The cloudification of German finance companies: the safe bank?
umlaut experts Lucas Heckmann and Florian Krengel on the opportunities and risks of the cloud – and what that has to do with pizza
Mr Heckmann and Mr Krengel, Among financial services providers, the issue of cloud services is rather contentious. Why is that?
Lucas Heckmann: There are four major providers of cloud services. Three of them are based in North America, that's Amazon, Google and Microsoft, and one is based in China. That's Alibaba. In Europe, we have GAIA-X; they may soon become a serious rival. This means that, as a bank or a financial services provider, I am making myself dependant in geopolitical terms on a company that is situated outside of Europe, that in case of doubt is not subject to European jurisdiction – and that has a different approach to dealing with my data.
Florian Krengel: When financial services providers move data to the cloud, the cloud service providers are able to improve their algorithms. This may on first consideration seem to be a good thing. But what will actually happen when Google or any other cloud service provider starts thinking about using these highly advanced algorithms for their own purposes? Along the lines of "Hey let's set up a bank too". But in this case they will be in possession of as much knowledge as all the other banks in the world put together. In such a context, a German bank would never be able to become a relevant player.
Up until now, we have had Amazon Pay, set up by Amazon. Google has Google Pay. Is there a real danger? Do they really want to become banks?
Lucas Heckmann: From the point of view of a bank, I must at least consider that this may be a possibility. There is a certain danger that Amazon or Google will at some point think to themselves: Do you know what, here's a credit card, that's our bank account, and anyway our analyses are much better than all the others put together. If they want to buy market share for themselves, they can do it.
Let's assume for now that this isn't going to play a role. How do we go about moving to the cloud?
Florian Krengel: We start with a well formulated strategy that highlights in concrete terms the ways in which financial services providers can migrate their IT systems, processes and applications to the cloud – and looks at whether this would even make sense. Advantages and disadvantages are weighed up, value added is questioned. What are the customer's goals? Why do they want to move to the cloud? We also need to ask what kind of cloud would be worth considering.
Cloud services are often compared with “Pizza-as-a-Service”. Why?
Lucas Heckmann: In simple terms, cloud services are a bit like pizzas in that there are endless possible options. You can buy a frozen pizza, i.e. a ready-made product – known as Software-as-a-Service. But you can also get a ready-made pizza base, thereby making use of existing platforms and ideas, and then use your own toppings to create your pizza. This would equate to Platform-as-a-Service. Or you could make the pizza dough yourself – or even build your own oven!
Last year umlaut was approached by a bank. The assignment: to examine whether moving to the cloud would pay off.
Florian Krengel: We are currently hearing a lot of people saying: Move to the cloud - it'll solve all your problems. We are examining the idea on behalf of various customers from the financial sector and going through all the possible scenarios. But we are not working with standard solutions. Staying with the idea of pizzas: we are the ones who know how to buy the right kind of pizza, how you should bake it and how the recipe should be improved. That may sound easy – but it involves some major requirements and challenges.
What kind of challenges?
Lucas Heckmann: Imagine it were to come out that a bank had disclosed all the personal data they hold and that suddenly the whole world knows who has paid off their mortgages and who hasn't, and precisely how much Great Aunt Erna has in her savings account – this is something that a smaller bank simply wouldn't survive. Looking at all the possible scenarios and use cases in order to prevent such a situation from arising is very time-consuming. Banks have to consider very carefully whether the effort will be worth it.
Assuming that a bank is prepared to accept the time and effort involved. How can such problems be resolved?
Lucas Heckmann: Data can be clustered and anonymised. Taking the example of personal data, this would mean: Has Great Aunt Erna – or another elderly person in the Düsseldorf region with an alias – taken out a loan from the bank: yes or no? Yet with anonymisation, personal data is lost, and data is of key importance to the bank. Finding a secure middle course between the two extremes is possible, but very time consuming.
Is it then better if I use lower quality data but better algorithms?
Lukas Heckmann: That's something we examine on an individual basis. The core questions are as follows: How and where should we make use of the cloud? Which applications work with the various different kinds of data? And how many applications will actually still be of relevance for the cloud once all these strategic considerations, opportunities and risks have been taken into account? With the bank, it ended up being around five percent of their applications.
Five percent: sounds as though it's not really worth it. So what are the advantages of the cloud?
Lucas Heckmann: If you migrate everything to the cloud, you can be significantly more flexible; it is easier for you to scale things up; you can analyse data faster and more effectively. You no longer need to have your own data centre – so it saves time and money. It is more and more frequently the case that financial services providers need quick access to a test environment, above all in development. Procuring computers and hiring staff to do this and repeatedly having to set up new systems involves a lot of work. In the cloud, I can write a new script and the server environment is already set up. Within just a few minutes.
Florian Krengel: Many financial institutions are still working with old infrastructure. Everything centres around the so-called mainframe computer. The mainframe is a central computer that is incredibly fast at processing accounting entries and carrying out money transfers. In simple terms, it consists of several hundred thousand lines of code that host the data and process it. Of course this doesn't make it easy for a bank to migrate everything to the cloud in the blink of an eye. But it's almost impossible to get around using the cloud – and we also need to consider the fact that there is a severe shortage of new skilled workers.
Why should we even consider replacing the mainframe with the cloud if it's so efficient?
Florian Krengel: Quite simple: no-one is learning how to do programming any more. There are no new developers that have the right skills. The mainframe specialists from the early days are in some cases already in retirement. Just to get things in context, a mainframe is three metres long, one and a half metres wide and two and a half metres high, and is stuffed full of technology. Maintenance is impossible if you can't find people with the required skills. It's as if you wanted to repair your car but there are no longer any tools available.
Would either of you like to take a look into the future?
Lucas Heckmann: I have this picture of gliders in my mind. A glider stays up in the sky because it uses thermals – without them it can't gain height. In the past few years, the bank landscape has descended into a valley. Ten years from now, there will far fewer banks than there are now. But there are areas of potential - some banks can still manage to get their gliders out of the valley if they develop digital concepts that are viable for the future.