 Live from Orlando, Florida, it's theCUBE. Covering Pentaho World 2017. Brought to you by Hitachi Ventura. Welcome back to theCUBE's live coverage of Pentaho World, brought to you by Hitachi Ventura. I'm your host, Rebecca Knight, along with my co-host, James Kobielus. We're joined by Dr. Ella Halal. She is the Director of Innovation at IMS. Thanks so much for coming on theCUBE, Ella. Thanks, I'm excited to be here. So you described your mission this morning as the mission to enable the connected car. Tell our viewers, what is the connected car? Oh, that's a very interesting question. So to us, to us at IMS, we define the connected vehicle in a little bit of a different way. So most people define it as being connected to the internet, but having it connected to the internet is not very useful to us drivers. But having it connected to you, the driver, is the key, is the essential point. And this is how we define the connected vehicle. So if by connecting it to you, we need to connect it to the internet, then that's a byproduct. But the key is giving you actionable insights as you're driving along, doing your daily commute. And as I mentioned this morning, you spend about 4.5 years of your life in a vehicle. That's a long time. That's a lot of time on your behalf. So if we are able to make this commute or your time in a vehicle more productive, then you get to enjoy this ride a little bit more. So augmenting the driver or passenger's experience with analytics, as opposed to what people usually think of which is self-driving autonomous vehicles. So it's one step of the way. You cannot have an autonomous vehicle without having connected vehicles. Because if you think about it, if you're having autonomous vehicles that has a horrible user experience, then what are you really doing, right? Nobody will want to ride it. So talk about what are some examples of these actionable insights that you could give someone as they're driving along. Okay, so imagine this. So if you're driving in the middle of highway and we know your destination in advance, but we know that there's no parking space and we can redirect you to another parking spot. That's an actionable insight that would be useful. If we know that you're driving and because of the way you're driving, your premiums will go up because you impose a little bit more high risk. We can give you coaching and feedback on how you can get to be a better driver and save some money. Think about it another way. You can be driving in a harsh braking, harsh exploration, imposing wear and tear on your tires that will cost you money because you would need to change them. If we give you this information early on, your incentive is to change your behavior a little bit to prolong the lifetime of your vehicle as well as save some gas. So IMS is a longtime IoT customer. Can you tell us how you've been able to stay relevant? Oh, that's a very interesting question. So definitely it's been an interesting, ever-changing market. So we focus on delivering a suite of services, not just one service with one provider. We actually provide a suite of services and we can enable different ones at different times. So we're not just a usage-based company, we're a connected car company. That means that we enable road user charging. So you know road user charging, right? So like multiple states in North America, as well as in Europe, different countries are focused now on having road charging. Instead of you paying the gas tax at the gas pump, every time you put gas in the car to offset the cost of the infrastructure, you pay the road user charge. At all. Well, similar to at all, but it's different because you're already paying it somehow. So at all is its choice. You need to take this road, you pay the toll for it. But for road user charging, it's trying to have a fair system to offset the cost of the infrastructure. The way it was done before using the gas tax, everybody had to use gas, everybody buys gas, and then they pay a little bit of money that goes to the infrastructure. Now you have hybrid vehicles, now you have fuel efficient vehicles, as well as you have electric vehicles that are imposing wear and tear on the roads, but there's not money coming to the government to help offset this cost. So they are trying to have a more fair system where we all contribute to the roads that we're driving in. So what's the metering infrastructure to enable road or usage-based road charges? Okay, so road user charging is actually quite interesting. So you think it has a lot of different additional overhead that you need, but actually it's not. You can, we as a company, enable road user charging through an OBD dongle that you add on your vehicle, and that's enough for us to get all the information needed, whether it's just mileage information without GPS. Again, it's a diagnostic port, yes, yes. So it has multiple ways, right? So you can enable it. Road user charging has multiple flavors of it. So one of them with GPS information, so we only charge you on public roads, not private roads. So if you're driving on a campus or a big campus at work, you're not charged for that. You only pay for public roads. If we don't have GPS, we do a mileage-based approach where we collect this data and we provide it to the government to charge you for it. And the nice thing about it, they actually do a gas rebate. So gas tax rebates. So you claim these mileage, you claim what you're paying for road use charging, and you rebate your gas tax. Another flavor of it would be based on an OBD2, sorry, other than the OBD2 is a mobile phone. So we can use a mobile phone to collect similar data and again, understand where you are and accordingly, send the information for the government to charge you as such. As it relates to the internet of things, then are those our approaches that would you regard? Those are both as IoT related approaches. Is there any other like metering technologies that you're exploring for gathering this data in a way that's more or less invisible? So I would definitely consider this as an IoT because again, the IoT is having the sensors embedded in multiple services. And so definitely to me, that's an IoT application. That being said, there are the existing tolling approaches which are like cameras and sensors at entry points and exit points. These are roadside infrastructures. You can also have like lane, high occupancy lanes where if you're in it, they can take a picture or sense how many people in the vehicle. So there are a lot of technology that enables road usage charging. That being said, I think using an OBD2 or a mobile phone is one of the most seamless things that you can do simply because you plug it in once and you don't have to interact with it. So how is Pentaho, how are you partnered with Pentaho to manage all this data to drive these programs? Actually, that's an interesting question. We're at Pentaho World, so let's hear it. Yeah, exactly, this is the right question to ask you. So Pentaho has helped us accelerate the ETL, the extract, transform and load process. Especially that we're collecting data from diverse sources, from heterogeneous platforms, whether it's from an OBD2, from a mobile phone or even from vehicles themselves. So collecting data from all of these different sources, Pentaho enabled us to ingest it fast, extract it, transform it and load it. It also helped with data integration. So the Pentaho data integration platform helped us