Albrecht Ricken: 0:04

installing these things may take two or three weeks until you have it in the test system until you started initially testing. But then adding a robot from any vendor is really easy. It is just connecting it to the cloud. And then it is available as a resource in extended warehouse management.

Tom Raftery: 0:28

Good morning, good afternoon or good evening wherever you are in the world. This is the digital supply chain podcast, the number one podcast focusing on the digitization of supply chain. And I'm your host, global vice president of SAP. Tom Raftery. Hi, everyone. Welcome to the digital supply chain podcast. My name is Tom Raftery with SAP and with me on the show today I have my special guest calibrate. Right, would you like to introduce yourself?

Albrecht Ricken: 0:55

Thanks, Tom. I'm a program manager for robotics at SAP. And today we I'm going to share some details on the new integration of robots that we have for extended warehouse management.

Tom Raftery: 1:13

Lovely, great. And Albreact, let's let's take a step back and put a bit of context on this first, robots in warehouses are becoming more common, why?

Albrecht Ricken: 1:26

There's a new class of human safe collaborative robots that is now available from hundreds of low cost vendors. And what they do is they defy the traditional rule that automation pays off at scale only. This used to be true for industrial robots. But with this new class of collaborative robots often dubbed as cobots. The rule changed, they make automation affordable, because they can sense the environment and they gain a semantic understanding of what is around them. This means they can work side by side with humans. And when they encounter a human that can brake, they can go around it the way they can do it with any obstacle that they find. And in addition, it makes them useful for one off tasks because they come with more advanced programming environments. Yet, challenges remain. And this is particularly true if we think of integrating such warehouse robots that carry things around in warehouses. When we integrate these robots into a warehouse management system. There, you encounter expensive system and process integration with high upfront costs. So system integration is bespoke time and material work that traditionally results in expensive and highly inflexible solutions when you change over from one robot vendor to another, or even when you just want to add robots. Another integration project easily starts. The other thing that vendor lock in, because vertically integrated offerings provide more generalized solutions with high dependency, but very limited flexibility. And finally, you need to think about IT strategy and governance. What happens here is that every robot vendor comes with a stack. And this is this means another thing IT needs to maintain IT needs to govern, you need to create again a set of users, you need to think of security, what happens if somebody just enters for instance, a USB stick into one of those robots with a virus on it does your security concept cover this? So it needs to be part of it. And the work multiplies with a number of robot vendors you employ. And this is why we came up with Warehouse Robotics, which provides a new way to integrate robots from very different vendors in one single solution. And what we made sure is you can do it it is a plug and play robot integration, a setup that takes about three weeks, four weeks maybe for the initial setup, and that allows once it is set up to add new robots on the fly and within minutes, and this is a tremendously different because a solution like this can easily scale. Think of events like Cyber Monday, Black Friday or just recently COVID-19 we have a wholesaler as a customer whose sale spiked when COVID-19 arrived because they are a wholesaler for IT equipment. And they have to cope with increasing demand. And how do you do this, when you have limited capacity in the warehouse?

Tom Raftery: 5:10

Okay, so let me walk that back a little bit first, to make sure that I that I understand it, essentially, these new cobots that have come on the scene are better and cheaper than existing ones because they can sense their environment around them. And consequently, they are safer to work with. And they're cheaper for people to buy. So people are buying more of them. But the problem with them today is that it takes time to and money to integrate them into the existing IT landscape that they are walking into, are rolling into in a warehouse, right? So can you give me an example of you know, how long or how much it would cost to integrate, you know, current cobots into an existing landscape.

Albrecht Ricken: 6:12

We asked customers, and they told us anything between half a year and a year and a half depending on the size of the warehouse, depending on the number of warehouses they run. So this is a substantial expense. And what on top off this is that once you have these robots, and you have them possibly from different vendors, what you need to do is maintain all these stacks, which puts another burden on your IT departments.

Tom Raftery: 6:44

So just just to clarify that, what you're saying is each one comes with its own hardware and software stack that is proprietary and needs to be maintained. So if they do a software upgrade on one cobot from one vendor, you have to then make sure that you roll that out and that it works OK, for that series of cobots that you might have from that vendor. But another if you have another series of cobots from another vendor that hasn't rolled out an update or has rolled out a different update, it's you you've got to manage those. And then if you get a third vendor supplying cobots, it becomes even more complex again, is that it?

Albrecht Ricken: 7:24

Yes. And the reason to have different robot vendors is because of these collaborative robots are highly specialized. So there are some who can pick items only of a certain size. There are others who have a limited payload, or can work in more punishing environments, on certain types of floors, but not on other types of floors. All these problems and challenges may drive you to procure robots from different sources. But then, again, how do you do the integration? So this is like a blocker. And this is one of the problems we try to solve. And in Warehouse Robotics, we allow heterogeneous fleets to be managed in a single stack.

Tom Raftery: 8:20

Okay, so when you say Warehouse Robotics, you're referring to this new product that SAP rolling out, which allows us to, for our customers, add robots from multiple vendors. I won't say at the drop of a hat, but you're talking days or weeks to roll it out rather than weeks to months to years. Yeah?

