Masterclass #3: Harnessing an OCPP-Broker for Flexibility

Welcome everybody to today's masterclass. Welcome everybody to today'smasterclass. Take your seat, make yourself comfortable, take a cup of coffee. We'll start in 45 seconds. And welcome to the third installment of our service of for master classes about OCPP broker architecture is great to have you all here. In our last session, we took a deep dive into the architecture of the CPP broker exploring how to give CPO the flexibility to scale the network without vendor locking. Today's sessions is all about unlocking new opportunities for CPO's from AI driven, predictive maintenance to that I make load management and energy flexibility services. We'll show you how to unlock new opportunities for CPO's. And how the OCPP broker acts as a bridge to these advanced services, making your operations smarter, more efficient and future proof. Professor Robert Brame, CTO of FlexiCharge, will be giving the masterclass. And in today's session, he is joined by another leading profile from the industry, Stefan Herr, CEO of Evailable, formerly known as Evercharge. Welcome Stefan. We have a great session ahead of us with a lot of practical strategies and real world applications. So let's dive into this. Robert, the floor is yours. Yeah, thanks a lot, Chris. And also welcome from my side to everybody to our third masterclass. I think we had two first or two classes. Quite, quite interesting. I will recap a bit on that. And then, as Chris said, today, we are going to dive into how to integrate third party solutions and in part particular, also one service, software as a service solution, the one from Available. And I'm very happy that we have Stefan Herr here. He will present about Available. And I will give you a bit of an intro on how we actually integrate with software as a service solutions in general, but then also in particular with Available. So I will quickly share my screen. Am I sharing already? No. Yes. Is the screen available, Chris? Everybody can see it? Yes. Yes. Yeah. So I said the agenda for today is do a quick recap on what we have learned and looked into in our first and our second masterclass. Integration of third party solutions. And then we'll have an introduction from Stefan about Available. So in our first masterclass, we discussed a bit the current challenges for ChargePoint operators to technical challenges with respect to how to orchestrate systems and technologies and how to set up a future proof technology stack. Which allows, for example, for easy integration of third party solutions, but also being able to orchestrate different ChargePoint management systems. We see a lot of CPOs in our customer base, which are actually doing that, especially international players, but also using CPMS systems for different applications, whether it's public charging or destination charging or even apartment building charging. So there's different needs on the CPMS and often they run multiple CPMS systems in their technology stack. And then on the top part here, there's often more and more needs now for integration of system operators or network operators and different levels on the CPMS side, but often also locally. So we discussed in the first masterclass how broker can actually help to overcome some of these challenges and try to reach the KPIs, which are currently important for CPOs, which is cost revenue optimization, scaling, rollout, site planning, operations, maintenance, support, user experience, excellence, security, innovation, of course, how to differentiate from others, internationalization, as I said, specifically with CPMS, different CPMS systems, different rules and regulations from the DSO, DNOs, and then obviously also load energy management to overcome grid congestion problems, limited grid connection, availability, integration of batteries, and so on and so forth. And the key to sort of, or one of the keys to solve these challenges is data. And therefore, a data broker or general OCPP broker can help to actually solve these challenges and overcome on the basis of data and make decisions for strategic decisions or improve operations and so on and so forth. So in the first masterclass, we've sort of derived this more future orientated technology stack based on the standard OCPP technology stack, which everybody knows, where you have the EV on the left-hand side, you have the EVSE. And then typically, you do not have the broker in the middle, you have the EVSE, so the charging stations, the hardware in the field directly connected to CPMS. And then from the CPMS, you have the integration to clearinghouses for roaming, e-mobility service providers, and so on and so forth. DSO connections to different sort of nodes here in that architecture, either directly to the hardware or through the CPMS. And obviously, a broker, a data broker will help CPOs to become a bit more independent and sort of build in a layer which abstracts the CPMS, but also abstracts the EVSE, so the charging stations. And basically provide a tool for easy integration and orchestration of these different systems and different challenges. And one of these challenges we are going to look into today is how to integrate, for example, systems to enhance operational excellence based on artificial intelligence. Stefan will talk about that. But on the other side, this is where we come from. I mean, at the end of the day, FlexiCharge provides a software as a service solution as well. Load energy management, grid integration, grid services, frequency containment reserve with chargers and so on and so forth. And we are doing that through our own data broker. And we have been doing that for a long time since the beginning. So we always wanted to provide an open platform, which is independent of proprietary integrations via Modbus, for example, into the charging station. So an open platform. And we have decided at the very beginning with FlexiCharge that we're going to provide the platform or integrate with the charger hardware via OCPP, because OCPP is the protocol which is made to basically communicate with chargers. So we always had a broker as part of our general architecture. And to that broker, we have connected our load and energy management services, VPP aggregation and so on and so forth. We talked