Intro: You said the cloud becomes the brain — the aircraft is, in the future, just an edge node for the cloud. The passengers are happy to have this connection, but from my perspective, that's it for the airline. I do not integrate the cloud into my office floor. I integrate my office into the cloud. Where do you see the border between the cloud and the aircraft? The internet is free — everybody can go anywhere and call any URL. The aircraft should not do that. You as an airline should own your software ecosystem. In some years it will not anymore be a nice-to-have. The airlines that do not go this way down the road will really struggle with efficiency. What it actually enables is getting the cloud in and putting the cloud first, and driving cabin and airline operations in the air out of the cloud.

Stefanie Schuster: You are listening to LOADED: Talking Software for Connected Aircraft. My name is Stefanie Schuster. I'm here with Ralph Wagner, and the title of today's episode is “The Cloud as the Brain.” We had this statement in one of the last episodes — you said the cloud becomes the brain, and the aircraft is, in the future, just an edge node for the cloud. So what do we mean by that? Let's dig deeper today into those statements and what it actually means for airlines.

Ralph Wagner: Let's have a look at the cloud and what clouds have in common — and what that means for the cabin and the cabin infrastructure. The statement that the cloud is the brain is true for almost all digital services we know today. We have services running cloud-based, in a private cloud or in a public cloud, but it's cloud-based services that actually orchestrate all our business processes. That is a very normal thing to do. You do not have an office where one floor has its own infrastructure, its own server, its own recommendation engine, its own whatever. It would look awkward today if that were the case. When you think connectivity, it's very popular to say — okay, everybody has a great connection in the aircraft, and it's good for the airline because the passengers are happy to have this connection. But that's it. From my perspective, that's it for the airline.

Stefanie Schuster: What do you mean by “that's it”?

Ralph Wagner: There is no more benefit out of that —

Stefanie Schuster: — if the passenger would just…

Ralph Wagner: Watch Netflix and do some vibe coding while flying. The real benefit is that the airline has the possibility to connect their cloud, their infrastructure, which they use to orchestrate their processes. They have the possibility to integrate the cabin into this, utilizing the connectivity they get. And this is by far more relevant for their business, their revenue, their profitability, and their future in general.

Stefanie Schuster: I like that you say “integrating the cabin into this,” because typically it's formulated the other way round — you integrate the cloud into the cabin infrastructure. But actually it's exactly like that: you're integrating the cabin into the cloud infrastructure.

Ralph Wagner: Yes, exactly. Getting back to the office and the floor — I do not integrate the cloud into my office floor, I integrate my office into the cloud. That is the real essence of it: to really think about connecting your aircraft, connecting your infrastructure on board, with the cloud. But this is a lot of work to do. Where do we start?

Stefanie Schuster: Before we go into the work to do and what is actually needed — where do you see the border between the cloud and the aircraft?

Ralph Wagner: The border between the cloud and the aircraft is very similar to the usual business processes you have. I would say everything data-driven. For example, you utilize data, you store data, you process data — this is for me all cloud-based. When somebody is authenticating, when you have to process a payment, if you integrate devices on board that transmit data — this data needs to be in the cloud, needs to be processed and analyzed in the cloud. The devices on board, I don't think they are dumb devices. You do not need to have all processing on a device on board, but you need a service running in a smart trolley doing the job there — scanning items, calculating what's left — or a smart oven understanding if the oven is still functioning right or not.

Stefanie Schuster: But do they calculate, or does the cloud calculate?

Ralph Wagner: They need to calculate, for example, the temperatures — and if it goes off temperature, that they turn off, or something like that. But all the data they collect — for example, how did they do the last three months, and is there something inside that does not look right — this data needs to go to the ground, needs to be analyzed on the ground. The same for me for the entertainment system. The seatback screen is something that needs to care that I can scroll and watch something, that the usability is right, that the interaction really flows. But the data — for example, what is shown — that is ground-based, that must be ground-based. On the ground you decide what I recommend to you, where I stopped watching, and so on. All of this is actually ground data, so the screen should not care for that.

Stefanie Schuster: So real-time decision making on the ground. I mean, there might be cases where you still need it on board, if I understood you right, with the IoT devices. But you can always ask yourself: what of this can move to the ground, and where would I have some benefit? Because on the ground I have much more power to…

Ralph Wagner: Absolutely.

