← Writing

Digital Sovereignty for CIOs: From a Vague Fear to a Board-Ready Decision

Many CIOs are asked how dependent their company is on US cloud providers and what they should do about it. A dependency on its own is not a risk, as risk only arises in context. Enterprise Architecture can make this context visible, so that the board can take a decision.

Let’s say you are the CIO of a company. In the news you read about data protection frameworks, about a price increase by a cloud provider, and about a service which is going to be discontinued. A board member forwards one of these articles to you and asks: “How exposed are we here?” The board expects a well-founded assessment and a decision paper within four weeks.

This is an uncomfortable question, as the first answer is “it depends”. This may sound evasive, but it describes the problem quite well: a dependency on its own isn’t a risk. Risk only arises in context. Relying on a US hyperscaler can be dangerous or not, depending on what you use it for, how critical this is for your business, and how easily you could do the same thing in another way. So the job is to make this context explicit, quickly and in a way you can defend in front of the board. I would say this kind of translation is what Enterprise Architecture is there for.

Why Is This Question Asked at All?

There are several reasons why boards are concerned:

  • Legal exposure: the CLOUD Act can oblige US providers to hand over data even if it’s stored in the EU. The legal basis for data transfers between the EU and the US has already been declared invalid once (Schrems II), and nobody can guarantee that the current framework will last.
  • Market concentration: a small number of hyperscalers hold a lot of power, which leads to lock-in. The more you use their unique features, the harder and more expensive it becomes to leave.
  • Operational and strategic control: outages, discontinued services, or a provider changing its direction can make critical processes depend on decisions which are taken far outside of your company.

Many people treat this as an IT problem. I think it is a business problem which is described in technical terms, and this is why management needs someone who translates between the two.

Four Scenarios

Discussing in general how worried we should be doesn’t lead very far. I prefer to look at concrete scenarios and ask what each of them would mean for the business:

  1. The data transfer framework falls, and transatlantic data flows lose their legal basis.
  2. Prices rise, as a provider increases the price of a feature you can’t replace easily.
  3. A service you depend on in your daily operations is discontinued or withdrawn.
  4. Nothing changes at all. Even then the CLOUD Act exposure for data stored in the EU remains.

Each of these scenarios comes with a different probability and a different impact, so it makes sense to plot them individually:

Risk matrix: US-cloud dependencies Mitigate Act now Act now Monitor Mitigate Act now Accept Monitor Mitigate Scenario 3 MEDIUM Scenario 1 CRITICAL Scenario 4 MEDIUM Scenario 2 HIGH Low Medium High Probability Severe Medium Low Impact Risk level (also named on each bubble) Critical High Medium

The Enterprise Architecture Approach

So how do we get to such a picture? I usually go through the following steps.

Clarify the Scope

Four weeks is not much time, so the first step is to narrow down the scope. Which company goals are affected? Do we look at infrastructure only (IaaS/PaaS), or at SaaS like office and collaboration suites, too? Which providers? Which kinds of risk: technical, financial, legal, strategic? Without a clear scope you won’t be able to meet the deadline.

Make the Dependencies Visible

You can’t assess what you can’t see. In most companies the required information is already available somewhere: in an EA management tool (LeanIX, ardoq, or similar), in the configuration or asset database, and, most importantly, in the heads of the architects and developers who know how things are connected. So you need to talk to them.

Assess Each Service on Two Axes

For every relevant service I ask two questions: how business-critical is it, and how easily can it be replaced? With these two axes a flat inventory turns into an assessment, as a critical service with an easy alternative is a very different problem from a critical service without one. For a single business capability the result can look like this:

Business capabilities & US-cloud services Operations Product procurement US-cloud service (dependency) Sovereign alternative (replaceability) AWS EC2 MEDIUM OVH HARD Microsoft 365 CRITICAL Nextcloud Office MEDIUM Miro LOW Conceptboard EASY Dependency (fill, also named on each pill) Critical Medium Low Replaceability (border, also named on each pill) Hard Medium Easy Potential replacement

How easily a service can be replaced depends a lot on the number of interfaces it has with other tools. In the example above, Miro is an online whiteboard, and replacing it with a European alternative means migrating some data to a tool with similar capabilities. Microsoft 365 is much more complex than such a collaboration tool, as it covers mail, calendar and documents, which are integrated with all kinds of internal services. Therefore it’s harder to replace.

Aggregate to Business Capabilities

Individual services don’t mean much to a board. Therefore I aggregate the assessment to the level of business capabilities and present it as a heatmap. Now we can talk about a statement like “our procurement capability has a critical dependency which is hard to replace” instead of a line item called EC2.

An analysis without recommendations doesn’t help the board. So the paper ends with concrete options, sorted by time horizon:

  • Short term: formally document and accept the risk for non-critical services, and commission a deeper review for everything in the red zone.
  • Medium term: make multi-cloud mandatory for new projects, and build an exit strategy or a proof of concept with an alternative provider for a specific critical application.
  • Long term: evaluate a sovereign or private cloud option for core data, and establish “sovereignty by design” as an architecture principle, so that new systems do not create the same exposure again.

Each of these options is linked to an official goal of the company, so that the board decides about the business and not about infrastructure details.

Conclusion

We started with a rather vague and slightly panicky question and ended up with an assessment on two axes, a capability heatmap and a list of options sorted by time horizon. The uncertainty itself hasn’t changed: the legal frameworks are still fragile and the providers are still powerful. But the organization is now able to act.

For me this is a good example of what Enterprise Architects do. We translate technical facts into business risks and options, so that management can take an informed decision.


If you want to go deeper on the CIO’s questions around digital sovereignty, INNOQ has two openly available pieces I’d recommend: CIO-Fragestellungen zur digitalen Souveränität and the Digitale Souveränität briefing.