagine you run a financial services company in Europe.

You have many difficult-to-migrate legacy systems and very sensitive customer financial data  – you don’t want to move all this to a public cloud.

But your IT team does want modern Kubernetes applications, automated DevOps, rapid provisioning and the ability to scale as the business requests.

A hybrid cloud architecture could therefore leave some of the most sensitive data inside your own data center. Other workloads could move to the IBA Prague infrastructure, and where rapid scaling is needed it will also be possible to use a public cloud.

ICDC therefore becomes the management layer that connects all this together, so it feels like a single environment.

Building a hybrid cloud is not just about having several clouds and trying to connect them. It is about focusing on the team that needs to manage the data, rather than focusing on the location of the data and infrastructure.

Think about the users first.

The real value of building a hybrid cloud is not that your infrastructure can exist in several different places – this has always been possible. The value is that the business no longer needs to treat those locations as separate worlds or entities.

You can place each workload where it makes the most sense for cost, security, performance and resilience, then give your IT team the tools to manage everything as if it is one environment. The future of cloud management and strategy is not about choosing a location or provider – it is about making the location matter less.

For more information on all the services available from IBA Group, including Data and AI services, please click here.