Stefanie Schuster: …run analytics, to grasp other data, to connect the data.

Ralph Wagner: It's similar to your phone, where you have your applications running and they have some power to do something, but they actually put all of their data into cloud services. This is where the decision-making process really runs alongside — do I need this decision made locally on the device now, or is it something that also needs historical data, that needs more information? So my total travel journey, having an understanding of where I'm coming from, where I'm going to, and so on — this should not be decided in the aircraft, that should be decided on the ground. All of this is cloud-based services.

Stefanie Schuster: Of course, we need to state that we are talking here about connected aircraft, and about which direction we should move. I think we are not yet there, but this is the direction we see now. Still, if we talk about a cloud-first architecture, we still have topics of losing connectivity, having some kind of need for onboard redundancy. What does that mean for where to draw the line between cloud and onboard?

Ralph Wagner: Yes, I think you definitely need to reckon with connection loss — that you are flying in a region where your connectivity is not available. You need to have fallbacks. But these are fallbacks. And I think you must be careful with making a decision like: because there can be such a fallback, I need to think everything in a different way. You need to really look at the use case. Okay — you are three hours disconnected. How do you handle this? How do you cache data that you want to transfer? How do you make decisions on board without having the cloud behind you? But this is an exception. I would treat it as an exception, because it is really something that will change in the future. Connectivity will not get worse, it will get better, and it's already now impressive. Maybe at one point you will also have different connectivity providers — providers that cover you in a certain region and in another region not. But it's still an exception. For me, this is all still an exception. You should treat it as an exception, and architect it as an exception, and not as the main flow.

Stefanie Schuster: Okay. Then let's move to the cloud side of things. One topic I always like to think about is: what is there which has been built and thought through by other industries, which you can use now? Because there are tons of cloud technologies which might open up new possibilities for airlines which were not there before. Typically for onboard software, a lot was built new, reinvented, just for the aircraft. Let's dig a little bit into what kind of cloud technologies can be leveraged.

Ralph Wagner: What are you thinking about?

Stefanie Schuster: Well, I'm thinking about technologies like API gateways. There are companies and industries which have thought through how to connect different backend services, how to make sure you consider that different parties will access the APIs — there need to be some security layers, there need to be some performance considerations. If I would like to have, on my front-end application, some information which requires data from different backend services, responses from different backend services, there is a performance topic to consider — to get that response fast, especially now if you think about latency and so on. In the cloud there are such technologies already, how to do that.

Ralph Wagner: Yes. Let's also have a look at this statement: when you connect from the aircraft to the ground, the internet is free — everybody can go anywhere and call any URL, and you cannot block them, or you should not, and so on. That's okay. But this gives a bit of a perception of everybody going everywhere by himself and deciding it. The aircraft should not do that. When you have services on your aircraft that connect to the backend — be it running authentication or transferring data, or a very interesting area, avionics data, because you now have a real pipe to the ground — when you get all of this data, consumption data, avionics data and so on, they should not all call different services. They should not have hundreds of entry points. If you are considering to really orchestrate that, you use something exactly like an API gateway, like a federation layer, where you know who is actually calling your services. You use an API gateway to integrate your services within the company. This makes it a bit easier for you to not have so much service-to-service communication. So if one service needs information from another, an API gateway could make this easier to handle. Services that sit behind this gateway in the cloud can more easily trust the connection, because it's your API gateway that is integrating the services. And you can integrate third-party services with that — so if you use a third-party service for recommendation, for example, or for data analytics, you could also route this over an API gateway. That gives you more control. You understand the requests to your different services, you can use it to join your services in an easier way, and you can integrate partners into this infrastructure. And this gives you more resilience regarding the connection between the aircraft and the ground. This is one aspect of it. The next aspect is: what kind of partners, what kind of services would you actually integrate into that?

Stefanie Schuster: Yeah, because all this opens up the possibility — we touched on that in one of the episodes already — that you can integrate very specialized software services. There are tons of companies out there developing such very specialized services: for kids, for accessibility, for recommendations, for whatever. You as an airline should own your software ecosystem. I still believe that owning also means deciding what kind of partnerships to choose and partners to integrate for such specialized use cases — not to own the infrastructure, not to own the vision and the whole ecosystem, but to add specific functionality where you as an airline benefit from somebody really specializing in this specific field.