about that in the last masterclass. And we have decided now in the last summer that we'll make that broker available for customers which want to go with FlexiCharge, which are looking for solutions, primary load energy management, solutions, and then get the broker as an extra. So what we can provide with the broker is obviously, or what we're doing or what we have been done the whole time always is obviously to provide the link between the charging stations, the hardware and the CPMS. So that's a completely transparent OCPP proxy, which sort of gates the data from the chargers to the CPMS and from the CPMS to the charging stations. It's also connection transparent. That should not be the topic today. So data goes through the broker. And since the data goes through the broker, we can sort of fork it to other software as a service solutions. We are doing that with our own solutions, with our load and energy management system. But we can also gate the data into data warehouses from our customers. So if they want to build their own data lakes, they can get the data. But we can also forward it to third party solutions, software as a service solutions like available. So the broker was always there. So it's a solid solution which we have developed over years. And we have been using since a long time. And most of our customers, like 95, 98% are basically using the broker as part of our load energy management solutions. So how does it work? We talked about, we had a deep dive last masterclass where we looked a bit into the specifics of our architecture and also how we differentiate from competitors. Just to recap that a bit. So we are having these, what we call microservice stacks. We run a uprising model of integration美其 ssusli uh load in energy management services which then consumes the data make decisions and themselves will also publish information which should be sent to the charger and all these services they are encapsulated in self-contained microservice stacks so we group charges behind the grid connection point into these microservices so for each site for each infrastructure for each group of chargers behind the grid connection point we are operating one microservice stack and that stack can run in the cloud so on our server cluster but it can also run locally on our gateway connect if it runs locally then we can integrate local power meters we can integrate batteries we can integrate pv panels and then orchestrate yeah the load management and the charging at at the site a central element of these microservice stack is our local message broker here which exchanges information or it's the message delivery service for exchanging information between these individual microservices operating here or running inside the stack but the local message broker is also connected to a central message broker yeah so as i said each microservice stack basically operates a site between the grid connection point for which we provide load or energy management and all the sites and all these stacks and the local message broker they are connected to a central message broker and that central message broker we call um our ocpp message or data broker um it cannot only um provide data from ocpp um message flow so from from chargers and cpms if there's local power meters batteries pv panel installed then it will also provide or make data available from from these um local assets so basically we can mirror all the communication in the microservice stack here into our central broker and then we can make it available by an api or direct message bus access and that brings me to the point i mean we can we can have data flow in one direction so from the microservice stacks here so from each site um into our message broker but we can also have messages being gated into injected into the local message stream from a central connection point here so via an api you can basically inject the message into an ocpp message stream and for example restart a charger or change the configuration or have even a remote start transaction and so on and so forth so it's a directional message flow um we can we can integrate via an api quite easily so if you have your own sort of system operation system for example monitoring system or whatever you we can make the data available via an api again bi-directional but we can also do um and it's not a secret to say that this system here is an avs so amazon web services um backend system which um basically does the the the device management here for our microservices and also stores data so their databases and so on and so forth and if you're a partner or if you want to integrate a partner which is also using avs then we can even do an internal and we call it direct message bus integration and that is basically what we have done with the testing from other devices for like sero leaflet apps or here we have multiple inject OCPP messages through our central broker into specific OCPP streams here from chargers. So the integration is basically relatively straightforward. Again, we offer different ways of doing it via API or via direct message bus access. So it's typically done within a week or two to integrate the third party solution. And with that, I would basically like to give to Stefan to tell you a bit what available is, what they're doing and what they can provide for operational excellence. So I will change the presentation. Maybe just a side note before we dive into the presentation by Stefan, can you just give us a little bit of context on how the collaboration started, Robert and Stefan, maybe also both of you? Yeah, I can give a quick answer to that. I mean, obviously, we have the same customer segment. So we are both looking for, or we are both serving charge point operators, which have the challenge to, yeah, what I said, orchestrate systems. They are in need of special systems for, in our case, obviously for load energy management. And then in the wider perspective, or in the longer perspective to also create VPPs, provide web, not web services, provide grid services through charging infrastructures or with charging stations. And yeah, we have met, we have known each other for some time. Yeah, we got the first customer and then actually we made the integration. That's from my perspective. Yeah, it has been pretty