Albrecht Ricken: 8:42

Yeah, the initial initial things that you need to do about extended warehouse management, you need to install an add on, you need. If you don't have yet you need a Cloud Connector. installing these things may take two or three weeks until you have it in this test system until you started initially testing. But then adding a robot from any vendor is really easy. It is just connecting it to the cloud. And then it is available as a resource in extended warehouse management. And it is up to a regular customizing task in AWS and extended warehouse management to assign the robot to tasks and by what rules this goes. This is standard and we don't change anything about extended warehouse management. So you may have different tasks by payload. You may have different tasks by what the robot can actually do. The size of boxes, you name it.

Tom Raftery: 9:47

Okay, interesting. Just just to get my head around it once it's once it is set up. How long does it take to add a cobot to the system typically?

Albrecht Ricken: 9:56

It it is a matter of minutes. Once you have set it up adding any robot and let me give you an example that we played through. assume you have a warehouse that has two different areas. One is variable sized boxes, small and big in one area. And the other one was just small load carriers. And initially, you have a robot from a vendor of, let's say, mobile industrial robots, Danish company or robot from by fetch Californian company. And it serves both areas. now all the sudden you encounter a peak demand, and you need an additional robots. Now, we are just entering an era era of robot as a service. So getting a robot in two or three years from now will be almost as easy as getting, as renting a car from a car rental company. So you call them can we get an additional robot for our warehouse. And they may tell you eh sorry, we don't have an additional mirror or fetch robot, whatever you have. But we have one that comes for instance, from magazine Oh, now magazine, oh is a Munich robot company. And the difference with the special thing about the magazine or robot is that it has arms, so it can pick parcels it can pick handling units, and then support and then unload them. The problem with a magazine or robot is it can do this only for small load carriers around shoebox sized items, okay, which makes them very popular with these Zolando and shoe makers of this world. Now, you just add these robots. And then warehouse management immediately assigns the magazine a robot to the smaller carriers which it can do fully automated, and the mirror robot to the other area, which always need somebody to load and unload it. And that brings us to the various scenarios that we support. We implemented so far three very popular scenarios. The one is cross docking, cross docking means you get a goods receipt at the warehouse. And rather than moving it on into the storage area, you move it right away to the goods issue area where it is put on another truck. This applies in environments where you have happen spike logistics, is something that very much the logistics service providers the DHL's, the UPS, the FedEx of this world do. Sure. And this is something that where a robot gets loaded, and they roll and a half on top of some kind of movable shelf. So it can be loaded into the goods receipt, like they load a shelf, it moves the shelf to the goods issue area, and there it is put onto the next truck, then put away the second scenario. This is the classic scenario where the robot moves an entire shelf into the warehouse. And the third one is a little complex. It is called robot unable to pick pack and pass. Now, anybody in warehousing knows what pick, pack and pass is it is we're a a cart, or also a conveyor belt across several areas in a warehouse, and each area has one person for loading and unloading. So it makes like a rendezvous in each of these areas with the individual pickers. And we can have a robot do this do exactly this. And why we do this the handheld devices anyway, integrated with the extended warehouse management and we do this for many years. This all harmonizes them into one single solution fully controlled by EWM.

Tom Raftery: 14:21

Obviously, in any of these discussions around robots, people are going to be wondering about, you know, warehouse workers and you know, are they going to all lose their jobs? Or what what what's your feeling on that with the advent of you know, lots more robots coming online, as you mentioned in the next couple of years, will warehouse jobs go away?

Albrecht Ricken: 15:18

The motivation for our first customer for Bechtle in Nekazoo had something to do with with warehouse workers but it was because they can't hire enough people. In Nekazoo where Bechtle is located. They have the worldwide headquarters of Lidl, a big retail chain. And they have one of the largest plants that Audi has a German carmaker. Then south of Mecca zone, they have a big Porsche plant. At Porsche they pay for warehouse workers double the tariff that has been agreed on the national level, because they simply cannot get enough people. And Bechtle is in the middle of all of them, and they don't know how to get how to hire sufficient people for their warehouse operations. So all the people in the warehouse, we're very much in favor of getting support from a robot. The the business side is beyond the viewpoint of the of the workers business side is that Bechtler has ambitious growth plans, and they knew to support these growth plans, they need more automation, not because of the cost simply because of the capacity.

Tom Raftery: 16:44

Okay. And I got to think as well that cobots are less likely to be spreading Coronavirus than people who are in close proximity to one another.

Albrecht Ricken: 16:54

By all accounts, but this this when they made this decision that was before Coronavirus, but you're right. That's, it does. That doesn't mean that working with robot does need some accommodation in the business. It actually does. Like they had to tell forklift drivers. to slow down when the robot comes, they had to modify the robot a bit so that it has now laser beams on the floor so that everybody can see it from far away. They have it beeping. All these are things to consider when having a robot. But these are the same things that any company needs to consider when acquiring a new machine. It is how does it work in the environment?

Tom Raftery: 17:43

robots with lasers that sounds like something from Skynet, or some kind of science fiction movie.

Albrecht Ricken: 17:51

It is, it sounds almost like Star Wars on the warehouse floor.