Ralph Wagner: Yes. And if you look in the past, at the cabin — for example, you made recommendations within the cabin with software that was running in the cabin, if you did it at all. So you needed the knowledge of how to get the software there, and so on. Actually, these companies who did recommendations, or tried to, were not companies whose main job is thinking about how recommendations work, how to integrate AI into that, and so on. Companies that do recommendations and specialize in this, and in data analytics, and in AI integration, are typically cloud services. And they are rather agnostic to an industry. They may serve the travel industry, but not necessarily somebody sitting in an aircraft. So such services are by far more straightforward to integrate when you do this out of your cloud. When you're adding these services, you look for the specialized company in recommendations — not a specialized company knowing how to run a service on board of an aircraft. If you do the connection from your cabin to the cloud right, if you set up the infrastructure the right way, then it will be much easier to integrate such companies within your infrastructure. Also to integrate your own systems — your booking systems, your upgrading systems, and so on. And even whether you collect your flight data, how you do your electronic flight bag — this also gets easier if you are able to use the pipe.

Stefanie Schuster: Which is already done more strongly for ground operations. In the last few years, airlines moved already more strongly into cloud ecosystems, cloud infrastructure. Now, what if an airline is not really set up for that yet and does not feel ready? What's the minimum, or what to do if you as an airline are not yet into a cloud setup?

Ralph Wagner: It comes again to the question of starting with a holistic picture or starting with a use case. I would rather say: build up the right infrastructure first. Have a look at what kind of cloud services you would want to do, and then build up, for example, accessibility to your cloud infrastructure to store data, to analyze data. Maybe start first with analytics tools, start first with what kind of data you would like to collect on board — be it passenger data, consumption data, maintenance data, crew information. Getting this data to the cloud, and then building processes to analyze it, to understand what you get there. And from there you can expand. You can add other services to it — you can look into entertainment services, booking services, onboard ordering services. You might start with something that is nearest to you. So if you have onboard point of sales, you can think about: can I do this over a cloud service? All of this — you need a connection for that, right?

Stefanie Schuster: Yeah. And it will, in some years, not anymore be a nice-to-have. Right now it might feel like a nice-to-have — you can still survive without it. If you don't feel ready, if you don't have the people for it, if you don't have the budget for it, then okay, let's go without the cloud. But it won't be like that in the next few years, because the airlines moving ahead with it will have quite some competitive advantages — in efficiency, in how they run their operations, in how they can serve passengers and keep passengers engaged, know the passengers, use the data they have. So right now it might feel cool because you can do cool things there, but probably in the next few years it is really needed to keep your competitiveness. Even if you need just a few people and just a little budget, get started somewhere just to start setting up this infrastructure.

Ralph Wagner: The airlines that do not go this way down the road will really struggle with efficiency. They will lose the data and lose the processes they need to be more efficient and to serve to the point. And this will make it very, very hard for airlines to stay competitive. Even starting small — having first experience, basically creating a round trip, having some process on board that you transfer to ground and handle on the ground — is something that will give you a competitive advantage. You do not need to start with everything at once, but you need to really get experience with it. After your passenger can watch Netflix — that's very, very important.

Stefanie Schuster: How is it for the different kinds of airlines? We have the very large ones, we have the low-cost carriers, ultra-low-cost carriers, mid-size ones. How do you see this?

Ralph Wagner: You can already see today that all of them make decisions for the connectivity that allows them to be that efficient. So all of them are going this way down the road. They might have a different focus, but they will all need to do technically the similar thing. One might focus on passenger engagement — I turn my seatback screen into something that makes it super personal to be on board, so that I can understand where my next gate is, whether I can upgrade my next flight, enjoy entertainment, even choose my own login to my streaming. Maybe all of this is cloud-based services you can offer to please your passenger, and with this, increase your revenue, because it's more likely that this person books the next flight with you. Now, if you look at a low-cost carrier, they might really look at efficiency — at avionics data, getting avionics data to the ground, understanding their aircraft, understanding if there needs to be some maintenance done, understanding how they can make the turnaround more efficient. Maybe the focus will be a bit different between the carriers — what they process and what they do on the ground. But I don't believe there is a separation of “the one does not need it and the other needs it.” Low-cost carriers are highly digitalized too and have a lot of digital processes. I think that's crucial.

