In the 81st episode of Something with Logistics host Thomas & Jens talk together with Wolfgang Hackenberg (SYNAOS) and Peter Neuling (Volkswagen) about the use of logistical complexity by means of data. Wolfgang explains not only the actual technical solution but also the potentials and the vision of SYNAOS. Peter, who is involved in the implementation at Volkswagen together with the SYNAOS team, gives an insight into the practice and the user perspective.
Listen to the episode now!
Click on the button below to load the content of something-with-logistics.podigee.io.
Thomas: Yesterday my forklift died, today stress with the shopping and now this shit. Who the hell put this up here? Don't you have a podcast?
Jens: Chief? Chief? What kind of podcast!
Thomas: Well, something with logistics!
Thomas: Something about logistics. The podcast with not always entirely scientific topics, all about the exciting world of logistics, presented by Andreas, Jens and Thomas.
Thomas: Hello and welcome to a new episode of Something with Logistics. I'm here today with Jens at the start. Hello Jens.
Thomas: And then we have two guests with us today, Wolfgang from SYNAOS and Peter from VW. Hello you two.
Wolfgang Hackenberg: Hi.
Peter Neuling: Moin.
Thomas: I would suggest that before I talk a lot of nonsense about what you do or maybe don't do, you both introduce yourselves first before we go straight into the topic.
Wolfgang Hackenberg: Yes, then I'll jump right in. Wolfgang. I am one of the three founders of SYNAOS and act in the role of CEO. So I have all the strategic topics - sales, HR etc. and was previously in the automotive industry for ten years in the field of logistics IT and did a lot with digitalisation. Then we founded SYNAOS, which we will talk about in a moment. Maybe that's all about me at the beginning.
Thomas: Yes, then I pass my baton to you Peter.
Peter Neuling: My name is Peter. I come from Volkswagen's IT department. I am responsible for a manufacturer-independent control system for driverless transport systems and have been with the company for a long time. I have 15 years of logistics experience and have been in IT for a few years now.
Thomas: Okay, SYNAOS is somehow a term that you don't know like that. A name that you don't know like that. What exactly is behind SYNAOS? What is your product? How do you earn your money? Tell us a little bit.
Wolfgang Hackenberg: Yes, with pleasure. SYNAOS is a software company from Hanover. We are now two years old and have 80 employees. SYNAOS stands for Synchronising Chaos, so we combine creative chaos with an ordering factor and have a very big vision. We develop software, an operating system, as we call it SYNA.OS, for the cross-site optimisation of intralogistics, both in factories and in warehouses.
Jens: That's where I'd like to jump in. Why would you like to synchronise chaos and not just, let's say, make chaos disappear? Isn't that actually the most obvious thing to do when you look in the direction of production, intralogistics, production logistics? Who wants chaos there?
Wolfgang Hackenberg: Yes, I think you can prevent chaos - chaos does not stand for real chaos - but you cannot prevent a certain complexity, if we want to paraphrase it that way. That's simply the way it is. Today you have the most complex supply relationships, the most complex tasks to manage in intralogistics and there is simply a basic complexity. And now you can decide in two directions: You can try to fight this complexity. Yes, this has been done for a long time and many of us are trained to fight this complexity. Or you can simply use what complexity brings with it, a great deal of data, in a positive way, to use this complexity in its strength. To use the positive power of complexity. We see a great opportunity in this, because with the demands that customers have today, which are placed on logistics networks. You will have complexity in it and you have to use it as well as possible. That is our vision and that is what we are doing, namely data-driven.
Jens: I think that's a pretty cool vision. To be honest, I've never heard of it before. Normally, as you just described, the approach is to say I'm reducing complexity and not getting the benefits out of complexity. Maybe one question before you explain, besides the vision, what your product is actually about in one sentence. How did you come to say that I adapt, I participate in complexity and chaos and I do not reduce complexity and chaos? Was there an initial push of thought, something that told you: OK, this is the right way and not an avoidance?