Tom Raftery: 17:59

Tell me something about the technology underlying this Albrecht

Albrecht Ricken: 18:02

The way we worked it out was that we that there is now a technology available that is called Kubernetes Kubernetes deploys, so called software containers. Now let me explain a bit where the container thing comes from. In the past, when you decided, let's say in the late 90s, you decided to install, let's say Acrobat Reader on your PC, you would insert the CD, and you would get an installation environment on Windows 98, let's say. And it would lead you through an installation procedure. And what it would do is, among other things, it would rename the variables provide a safe environment for your installation. So make sure that the program works on your particular PC. Now, when the cloud came up, there was of course, a question, how do we do it there. And then what started was first the concept of virtual machines. The problem with virtual machines is the size of every virtual machine because every virtual machine carries its separate operating system. So even if you have three virtual machines running on Linux, on a Linux machine, you're three times carry the operating system for each virtual machine plus the underlying operating system if you have one. And this is simply a question of economy. And then the containers came up. Now the containers are like a virtual machine but without the operating system. Well, that's not exactly true. But it comes close. And they leverage the underlying operating system. Now, how many operating systems are today there when compared from the late 90s. From the early 90s. Those days, we had Unix and we had all kinds of things. Today we have three we have Linux, we have Windows, and we have Mac OS, of course. And that's it. And with this, you can economically make that sure. Now there was one problem, and this is how do you distribute these containers if we know how to install them. And this is when actually Google invented coordinators, which is a Distribution engine. And now, the way we look at a collaborative robot as an autonomous and this is important here is by saying the autonomous robot is just a Kubernetes cluster, or like a server that does a very specific viewpoint. But what it does is, most of the autonomous robots, they, they come with pretty much the same basic stack, so they have Linux. And then they have on top of a system called Robotic Operating System, is an open source operating system developed about 15 years back by a company called Willow Garage in California. And I am not aware of any startup in robotics, not leveraging and not taking advantage of ROS. Okay. And this means that Linux, of course, we can deploy our software containers, and the beautiful thing is, it is always the same software container regardless of the of the robot make, okay. And that's really the secret sauce behind it. That allows us to do this. So we have a, an order manager somewhere in the clouds, the order manager deploys, or dispatches orders, to the robot, if and only if a robot pings for work. So the robot could decide I'm not going to ask for work, because I ran out of battery, because I am busy, because I'm working on a different order. And there is another interesting impact. And this is, we are used in the warehouse management interaction with subsystems that you need response times in the millisecond area. In our case, we download onto the robot into the robot container, or the warehouse order. This means the warehouse robot knows what it does. Now think, don't think of a warehouse in the glamorous brochures, but think of a real existing warehouse, where you have typically good Wi Fi and the goods receipt goods issue area, no doubt. But when it comes down to the back corners, poor Wi Fi signal, right, no wifi signal at all. Now, with that approach of containers on the robot, the whole warehouse order on the robot. The robot doesn't care, okay, it loses connectivity, yet it knows what to do. And when it's done, when it then regains connectivity, it can update the warehouse order into EWM. What what happens if there is a time delay of 60 seconds? Doesn't make a big difference. EWM will not collapse from that and the robot won't collapse. And this is something that traditional robots integrations can't do.

Tom Raftery: 23:47

we're coming towards the end of the podcast Now, is there anything I haven't asked you that you feel I should have any topic that we've not covered off that you think it's important for people to be aware of

Albrecht Ricken: 23:59

the Warehouse Robotics today is open source for anybody who would like to host by themselves. But what we are planning is to come up with a hosted offering, where we take care of all those things, in particular for customers who say, somebody needs to assume the responsibility for the Kubernetes clusters and all those kinds of things. And we got a lot of requests around this, say, Don't make us look into these technologies. We want to run our warehouse, we don't want to do technology. yet. Anybody interested in downloading the code. It is on the GitHub, anybody can use it, take advantage of it. But wait until possibly the end of the year. And you may expect a good offering as a service.

Tom Raftery: 24:55

Nice. Nice. Albrecht if people want to know more about yourself or about the Warehouse Robotics or any of the things we've discussed on the podcast today, where would you have me direct them?

Albrecht Ricken: 25:08

I can send you a link.

Tom Raftery: 25:11

Lovely and about yourself.

Albrecht Ricken: 25:13

The same thing, they can do they can get in touch with me no big deal. Yeah. albrecht dot ricken at sap.com happy to reply to questions from either interested customers or from consultants who want to take advantage of this from out partners. Absolutely.

Tom Raftery: 25:35

Superb, super Albrecht. That's been great. Thanks a million for coming on the podcast today

Albrecht Ricken: 25:40

was such a pleasure.

Tom Raftery: 25:42

Okay, we've come to the end of the show. Thanks everyone for listening. If you'd like to know more about digital supply chains, head on over to sa p.com slash digital supply chain or, or simply drop me an email to Tom raftery@sap.com. If you'd like to show, please don't forget to subscribe to it and your podcast application of choice to get new episodes as soon as they're published. Also, please don't forget to rate and review the podcast. It really does help new people to find the show. Thanks. Catch you all next time.