on the same page. We both, I mean, you had the broker approach to enable your own solution. And we said, we, and then our conclusion was why you're not offering to the customers, to other CPOs, because you need data first to control your business and second to enable us other service like, like available. And that, that has been the starting point for, for, for the corporation. I think last year in ICNC, we, we detailed that approach and now, yeah, we offered to the market or you offering to, to, to every third party who was interested in that solution. I think that's a cool step. Yeah, it is. All right. So let's dive into available. Yeah. Let's dive into available. And yeah, I will introduce, introduce myself very shortly. When I started in 2007 and immobility, I mean, the landscape looks really quite different than today. I was involved in the first AC charger development and also in the ISO 15118 definition. And since the initial idea of available, I've been responsible for that. And yeah, today I have the pleasure to lead that topic as the managing director. So let's jump in. What is available? What are we doing? Can you share the presentation? Perfect. Where does available come from and why do we exist? And shortly in a nutshell, we are powerful. I tool for, for operation of charging infrastructure. And we predict events before they occur. We can help save operation costs and we improve the availability of the charge points. And the product was firstly developed within the own group. Because it was a combination of in -house expertise in AI and an immobility. What we really developed a unique product for charging infrastructure operation. And that value has been demonstrated with, with the charging infrastructure, the internal one. And it comes to the decision that we want to, to offer that to the external market. And therefore we become a known company. And what we are presenting today. Already in operations in eight regions at the moment. So next, next slide, please. Well, the last year it happens a lot. And with a lot, I mean, the sheer boost of a buildup of charging infrastructure. And that is pretty good or great achievement from, from all CPOs and hardware manufacturers who made that really possible. And it's also a huge step in the entire industry. And when we look at the figures at one hand, I think the more, the better in terms of infrastructure, but however, there are also side effects of large infrastructure. And we all know that with large networks comes great responsibility. And I asked also myself, what are really the downsides of, of a larger networks? And then we come to the next slide. What is really the challenge at the moment? And when we look at the figures at one hand, when we look at people who, who drive electric, we see the following picture that only 13% of the individuals have experienced no problems at all. And that's just 13%. I mean, imagine you're, you're, you're buying a cell phone and only 10% can do proper calls. I mean, that is something what, what should not be. And most common reasons are issues with starting a session, charging performance or unsuccessful sessions. And for me, that, that really looks quite alarming that so many EV drivers have such a bad experience. And at the same time, I'm asking myself, is it really the situation we find ourselves in the industry? And to answer that questions, let's have a short look at the uptime values of the, of the CPOs. And what about next? For the last couple of months, we analyzed some uptime data from different CPOs and we focused on DC because of the really high investments. It's really quite interesting to have a closer look on that. And it turns out that, that some CPOs are performing pretty well, according to the OCPI data and some other flagging behind, but overall, the picture doesn't look, look too bad. Is it? And, and that now it comes back to the EV drivers. What the hell are they talking about? Next slide. So who, who is right now? And who is right now? I mean, I got the impression that it's uptime values are not really the point. What, what really counts for the EV driver? And that also we see, I mean, you have, you have made some, some similar experience with that. And that is exactly the picture on the next slide that uptime does not equal reliability. And, and that is really the point. What really counts is successful charging sessions for the driver, not only for the driver, but also for the CPO. And at the one hand we see uptime, whereas, which are pretty high, but, but they are not really resulting in, in successful charging sessions at the end. And when we have a look on the picture and the data we see on the right hand side, that are two examples from on the left that are test drivers from P3 automotive across Europe. And they did tests with the 5,000 charging sessions and, and they found out that that is not working properly. And on the other side, you see data in our system when customers start using Souped. And if they want to be involved, that has already been 변 reward them. It's not going anywhere. give you the Ann座rthes, 其νэс aanгк е 시 вс bread Olive, which you already halt an AU289 digital gnome показ das been really quite clear when you look at the data. For example, an HPC charger is sending nearly 700 messages a day. And that means when you operate a charging infrastructure of, let's say, 500 stations, you have 10 million data sets a month. And now it's getting really tricky to solve that with Excel. And when we look at the data level of incidents, that's just more than a single error code. And it could have at least three dimensions. It's the sheer number of errors. The faults could be hardware specific, or it could be a combination with external factors. And then it becomes really complex. And to make it even worse, there are endless combinations possible. And that is where AI can really help. We, as a variable, our vision is that everyone deserves to have a successful charging session. Can you jump on the slides, please? Next. And our solution, next slide, please. Our solution does basically four jobs. And first, we identify and unify failure information, and that it makes easy for an operator to understand the condition. And it's visible in a graphic way without reading the code. And additionally, we