Wolfgang Hackenberg: Yes, exactly. We have a background in the automotive industry and I'll just explain it with the example. But you can transfer that to warehouses in the same way. Due to increased customer requirements over time, due to development, the megatrends that we all know about, the level of complexity has increased greatly. If you have a vehicle today, you have 200 different steering wheels that have to be brought to the line in this example now. So you can no longer say to the customer, as Henry Ford did: You can have the car in any colour, the main thing is black, and we have zero complexity in it and I can bring huge containers to the line. But you have complexity and it's the same in the warehouse. You have shorter and shorter order times, you have to deliver faster and so on. And now, of course, you can think about it all day long, so we don't want to promote complexity, but you always have a certain level of complexity and that's where data simply accrues. The new technologies that exist, the digitalisation technologies. If you use them skilfully, they help to create a great deal of flexibility in this complex environment and to really make use of it. That is exactly where we are starting from and that is our vision. If I now go into your second question, it is to do this for entire locations, entire factories, entire warehouses, and to really offer a holistic solution - holistic optimisation - from incoming goods to outgoing goods with new planning algorithms. That is something that is possible today. We'll get into that here in a moment. Cloud technology allows us to deal with this complexity in a completely different way, and we do it for mobile robots, we do it for human-guided systems such as forklift trucks, tugger trains and we do it for people. We integrate these three big building blocks into a holistic planning and optimisation and that makes us unique. Our first product, which we call SYNA.OS for Operating System - for Logistics, for mobile robots - is a solution that can control and optimize large fleets of robots, mobile robots, i.e. AGVs or AMRs, independently of manufacturers and with a focus on process optimisation. That is very, very important. Not like many systems do today with a focus on hardware, but on process optimisation.
Thomas: Are the points you just mentioned at the end: process optimisation and consolidation, cross-manufacturer AGVs, AMRs and so on. Is that exactly what differentiates you from other systems? Because at the end of the day I always thought you had an AGV control system somehow. It's a transport management system like any other, which integrates with a warehouse management system. Yes and I thought there are already a few of those or what is it exactly what I said at the beginning, your distinguishing feature plus also the complexity of the data etc.?
Wolfgang Hackenberg: I'll tie that in again with the data, because I think that's very, very important. Of course, in a first step you can say that we have developed something similar to a manufacturer-independent AGV or FTF control system. But the vision, what is very important to us, is much bigger. It also includes forklifts and people, an integrated optimisation. That has very, very huge levers. What distinguishes us: we are independent of manufacturers, we have a very strong - we will perhaps hear from Peter in a moment - a very strong process view of things. We don't just look at any vehicles, but when you have 200 vehicles on a map, you are no longer interested in each individual vehicle, but whether the deliveries are on time. Are the orders there at the time they need to be there, and we have a new planning approach and that makes it special. Now we can also make the conclusion about complexity and data. Today, planning is done in waves, in the different systems. Once a day, once an hour. What we do is a constant comparison between the real-time picture and the plan picture, with a constant optimisation of the plan. To do this, we calculate new solutions many thousands of times a second. Today, quite a lot of decisions are made based on KPIs aggregated from some millions of real-time data, and people then have to make decisions quite often. What we do: We take all the data we have into the decision, at every point in time and really also on a smaller, smaller scale, we calculate 250,000 solutions per second and we always take the best one. That means no emergency strategies. All that is omitted and we always have an optimal solution.
Thomas: Peter, I would like to have another... Hello to you. I'd like to take a look at what Wolfgang has just described, and both very clearly and with a lot of conviction. Wolfgang, it is clear that you are not only the CEO of the solution, but you love the solution, you live the solution. Do you agree with him that this is exactly the right approach? Actually, to embrace the complexity, to pull the best possible out of it and, above all, to put a higher-level control optimisation on top of the processes that happen on the shop floor.
Peter Neuling: Yes, I think there's no getting around it. I would like to be a bit more specific. We have a project in Hanover where we want to drive 200 AGVs, 175 underride and then also around 30 tractor units. You can imagine that a lot of data will be generated. Especially since we also have to communicate with some of the plants or transfer stations. Of course, huge amounts of data are generated. In this respect, I agree with what Wolfgang said at this point. You also have to look at the issue of higher-level control systems. We have it. We have been pursuing this topic for a little longer. You are certainly familiar with this VDA 5050. As I said, if we have two different types of vehicles that we operate, also from two different manufacturers. Because they don't come from one manufacturer, which is what we need. So a higher-level control system that operates a standard interface is indispensable, otherwise it won't work. You can imagine that the hall is 250 by 250 metres, and 200 vehicles are travelling there. You have a flurry of activity on the screen and that's it.
Jens: But we are talking, just to understand again, when we talk about control and complexity, we are talking primarily just so that I can classify it correctly again. Especially in the area of AGVs, AGVs, self-driving vehicles in direct contact with people on the shop floor. That is where you are currently active and where you are also looking at, where the complexity lies from your point of view. Right?