work with multiple sources, get stuff fast, get it ready, and above all, it helps with the visualization because we work with different clients and each of them require a different report or view of the data in an aggregated way. So Pentaho definitely helped us accelerate and adapt fast to the requirement of our clients. Are the clients, are they fleet managers? Are the clients insurance companies? Give us a sense for the sort of dashboards you provide to them. And I'm using dashboards in a double entendre sense. To what extent can this technology be embedded in the dashboards of the future, connected cars to help drivers and passengers to modify their behavior while they're using the road system? So I will answer that onto two parts. So first, who are our clients? So we work with definitely insurance companies, some of the top on in the world, which would need data in a different form. We work with governments. We provide them for road user charging, for example, work with governments. So we provide them a different view of the data as the requirement. We work with fleet managers, a fleet insurance company, which is commercial lines. We also provide information to the end driver, to the end user, because how can you change, help them change their behavior? How can you give them actionable insights if you're not interacting with them? So all of these are different end points to our data and how we're exposing it. Regarding what can we show in the dashboard? If you think about it today on some sense, we're showing some information. We're showing actually a lot of information. So we have the mobile app that acts as an interface or a touch point between us and the end user. Because at the end of the day, the end user is the one who owns the data. It's not IMS, it's the end user who owns the data. And he's allowing us to use it to give him insights to get insurance discounts or know how much he's being charged for road user charging or enable services like roadside assistance and others. So the mobile app is our interaction point. And we have screens that show the logs of your trips and what good did you do, what bad did you do? We have analytics on this behavioral side. Where are you in terms of percentile on all different drivers? So that also gives you an encouragement and we always focus on positive feedback to help you enhance and change your driving to the better. What are you doing in terms of data masking and anonymization to protect the privacy of this data that's being processed through your applications? Definitely, I- Very privacy sensitive, obviously. No, yeah, and we are very, very aware of it. We're actually- And how are you using Pentaho in that regard? Okay, and we're very, very aware of it and we're very, very security conscious. So if you think about it, who are our clients? Our insurance company, who are our security focus and then governments are security focused. And so when you work with such big companies and big institutions that are very aware of security, you need also to step up and show that. And this is why we're ISO certified in many, many areas. So we're very aware of privacy. We never use any PII and our PII officer, we have a security officer that is very, very, very strict. Let me tell you that. And when we use data, we use it in an aggregated and anonymized format. So you cannot, and we use differential privacy on it, so you cannot identify one person added or removed out of it. So we use all of these different measures. And all the data that is being sent from the device is double encrypted on a VPN as well as sent in a binary format to our backend through a secure system. The devices are unhackable because they are designed such that they cannot receive input. They just made to send out input. So we work on privacy and security. We are actually privacy and security focused in Institute. And this is why we have been chosen by top tier insurers as well as governments to work with. So how far are we from fully autonomous vehicles? I mean, in your keynote, you talked about how actually people think we're further along in the journey than we actually are. But can you walk us through the next sort of next steps and then give us an estimate? Tell me when to ditch my car right now. Yeah, exactly. That's what I want to know. Okay, that's an interesting question. And I'm sure it's a very controversial one because everybody would have a different opinion. I know somebody in my team and if he's watching, he would say like in the next three years, I will have my next autonomous vehicle. And it all falls back to the definition of autonomy. As I mentioned this morning, there are five levels of autonomy. So level zero is having no autonomy whatsoever. So it's like your 1970 or 1960 car that you drive, you enjoy, but it does nothing except it enables you to drive. You have then your level one autonomy which will enable one feature only. So it's either cruise control, automatic braking, one thing to assist you. So it's one thing. And then you have level two, that enables two or more things in the same time, but you need to be fully alert and aware. Level three, well, it can drive a little bit autonomously, but you need to be alert, fully engaged and ready to engage at any time, ready to go at any time. Level four, it is autonomous under certain conditions. So for example, autonomous on highways or autonomous in specific cities, but not autonomous in others. Level five, it's autonomous everywhere all the time. This is what we're all waiting for, where we can sit. I'm on tenterhooks. Exactly, where you can sit. Yes, I want to sleep while I'm driving. I want to binge on Netflix or catch up on all the readings. Right, exactly. I got a lot of Game of Thrones on my face. Exactly, yes, exactly. So it depends on how you define autonomy and this is where it defines where we are on the progress. So if you look at Tesla and Google car, we're actually somewhere between level two and level three. Multiple systems are engaged, but you need to be fully alert and ready to intervene at any time. We're still not at the phase where you can lay back and relax and sleep. So what is your opinion finally? How many years are we looking? Okay, it depends on the level. So if I say level three, yeah, well, we have it. Ella! No! If we're talking about. You're hedging. Level four, I would expect, okay, so level four, level five has challenges. Level four, I would expect it to be between five to 10 years, somewhere in between, but level five is a little bit further and the reason is multiple things. I would say 15 to 20 and I'll tell you why. Number one, you would have multiple cars coexisting on the road and humans' decisions are subjective and are not always predictable. So you will always need to default to human intervention when needed. Road infrastructure takes a long time to be developed and for government investment. Third one, you need human acceptance and trust into these systems so that I can trust my six-year-old daughter to sit there and I will not be afraid for her life. So these things take time to develop and this is why I'm saying 15 to 20 years. Okay, you heard it here folks, all right? 15 to 20 years. Great. I'm all for it. Ella, thanks so much for coming on the keep. Yes, awesome. Great conversation. I really enjoyed it so much. Thanks for having me. I'm Rebecca Knight for James Kobielus. We will have more from the cube at Pentaho World in just a little bit.