combine that also with field experience. Second is that it's connecting the dots and the P's and give data meaning. And today, we know it's still code-driven, and therefore, a high effort to get the essence out of data. And for us, it's instead of leaving the operator alone, we use machine learning to get behind the meaning of the data. We also do, with the power of AI, we predict failures up to one week in advance. And that is a pretty nice thing, because it's pretty good when you detect things before the customer calls. And a currency is a really good level of 40 % to 50% two days in the head of an event. And we solve. That's the last thing, but not the least thing. We automate operations with our healing automation. And we do it really successfully 100K times per month. And that is simply carrying jobs you don't want to do manually anymore. Another thing is automation is also ticket creation. We know ticket creation is really a pain and not really a pleasure. And we also know that the quality of tickets are limited and often. And that's exactly why we automated event-based ticket generation. Everything at the end, less trouble and more revenue. On the next slide, you can see that we have four products in the line up. And each product is tailored to the needs of CPOs, hardware manufacturers, or CPMS. You will also ask, what is really the difference? On the next slide, please. You might ask, yeah, what kind of features are behind this product? I would not dive really deep into that today, but just mention a few topics. First, our dashboard has all relevant KPIs inside in real time. We use healing automation, what I mentioned before, where software learns when, for example, a restart should be triggered. And could solve a problem that we do automatically. We also have machine learning models which analyze failures and charging behavior that you see, I would say, invisible problems. That is like a pile of snow is blocked or vandalism and all that things. And last, not but not least, the AI protection for protection of failures one week in advance and so much more. Let's have a look at the available effect on the next slide. And that starting point is today's, the incident process. And it goes from the first call via call center to the CPO and so on and so forth. I mean, you get the point. That's a pretty long process. All in one up to nine working days. And using available, we saw that especially in the first part, it helps to see issues before the customer does. And solving these issues is also quicker, thanks to the clear indicators and evaluated data and recommendations. So at the end, it means the more issues you can solve remotely, not sending out the field service. And bottom line that that will save you, yeah, two days in the process, roughly, and makes your charge or earnings revenues really earlier. However, the ultimate figure for availability, as mentioned at the beginning, are successful charging sessions. And let's have a look on this on the next slide. We know that 20% of the charging session fails or more than 20%. And we are proud to see that our customers have really significant improvements and successful charging rates every month. And actually, every percentage of improvement is really valuable and available. And depending on your network size and utilization, it's something about 700k million euros cross marching every year, and for every percentage. And this is why we are so happy that that one of our customer recently reached 92% of successful charging sessions. And yes, it's not only onboarding of available, but the transparency and and available helps to improve the situation. So thank you for introducing available here as an add on solution. Yeah, to your program. Thank you, Stefan. You're welcome. So we have prepared a couple of questions for you, Stefan, Robert. You want to start with the first question? Yeah, I want to make fun some some comments first, because I myself also find the solution really, really exciting and so useful, because I know that operations is one of the key challenges for CPOs. And we can also see from our customers that where they really differentiate is in operations. Yeah, because there is so much as Stefan explained, there's so much you can you can lose when you're when you when you don't have your operations in under control. And ultimately, and this should be close to circle with respect to why a broker architecture could be an interesting alternative or is an interesting alternative to the standard architectures or the classic architecture. But I think the key question is, is that, of course, it's all about the data, right? So if you have the data, then you basically you can apply methods, you can apply AI or whatever. But at the beginning, the core, the basis of it that you have data transparency. And we hear that a lot from from CPOs that they say, but but we don't have the data or we can't get it or it's so difficult to get the data or we are not sure the data is really transparent. And I mean, in operations, it's, it's not always only the OCPP flow or locks or what you get by OCPP, but it's also about the the connectivity, right? So is the charger? So who did actually close the connection? Is it a charger? Is it a back end? And I think that information? Yeah, you can at least get when you put a system in between, which can can can can gather the data and aggregate that including connection data, or I think you also mentioned sensor integration through through a broker. That just as a comment. So yeah, I have some questions. The first question is, so from your standpoint, why, why do you think and I think I basically answered it already? Is data the only oxygen for for for charging business, but maybe you could give your perspective I gave mine and maybe you can give yours now. Yeah, well, I was about to do to add something. What what you said, and you're absolutely right. And we are data, we say data is the oxygen oxygen for the charging business. And the good thing is, and that is also mentioned, actually, we have more data available than we actively use as a as a CPU. And having said this, we see that the lack of data driven business models leads to an efficient business operation. And I mentioned the example before, imagine you, you are responsible 500 charge