Peter Neuling: Yes, of course, the human being is far from gone because of that.
Jens: Exactly. And then the question: Where is the limit of complexity at some point? The controllability of complexity when other aspects play a role. For example, I am now thinking quite bluntly in the direction of intralogistics in picking robots or sorting in conveyor technology, in perhaps also between buffers, with shuttle vehicles with robot-driven storage systems and so on and so forth. If I always expand, expand, expand these AGVs by one more component, is there an end somewhere because of the complexity? Or are you set up and structured in such a way that it is not so important how many different types are considered, but is the result always optimisable?
Wolfgang Hackenberg: That is exactly where our expertise lies. It has to be said clearly. We have the algorithms that are in the background, we have developed them over years and have brought them together with very powerful computing technology, i.e. with cloud technology, which we can use at the location, but also in our own cloud. There are actually no more limitations. That has to be said quite clearly. That's, but also because the data we're talking about, that's data, that's status messages, that's events, there are battery levels and all these things that we have, or position data. We are not talking about any sensor data, any images that are transported through the area, but there are very, very many messages, millions of events, but they are very small in scope and that's why we don't see any limitations there for the time being. It's more a question of how you build up this algorithm so that it processes it optimally. How do you switch the whole thing on and off? At a later point in time, which is not the focus this year, we also want to incorporate machines, i.e. really machine occupancy planning, in order to be able to offer an overall optimum for a finished location. But this year it's robots, forklifts, tugger trains and people, and we can do that without any problems in the scopes that we know. We often simulate scopes of 600 vehicles, 12,000 orders and that is no problem at all.
Thomas: Keyword simulation. That's exactly where I wanted to hook in. I looked around a bit on your new YouTube video and the topic of simulation came up, and that you actually carry out continuous optimisation. Looking ahead to find the best routes. Does that add to the complexity? So that you don't just control the orders as they come in or how they have to be processed because of some transport orders and so on. So that in addition you are always looking further into the future or the question: How far do you look into the future at all? That you are still looking into it, looking ahead, in order to control the entire system. I think that's also a crazy complexity that you have there.
Wolfgang Hackenberg: I would like to break it down into two parts. We can simulate, we can emulate. That means that if a customer says, for example, this is my layout and I want to process the orders here and the orders there, then we can simulate that in advance with our software. We say you need so and so many vehicles and so on. We can do that and then we have, that was one point and the other point is that we always look ahead in the planning. But that is not so much a simulation. We do it with heuristics, with metaheuristics. We always look at how to parameterise the whole thing, and that depends on a customer scenario. But we always look half an hour, 10 minutes, an hour, two hours ahead and that helps us.
Thomas: So far?
Wolfgang Hackenberg: Yes, that helps us to plan the resources so well. Let me give you a banal example: you have a large hall and you have an order and you know that you need an AGV to pick something up in ten minutes. Today's systems, which somehow plan very sequentially, which plan double jumps, then start and drive through the hall for 10 minutes. But we know that in 10 minutes and 15 seconds another AGV will be available next to it and that's why I don't need this whole movement. I look into the future and know that something optimal will come out of it, and we do that hundreds of thousands of times every second.
Thomas: I think that's a very exciting approach. I also think it's exactly the right approach to avoid things. And then to give it priority by avoiding transport on the spot or driving. Let me ask you a straightforward question: How do you manage to create this added value, which is clearly imaginable? How do you manage to turn this added value into business cases? I mean Peter, who said earlier that this is somewhere more or less without alternative or is absolutely necessary, the question doesn't even arise. Nevertheless, you will often be confronted with the question: Is it worth it? And if so, from when or in what period of time? Is there an ROI? Is there no ROI? How do you do that with such a system, which depends a lot on optimising the actual state through very, very clever calculations? How can you tell where the added value really is before you even use it?