Stefanie Schuster: Sometimes even more, because they have to put their strategy on that — be efficient, be lean.

Ralph Wagner: Yes. So I think it's relevant for all the carriers.

Stefanie Schuster: Okay. What does it mean for the onboard side?

Ralph Wagner: Yeah, the onboard side is probably the most tricky part of it, because the experience we have right now is: you get the terminal and your access points, and that's basically it. The first easy-to-access person is me, with my own device, connecting to these access points — and that's it. Right now the cabin infrastructure is kind of not in this picture on the onboard side. So you need to think about how you connect this. How do you get your screens connected that the passengers are engaging with? How do you get your IoT connected? How do you get your avionics data connected? The first thing is to look at the different devices you have onboard. And I really would look at them as an infrastructure with devices. I would not distinguish an oven from a seatback screen, because you need to think about an infrastructure where your oven and your seatback screen are connected in a similar way — can authenticate themselves as a device and transfer data over a connectivity pipe. And this needs to work for your crew app, for your oven, for your lights, for your seatback screen, for the personal devices that work today. To design an architecture like that is a crucial thing to look at, so that in the end you can really connect every single device, get data from every single device through a defined pipe, through a defined architecture — not that everybody starts to somehow put the data together and transfer it somewhere. That will probably not lead to a lean, straightforward infrastructure.

Stefanie Schuster: So ideally the teams working for onboard use cases in the future should first ask: what does your cloud support? Before they start thinking about what to build on board and how to connect it.

Ralph Wagner: Yes, that's crucial. It's crucial to think about how you make the infrastructure work. You need to think of your devices, your software devices, in two layers. The first layer is how to connect all the devices with each other, how to make them all interoperable. And then you think about: okay, on a seatback screen I need what? Entertainment, booking, passenger information. What do I need on my personal device? I need to be able to pair it, to execute payments, to authenticate — and this I can also connect to the cloud. The crew needs applications they operate. The IoT devices need applications so that they are functioning, but also transferring data and connecting to the cloud. And on the cloud side you are basically accumulating all of this information. That is the way to go, and this is how you need to think a cabin. If you have a server in the cabin, if you have screens in the cabin — be it screens in the galley or screens in front of the passenger — if you have IoT devices, smart seat cushions, you name it, they all need to work on one infrastructure. And this is important to design. I'm right now not really sure how to do this, because right now this is separated, structured totally differently. Now there are devices and networks that are put together as one box — and this is a box called XYZ, that's it, and you have no idea how they intercommunicate, how they exchange data. This is super proprietary. And then you have the next box of maybe some IoT stuff, and then the next box of crew. That is something that is not really smart to do, because it's the airline cabin, owned by the airline. So you would not design it like that. You need to have, first of all, an architecture that is connecting, building the network. And after that you think about: do I need an app that does entertainment? Yes, sure — then build it on top of this network.

Stefanie Schuster: It will also be interesting to see how the connectivity providers will be involved in this game, what they offer. Seeing now already Starlink, Amazon LEO positioning differently on the market.

Ralph Wagner: Absolutely. Maybe one goes more in the direction of saying — I already offer you some infrastructure that allows you to build this infrastructure layer. The other might say — here is the connection, there you go. But in both cases you do need to build your infrastructure. Maybe with the one it's a little bit harder, but you have more freedom to do it; the other has a more defined setup for how exactly you should do it. But both are interested in building an infrastructure, a one-layer infrastructure into an aircraft — how to connect devices and how to collect data. That is the important way to go. Which one you choose is more your preference, I would say. It does not necessarily change the thinking on the common infrastructure on board of an aircraft.

Stefanie Schuster: Well, when we come to a conclusion, one topic here is going for cloud-first — thinking architecture and infrastructure always with the cloud as the brain and mind. It's also interesting to see that LEO connectivity enables that. When we talk about LEO connectivity, it's often about the connectivity — the passengers connected, bandwidth, latency. But what it actually enables is getting the cloud in and putting the cloud first, and driving cabin and airline operations in the air out of the cloud.