points, and you have 1010 million data sets every month. And that is just OCPP. And Robert also said, there's a lot more like sensor information and all that stuff. And that is quite valuable data. And in the most cases, we only use a fraction of this information, what is really hidden in that pile of data. And that means you really miss miss out quite a lot of insights. And at the end, I think you a lot of money will also use or lose down the road. So when when the customers approaching you, and they they realize that they, they realize the value of the data. So what what, from your perspective, is the biggest challenge they they experience these days? Yeah, that is also a good question. And they go through that moment a lot of times. But now, and there are some companies, they have the idea, wow, that's cool. Let's use our data and put it in a new service. And, and then what happens next? I mean, in many cases that they realize that a CPO, or CPO realize that this valuable data is somewhere in a closed environment. And as long your stick and the closed environment and with a given use cases, then it's fine, but there are no interfaces to extend your technical skillset. And let's, let's make an example. And imagine you have a robot, you have created some Word documents and stored in Microsoft Cloud. And in the past, you have been okay with that, because you're working with the documents in the Microsoft Cloud. But at one day, you decide, hey, I will, I would like to do something else with that, with your written content. And then what would you do if Microsoft would not allow to extract your data? And that is, that is not good, right? And I think that is exactly the point. To be honest, it's not all times that bad. We also have great examples where CPMs have a great API approach and where you can connect with third party services. But there are also examples, which come pretty close to my story that the data is locked. And that brings me to the basic challenge. If you have a business like a CPM, the CPM, the data, which your, your charge points or machines are created, needs first gathered somewhere in a neutral place. And, and I would say, make sure that the data is in the hub, enabled with, for example, Flexi charge, and then you can decide to what kind of service or which use cases you want to cover. And I think that is a really massive, disruptive topic. Yeah, yeah, absolutely. I mean, now, the question and then basically, now you we looked a bit into, okay, what is the challenge with with the with the CPMs systems, and I completely agree, it's not, it's not black and white, it's not all of them. And it also depends on, yeah, you have some legacy backends, and so on, and so forth. So it's, it's, it's complex. But what what we can see, and I would actually like to hear your opinion, because I know you have quite a long history in, in EV charging business. And what I can see is the big players, all the big players and the leading guys, CPOs, they use brokers. And there's a reason why they use brokers, either they use broker, or they've built their own CPMs. So yeah, maybe you can can give a bit on your perspective, how you see that, that that that architecture, I don't know if it's a shift, but at least what we can see is that more and more CPOs asking for broker solutions to basically build this independence. And so just wanted to hear your, your, your point of view on that. Yeah, you're, you're absolutely right. That is also my impression. The large of the CPS are the more they come and have their own data in their own systems. And that is for enabling use cases that but it's also to control your business. Sometimes they cannot ask Willie, they could not ask Willie the question, what is how the business is performing? Or what are every other questions and yours are not not only for available use, you need data also for to enable other use cases. And I think there's a broker helps really deliver a value for the CPO. And it's not for us a perfect match a broker, it's also I would say a perfect match for CPOs. And and that is really different now. And when you imagine, you have it in a central hub, and then you have the data at your fingertips, and then you can start immediately saving operational costs and, and for example, increase a successful start charging session that doesn't stop there. That is also what what you have landed, you can also do a smart restarts. And usually, it can take up a year until you get your data in our systems. But with your solution, I mean, that's something you mentioned, it's, it's a week. And I think I'm pretty convinced, as you also said, it's type of architecture will be in the future for every CPO. Yes. Yeah, I think we could be, I would really love to discuss a bit also the future vision of these architectures, what what role the CPMS will play in the future. But I guess maybe Chris, you can we can have that in the podcast. Yeah, or and you will will come back in the wrap up and talk about the fourth installment also of the masterclass series. There is one question from the audience. Alex Close is asking, is it possible to host the broker on premise? That's one for you, Robert. Yes, it is. That's what Alex short and sweet. That was the answer. If any question, it says the last few minutes to ask before we wrap up. So, so far, no questions. Again, you can always so you know, use thanks to our speakers for sharing valuable insights. It's been great to dive deep into how OCP brokers enable flexibility and innovation. If you'd like to receive the slides, please reach out. We're happy to share them. Or if you have more questions, of course, you know where to find us. Mark your calendars for March the 5th. That's when we'll be back with the fourth and final chapter of our masterclass series, where we'll take a closer look at the tight collaboration and interface between the broker and CPMS or CPMS is with a plural. So that's going to be an exciting session with our friends from Ivesto. Next longship. And if you're looking for more insights on the fascinating EV charging industry, don't forget to check out the ChargePoint podcast available on all podcast platforms. Thank you all for joining us today and see you next time. Bye bye. See you. Thanks, Stefan. Thank you. Thanks for the invitation.