Wolfgang Hackenberg: I'll jump in, Peter, then you can add to it afterwards. There are two developments: The market there is currently undergoing a major upheaval. You can feel that, and of course we can feel it, too. And it has to be said that the automotive industry, where we initially come from, is already a bit further ahead. They've had these autonomous systems for many years and they're now experiencing a pain that is increasing in logistics, across the board. What we also notice in many cases with logistics companies is that they have the same pain, which is a very simple lever. You have vehicles from manufacturer A, and now you have to add more vehicles. They may not have the capabilities of the other vehicles you need and they cannot drive on the same layout, because today's systems can actually only drive with a back-end control centre behind them. They can't drive on the same routes because they don't see each other. That is a very simple lever. You save on additional control centres if you have our universal control centre with which you can do everything. Now there is another development: not only robots are driving around, but you also have forklift trucks driving around and they often drive into each other. A forklift truck drives in and closes a lane and the AGV can no longer get through. There is some kind of deadlock and so on, and we can then also optimize that in an integrated way. That's a whole new category again. Imagine you have a warehouse with 30 robots and you can just throw 20 forklifts in there during Christmas and take them out again in January without anything happening. The orders are distributed and taken back seamlessly and that's the vision. We simply see that these are the points of contact. Then, of course, we also have points where we can say that through this optimisation we can also reduce the number of robots that are needed by 20 to 30 per cent. We can significantly reduce the travel movement, as I just explained an example. You can expand the layout much better, you have much less traffic and these are all things that have to be brought together. Depending on how the customer is positioned, we naturally present these added values.
Thomas: I would like to jump right in again, also what we just said about added value and optimisation of the systems and so on and so forth. Let me put it this way: If you only orchestrate the AGVs and forklifts etc. on the one hand, that's one thing, but I still have the human factor of influence in this overall structure of the warehouse or production. You said earlier that you plan an hour ahead or even ten minutes ahead. But now the human being comes into this whole game and does something that doesn't really fit into the simulation or emulation. For example, suddenly the human moves a pallet by half a metre or the human says: OK, I'll bring the order forward. Then this structure actually collapses again, or is everything already covered in this complexity?
Wolfgang Hackenberg: Well, the thing is, our system is not designed for that. That is, humans can intervene, but it is not designed for humans to prioritise around it all the time, but there are desired delivery times, desired points and we calculate new solutions thousands of times per second. You have to think of it like this: A supplier is late, a plant breaks down and an AGV breaks down. Something happens and we have already calculated a new solution.
Thomas: Yes, okay.
Wolfgang Hackenberg: It's all in there already.
Thomas: This is madness.
Wolfgang Hackenberg: Of course there are situations when a shelf falls over and three routes are closed, then there's not much you can do, that's clear, and then people have to get involved. We certainly don't want to take people out of the equation completely, but the idea is that that's kind of the point. When humans have to find solutions in this complexity, they only find permissible solutions because the computing power between our two ears is limited. We don't want permissible solutions based on experience. We want optimal solutions, and we want them every second. That is our starting point.
Jens: That's an exciting starting point. Peter, can you maybe talk a little bit about the secrets? What benefits did you see directly in there? And which ones have actually materialised? Can you confirm or add some of them?
Peter Neuling: Yes, I'll start in a different place: First of all, I can confirm that. The impetus in our house was that we wanted to free ourselves from the dependence on AGV manufacturers to a certain extent. Today, as Wolfgang has already said, if I have a manufacturer today, I actually have to take the same manufacturer again, because two master controllers next to each other and two different vehicles on one line simply don't work. The vehicles don't recognise each other, which is why there are always obstructions. It doesn't have to be an accident, but the vehicles are facing each other and simply don't move because they don't know anything about each other. This means that at some point the AGV manufacturer, who is at some location, dictates the price to us, and of course we don't want that either. That was part of the motivation. Then, in the first project we did, it also became apparent that there were situations in the degree of automation that could simply no longer be handled automatically because there were simply too many AGVs on the road. Through dynamic route guidance and the functionalities that we create the optimum through these many calculated steps that we get per second, we were able to regulate the traffic flow in the hall down to such an extent that we were able to drive a hall where we would not have been able to drive with AGVs at all in the first simulation. In this respect, this is really a valuable point for us.
Jens: That is very, very exciting.
Peter Neuling: And of course the whole thing pays off for us. In the end, less driving means a few fewer AGVs. Of course, that also brings money back into the coffers to buy good control systems.
Jens: I find that really hard to imagine. But what was the reason that was fixed by the optimisation, why an area that was not passable suddenly became passable?
