As Daniel Melin wrote here (in Swedish), I agree that it’s strangely quiet about the EU regulation on “digital operational resilience for the financial sector” (DORA). Financial organisations of critical importance to society that don’t meet the requirements when 2025 begins risk hefty fines, and those responsible for the organisations risk a business ban.
When I first heard the name of the regulation, I thought of a completely different DORA, namely the research report “DevOps Research and Assessment”, which among other things resulted, six years ago, in the book Accelerate. In a software architecture context, Accelerate-DORA is the exact opposite of EU-DORA; Accelerate-DORA has been talked about a lot for several years. Not least, it’s nice that, using scientific methods, it concludes that organisations that are successful with their software delivery are twice as likely to be successful overall (e.g. profitable). For those of us in the industry this sounds self-evident, but it’s nice to have it “proven” and citable. At the same time, it’s potentially liberating and joyful dynamite outside the industry.
One of the requirements in EU-DORA is to have an exit strategy for moving your system. I think it’s a good example of a requirement for resilience (the R in EU-DORA stands for exactly that, resilience), and I’ll use that requirement as an example for the rest of the text. When I read about the exit strategy requirement, I thought that it will probably take a huge amount of work for some companies to be able to move their most important systems within at most a day, so that they work as usual again. Since then, it has been mentioned, rather loosely, that the time limit is presumably half a year, not a day. The regulation doesn’t specify a time limit explicitly, but the intention will of course become clearer going forward. More about that shortly.
Let’s get back to Accelerate-DORA. In the years before the book, the researchers investigated which architecture was required to be among the companies considered to have good software delivery capability, and they arrived at the following (quote from the book Accelerate):
We set out to discover the impact of architectural decisions and constraints on delivery performance, and what makes an effective architecture. We found that high performance is possible with all kinds of systems, provided that systems – and the teams that build and maintain them – are loosely coupled.
I know, not shocking for those of you who have devoted your lives to software architecture. I’ve used that quote many times to motivate the use of Domain-Driven Design (DDD), a handbook in loose coupling (balanced with high cohesion). But the need for an exit strategy is perhaps an even more obvious reason to strive for loose coupling. What I’m thinking of is that even though it’s super important to avoid tight coupling between teams, it’s even more important not to be so locked in to a particular operating environment that moving the system to a new environment is considered impossible.
It’s kind of funny that there’s a correlation between the two DORAs. If the organisation has excellent order in its software delivery capability, including its architecture, it’s probably “just work” of a reasonable scope that’s required to meet the technical requirements of EU-DORA too.
I guess some of those affected by the regulations think it’s excessive and unreasonable. As a citizen, I think it sounds sensible, and close to hygiene. I wouldn’t appreciate a bank in the EU telling its customers that they can’t access their accounts for the next six months because of some veeery unlikely (Black Swan?) problem with the bank’s systems. A day of downtime would probably be enough to create a lot of turbulence in society. Society risks being destabilised long before six months if people can’t get to their money at the bank. Hence my guess that the pain threshold for the exit strategy, in practice, is somewhere around a day or so.
I realise that many organisations are seriously worried about how to meet the EU-DORA regulation. But I hope they can see it as an opportunity instead! This could lead to something good for them, beyond “just” getting the hygiene in order by meeting EU-DORA. I suggest that instead of angrily throwing money at the problem, they throw software excellence at it! On the other side, there’s also that little aspect of increased success (e.g. profitability) beckoning.
So let’s get the discussion going properly, together! How should the requirements be met? We all want to get over to the other side as soon as possible!
This is a translation of the Swedish original, which has also been published on LinkedIn. Join the discussion here (in Swedish).