Peter Neuling: First of all, quite bluntly, the amount of AGVs, because there were just so many. We have our factories and some of them are quite old and were not built for automatic transport. The roads are perhaps a bit narrower. Some of them have pedestrians on the road. You have to take into account that the roads are narrow and so on. We also have many crossroads. At an intersection, if there are AGVs coming from three sides, then of course I always have obstructions somewhere. Then there are standards that say that when the vehicle has stopped, it has to wait for two seconds, then it has to ring, then it has to light up, then it has to drive slowly and only then is it allowed to start again. Of course, that all hinders us. But if I have a control system that is capable of not bringing the vehicle to a standstill, but rather of letting it drive very slowly, I save those few seconds of time when the vehicle would otherwise have to stop, because that's the way the standard prescribes it today. We can get around all that because it has the appropriate intelligence, and that's why it really helps us.
Jens: Peter. How difficult was it to integrate this system into your existing IT and hardware landscape? Is it something where you need someone or a function like Group IT with a lot of know-how, with a lot of process and IT understanding or does it work so easily that you could also say: Any small or medium-sized company could adapt this for itself and use it.
Peter Neuling: Now, of course, we have the approach with us, we want to use it throughout the group and accordingly our step is that we also go into a cloud solution. We are working with a well-known cloud provider to build the so-called digital production platform. This system will then run on it and we can easily make it available to the locations. Wolfgang can perhaps say something more about what they have planned for themselves. There are also plans to offer every company the same size as ours without having to host it ourselves.
Wolfgang Hackenberg: Exactly, maybe I'll jump in, because this is very important now. At SYNAOS, we really do adapt to our customers' needs. Today, as I said, we come from the OEM sector. These are very large companies that often have very large fleets and, as I said earlier, have the skills to implement such things. Maybe Volkswagen will have a cloud now, while others will have to install something at their location. But the development we are seeing is clearly that we are bringing the system into a cloud environment in order to have these scaling options and this computing power. We also want to offer this to smaller customers via Software as a Service. So it's the solution that we have. It is just as effective or has many advantages for small customers and not only for large customers. We are in the process of expanding this step by step.
Thomas: I find this keyword of bringing together different manufacturers, different AGVs, very, very exciting. We experience it again and again in the world of logistics or warehousing, where it is incredibly difficult to bring together different manufacturers, including manufacturers of materials handling equipment, and to coordinate them with each other. So there is always a never-ending game of ping-pong. If you have created this with SYNAOS, that is of course a huge advantage. And that's where I'll be interested again now. Peter, you said earlier that you have something like 200 AGVs that he controls with it. How many different manufacturers do you actually have that you bring together? I guess a fleet of 200 or 200+ AGVs, you didn't buy or acquire them all at once, but they have been added over the years and you will definitely have different manufacturers. But how many are we talking about there? I would be interested to know, also to get a feeling for what is actually feasible?
Peter Neuling: At this point there are actually only two.
Jens: Oh, that's extremely disappointing.
Peter Neuling: Well, I'll put that into perspective a bit and it's also a new project. Of course, it only works if the AGVs we use can handle this VDR 5050, which is not that old yet and the existing projects very often cannot handle it. There are now the first manufacturers who go so far as to convert existing projects, but that is not the standard. As I said, we have two of them and the challenge for us is that we have under-run AGVs. The other is a large tractor unit from a well-known German manufacturer, which a company near Hanover has converted, automated and which then pulls up to three trailers with a towing transport. The processes are also completely different, the size is completely different, and that is simply what makes it so attractive for us.
Jens: Wolfgang, I would like to ask you, in order for you to work, or for the idea to work in general. You need at least two parties: the user or the customer who is really your customer, but you also need a certain openness of the corresponding interface to the various AGV manufacturers. We have already discussed VDA 5050, which you have already mentioned here. But the fact is that there is a lot of money to be made, especially in the AGV sector, the robotics sector. Not at all through the initial investment or through leasing the hardware, but mainly through software licences. How does that work, or how does the interaction with hardware suppliers and technology suppliers work? If this more or less permanent software update and also permanent software provision is often a not exactly small source of income, are you meeting open doors or relatively thick walls?
Wolfgang Hackenberg: I would say that we are meeting with open doors. I have to elaborate a bit, but there is simply a change in the industry. One of our core beliefs is that there is a separation of hardware and software, that is, hardware and hardware software, and real backend software that runs in a data centre in a cloud. Of course, it's easier for the development of a company at first. I have to say that clearly. There are a few examples. If we had also built hardware or somehow bought it in, then you would have integrated it straight away. But then you run up against the barrier that you can't optimize the entire system, that you create another black box, another island. Above all, you cannot achieve this optimum as a whole, and that is exactly what we want. We want to put this customer benefit in the foreground and we have clearly decided in favour of this. Based on the consideration that you simply have to concentrate on your strengths. There are many good hardware manufacturers in the industry today who have good hardware, who have very good software, but who have simply built something in the back end, i.e. in the optimisation software, over 10 or 20 years, which is also okay, but which does not fit into this world of the future that we now see coming. Let me compare, look at the consumer sector.
Thomas: Yes, I know. You mean.
Wolfgang Hackenberg: Everyone tells me Apple. Apple is really the only company that manages to be premium on the hardware and software side. But we believe that what has happened in the consumer sector will happen in the broadest sense, namely the separation of hardware and software via platforms, interfaces, and so on. This will happen and everyone will have to concentrate on what they can do. We are a top of class software company and that is where we have our strength.
Thomas: I still find it interesting. You just said yourself that the companies have all been developing in this direction for 10, 15, 20 years, of course in their limited field of vision of their products, but nevertheless. Nevertheless, I imagine it would be difficult - that's why I asked the question again - to say: Watch out, you'd better make hardware, the software, we can do that much better and then everyone says, Oh yes, you're right. Then I won't sell it any more. That's right, then I'll only sell the hardware. I think that's a bit too easy.
Wolfgang Hackenberg: Sometimes the really good solutions are also really simple, because...
Thomas: Oh yes, I want to adapt complexity.
Wolfgang Hackenberg: Yes, but of course it's not like everyone is saying that now. Yes, thank God. We have been waiting for you for 20 years. But they also realise that they simply don't have any more solutions. Of course, there are sometimes hardware-independent solutions. Then you put in some AGV from another manufacturer. But our solution is really developed per DNA, per default, hardware independent. That means traffic management - hardware independent, energy management - independent. That's a different class. And I think what is already an issue now. There are extremely many manufacturers out there and they have to differentiate themselves. If they come together with a very, very good baking solution, then that's also a strong proposition for the customer. There are many systems that simply bring both and would take them out of the market completely. We leave, let's say, from a cake of a revenue that such a manufacturer has with commissioning, maintenance, hardware, hardware-related software the whole localisation. We leave all that to you. In addition, we always offer really premium connection points to our ecosystem, to our operating system, where these can create additional added value for the customer, for example, by going into predictive maintenance and offering an additional service and so on. So that's already an offer that many companies find really interesting. We are approached by companies to ask whether they can adjust their software and enter into any kind of licence agreements with us on a permanent basis. There is really movement in the market, I have to say.
Thomas: Do you also see a trend towards this becoming more and more separate? So hardware and software solution?
Wolfgang Hackenberg: Yes, so sorry, for me this is an absolute trend and integrated systems will certainly survive somewhere. These large manufacturers, however, are more of a black box, partially fenced in. Of course there will be a lot of systems there, but on a broad scale there will be. I mean, we have 80 men and they make one software. Many companies have 30 men and they do projects, software, all mixed up. That's just, that's a development, you're just not going to roll that out. And now this deployment, this delivery to customers will be much easier through this platform, as Peter just said. There's a really big wave going on.
Thomas: First and foremost, I'm thinking about the complexity itself, which we've been talking about for almost half an hour now. I've been asking myself the question for a while: What kind of computing power do I actually need? You also mentioned cloud solutions and so on. I think you have to have the corresponding know-how. I would say that even a hardware manufacturer has his know-how somewhere in terms of conveyor technology or AGVs and so on. Of course, they would also have to provide the corresponding know-how on the software side or on the computing power side. So I would be interested to know again, especially with regard to this computing power, do I actually need anything as an end user? Do I have to set up a basement full of servers or something? Or does it all run on some kind of cloud? The keyword "Software as a Service" came up earlier.
Wolfgang Hackenberg: I would say that this whole issue of "do I have enough computing power" is solved for me today. If you think about it, a mobile phone has more computing power than a lunar module used to. So the topic is no longer topical.
Wolfgang Hackenberg: The question is of course, so how can you incorporate that, how can you integrate that? That is definitely a legitimate question and there are large companies, as we just heard from Peter from Volkswagen. They have their own cloud environments, they have their own online data centres, and you can bring that in.
Jens: If I may dig in for a moment. There are these big companies that have it all. But there are also the medium-sized companies or even smaller ones, the smaller SMEs, which may not have it and which may then shy away from it and say: No, I have to have server farms and back and forth.
Wolfgang Hackenberg: No one has to buy server farms. This is exactly the development that I have already described. We are also developing offers, both in terms of pricing models, but also in terms of technology, so that we can really offer SaaS. That we can say: You can really operate it from a central cloud that we provide, that you buy somewhere. And then that is also really attractive for smaller manufacturers. You don't need 200 vehicles to have this, but if you have five vehicles from one and five vehicles from another and then three forklift trucks, then it can be worthwhile. Then maybe that can be the right thing for you.
Jens: I also believe that this is an issue. But it's still good that we're addressing it, that we agree in our group, because it's something that, if you're not so familiar with the subject, you might be put off by the complexity. Based on historical experience, I'll call it that.
Wolfgang Hackenberg: Please put it in there again, because that is very important to me. We think from the customer's point of view, our focus is to make things as simple as possible for the customer. The customer should not have to go to any expense, any server farms. I'm going to use the word "have to buy"; instead, they should be able to book this service. The customer should want to be able to operate all this software easily. That means we are currently developing standard interfaces in the direction of WMS, in the direction of EAP software. That it is simple. We calculate the optimal solution in this system, but the customer can also intervene there himself and make changes. That doesn't necessarily mean that he now prioritises any orders through the area, but he can also adapt layouts etc.. We believe it is really a system where the customer can interact with it much, much better. That's why we have also invested an incredible amount in the user experience, in ease of use, in new interfaces. Why does new software that runs in industry always have to look like it runs in industry?
Thomas: Definitely! It's so crass when I think about how some WMS interfaces still look, like Windows NT for example. It's a complete disaster. I also don't understand why it's always, it's almost assumed. It has to look as boring as possible and as annoying as possible. But I think it's really good.
Wolfgang Hackenberg: They simply don't get anywhere. It's simple because the back end is not set up that way. Then you have to completely rebuild the software and that costs tens of millions of euros. Once you have developed such a thing. Who does that today?
Thomas: Yes, you have to. I think I would like to say again...
Wolfgang Hackenberg: Better than before.
Thomas: Yes, no, I'd like to throw another 5 euros into the phrase bank and say: If you're late, life punishes you. Wolfgang, maybe you could go into a bit more detail about the whole topic in relation to hardware. Do you have special partners and who are your hardware partners, for example, or how does the whole issue work in this regard?
Wolfgang Hackenberg: I briefly mentioned this separation of hardware and software earlier. Of course, our approach is based on finding strong hardware partners. Peter has already said that there is a movement towards this. I believe that more and more companies are concentrating on this, also on this new VDA 5050 interface, which allows this intersection between backend, software and hardware. We already have some very strong partners with whom we work. For example, MLR Systems, Tünkers, KUKA, Linde Material Handling, both for AGVs and for fork-lift trucks, which are automated in tugger trains, for example in Hanover with Götting. I can say here today that we are now publicising a very, very strong partnership with SEW Eurodrive from Bruchsal, one of the leading suppliers of drive technology for materials handling and industrial automation. They have great vehicles, a strong software stack and together we want to create a premium offer for our customers. Truly turnkey solutions from a single source, hardware, software and project management, is another step. We are also talking to many other companies who are interested in the partnership. And I think that this strong ecosystem that we want to build there, with more strong partners like SEW. This is growing and growing and we have more requests than we can satisfy.
Thomas: I would like to talk again about the special constellation in this round. I would be interested to know more, especially a little bit, after we have talked about the advantages in general. In general, I would be interested in your interaction with Peter and Wolfgang in the project, which was more or less created together. What were the lessons you learned from it? What worked very well? What perhaps didn't work out so well? Where, you Wolfgang, also drew out lessons again, what you would do better next time? How did it go in general?
Peter Neuling: That would be the one big project that we are just about to start. Of course, it's still hard to say, but in this respect I can't really give an answer. We have certainly underestimated some things, where we said that a layout in the hall is not that difficult. But somehow we also found out that an AGV is not an AGV, but we had to look for a solution where you don't have a control centre or a network in the hall, but have to put a network in the hall for each type of AGV. We certainly underestimated that and also different functionalities. Of course, we are working together with the AGV manufacturers. They are very different. You can see that some are very open to the idea of a standard interface. But some still view it with a lot of suspicion, because they also make money with future business when we then want to change a card. In the past, we had to commission an AGV manufacturer to make these changes. Of course, we don't want to do that in the future, because they hope to make considerable amounts of money. We say that we would actually like to save that money. That is of course a solution or an issue for us. These are all things, challenges. We simply have to see how we deal with them in the future. As I said, the manufacturers have their very own view of things and also very, very different know-how. Some are very cooperative, some have a much harder time.
Wolfgang Hackenberg: But if we look at these projects now. Volkswagen is now a partner or a customer that we won very early on. Of course, now with a very large group of companies, and there is already a very close exchange. Peter and his team are often with us in these coordination meetings and we agree on what is important for him and what is not important. That is exactly the new world that we have. We don't stand still and say the system is now fully developed, now it will be rolled out, but we continue to develop. In every sprint, in every fortnight, we develop it a little further. We listen to what our customers need. Now our first e-commerce customer has different requirements than an OEM from the automotive industry. We build in the formulas piece by piece and thus continue to develop this operating system. You're supposed to, of course that comes from the fact that card formats are not good and updates are not good and so on. These are all things that you learn. But overall I think the cooperation is very, very good.
Thomas: I find it generally, to sum it up again. In general, it's exciting how simple this system actually seems when you talk about it, but how much complexity is transported through it, is processed and also transformed into something proactive and something good. I find that very exciting. I also find it very interesting. At the end of the day, we have what lessons learned you have to deal with at that point. Somewhere on one side we are talking about hundreds, thousands, hundreds of thousands of algorithms that we calculate. At the end of the day, one thing where you were a little bit hacky, it was still a little bit good.
Wolfgang Hackenberg: You described it very well. Of course, you still have the project in there, but that is exactly the idea behind it. You need a system that is simple in its interfaces, simple in its operation and all the complexity is simply under the bonnet, is also processed to some extent by the system and used optimally.
Thomas: Yes, of course, that's how it has to be.
Wolfgang Hackenberg: That's how it has to be. Today there are often interfaces and people are simply exposed to complexity without restraint. That is what leads to excessive demands.
Thomas: Above all, there is no point in mapping this complexity in hardware, for example, so that you permanently have some very crass special AGVs for every single mini-step. The complexity must be as good as possible - which you also described once - with your separation of hardware vs. software, actually the complexity and flexibility must be provided via software. That has to be the future step. What you describe at SYNAOS is exactly what the right approach is.
Peter Neuling: What is also important from our point of view as users. It is simply the point. We know AGV projects today. That is always a project. That stretches over a very long period of time and the future for us, so from our perspective. We have a master control and we just buy AGVs in the market and the whole thing has to be so simple afterwards. I buy the AGVs, they are put in front of my door. Then I can activate them somehow and an hour later they are running. That's where we want to go. I think that is an important step. Perhaps especially for smaller companies that want to start with AGVs, this is certainly also a major obstacle. As I said, our approach is: it has to be simple. What is also important is that IT is playing an increasingly important role today. We notice that as a group IT. In the past, it wasn't such a drama, because you could literally put a server under the table. Then an AGV or an FTS course would run. Of course, that is far from being the case today. We have open source. We have the requirements, security requirements, which are gigantic. "Hacking" is in a newspaper somewhere every day. These are things that you have to deal with. Of course, that is also an experience we had to gather.
Thomas: Yes, also just what you say, that simply has to be. What we were also talking about earlier, how can we take away the users' "fears", so to speak, that they don't necessarily have to put server farms somewhere and so on. I thought it was very, very good, especially what you said, that Wolfgang thinks from the customer's point of view and that you take the idea on board and don't just lock yourself up in your office and say, "So, we're the great programmers here, we can do everything and we programme the greatest things here. Afterwards, the user sits in front of it and says: Yes, and what do I have to do now? Which buttons do I have to press now and so on. I think that's very, very good what you're saying: simple front end, complexity can hang inside without end and so on. I think that's very, very good at this point.
Jens: Yes, I think so too. Thank you again Peter at this point for the closing words you just gave us. That was again very important to show: Okay, we are not talking about a castle in the air here. Yes, where a certain limited number of OEMs have the opportunity to participate in something like this, but that it is precisely also a use case for companies that have possibly not yet had the opportunity to participate in such technologies. I think that was also important to conclude. Against this background, thank you for being our guest this time and I hope you enjoyed it as much as we did. Hopefully you were able to take away a little bit of our questions, just as we were able to take away a little bit more of your input. So thank you for that.
Wolfgang Hackenberg: Absolutely, thank you very much!
Peter Neuling: Thanks too, it was exciting. It was interesting.