 Apologize for the slight delay. Sean Nath from GE and I, I'm Sarah Cooper. I'm the Chief Operating Officer at M2MI, Machine to Machine Intelligence. We're gonna walk through how the combination of GE and M2MI and some other partners through an IIC test bed, Industrial Internet Consortium test bed, are using cloud foundry and actually a couple of different forms to build an industry wide platform for the aviation and airport industries. There are an incredible number of challenges, as we all know, in airlines and airports. Not the least of which is probably many of us are road warriors. I just know a number of you here who see it every conference. And things have actually keep getting worse. Despite the fact that we are progressing through time and civilization, it now takes longer to get from New York to Washington DC. It also takes about an hour longer to get from LA to Newark. And that's because of a number of things that are happening in airports today. There's certainly a lot of legacy systems. But frankly, there's a very difficult ecosystem in play. And this is true not just in aviation, it's also true in agribusiness. It's true in oil and gas. In the airport space and aviation, we have everybody from the airlines. They contract with baggage carriers and tarmac services and catering services. And there's the airport themselves, often a government agency. And then as you move from country to country, there are different vendors in each airport. All of these guys have to get together and be able to share information in order to make sure that your bag gets on the flight with you per often very strict regulations. So this is just sort of a look at where in the ecosystem there are interfaces. And there are a lot of them. There we go. Another interesting thing about the airport ecosystem, I'm kind of walking through some of the challenges. So you understand why we need cloud foundry and why we need to architect it the way that we have? I love the airport and airlines business model, mainly because they have continued to deliver a critical infrastructure service to all of us with some of the lowest margins I personally have ever seen. 4% net profit margin in 2015, that's globally across carriers. It was a bit higher in the US for North America, it was 7.6, which is a banner year that's actually above their debt load, which is about 6.8. And even 4% is a banner year. For approximately 60 years, the average was less than 1% profit. And the whole reason we go to an airport, obviously to fly. The guy selling you the $10 coffee, he's making margins. The airlines, not so much. So they should sell coffee. They do, they give you a cup of coffee. And I love this number, $8.27, that's what we are all worth as we sit in the seat. It's again, a little bit higher in North America last year. Fuel is a huge driver of margins in the aviation space. Whether you're spending time sitting on a tarmac, whether you turn on the air conditioning while your passengers are coming on, all of these eat into your fuel costs. And we all know that oil and gas have spent cheap recently. However, it still ends up being, even with the low prices, it still ends up being 28% of their operating costs. So when you look as an airline, you're looking at where do I generate new revenue? How do I affect my bottom line? Fuel has traditionally been one of the easiest places to go. The more savings folks like GE can bring into engine, the better the industry does. However, as we all know, we pay for bags now. There are more and more services that are coming online from Wi-Fi, to bag delivery services, to check your own systems. All of these come into the challenge of having to get transparency again across this ecosystem of loosely coupled partners. There's a lot of regulation that sits in the middle of being able to cooperate with your partners. There's also a lot of co-opetition, and that's really across the field. And that's been set up historically by dynamics between the landlords of the airport, the tarmac services, who's allowed to do what, when, and how. And the fact that it again changes at each airport you go to. So a system that is going to address this, that's going to allow information to be shared across this value chain in a meaningful way, has to be able to do that in a real time, and has to be able to take an advantage of all the policies to meet all the compliance metrics, but then also to pull that together in a way that can be seen and arbitrated in real time. So by that I mean, if SFO, which is fairly progressive airport, wants to be able to do just-in-time gate allocations, they've got to be able to figure out what are the resources available when a plane touches down, and who's on that flight. Who's on that flight is not something SFO knows. They want to be able to move a full flight coming in from a very long flight, very long flying time. They want to be able to move that right near all of the retail space, because those are the people who are getting off the plane and they're going to go buy something to eat, they're going to go kind of mosey their way down, and that increases the value of that real estate. Figuring out how to share information is certainly a challenge. Doing so in real time is even more so. IoT has been identified by a bunch of things, the internet of things, a number of players, airlines being chief among them, believe IoT is the answer. And not just because IoT is at the top of the Gartner hype curve, but actually because being able to measure physical entities in real time gives a single shared point of truth. So if I know where that bag is, because that bag has a GPS location, it's telling me where it is, I don't have to trust your 1970s enterprise software system that probably isn't up and running today. Or some guy who says, no, I totally saw it five minutes ago, I know where it is. And if we can agree on that point, then we can share resources and make sure that we're going to grab that bag and we're going to deliver it to the Uber that somebody paid 15 bucks to have their bag delivered to their Uber rather than having to run to the carousel. Frankly, I would rather pay to have my bag delivered to my car than pay the $25 to get it on the plane. There are lots of areas where IoT is already in the airport. Smart Baggage is something we're going to talk quite a bit about, so I won't go into depth there. There are beacons for a couple of different purposes. One is to help with the flow of traffic through the airport. Another is to manage the retail spaces and be able to justify the high rents that they are charged. Another is actually security. They want to be able to do behavioral analytics. Wouldn't it be great if there was no TSA? If you just walked through the first part of the airport and based on how your behavior and who they've identified you are, gives them a risk profile by which they're then watching every move you make. It's already happening. They're already watching. It would be easier if they could figure out your risk assessment profile. And then I don't have to stand in line in TSA for three hours. In-flight infotainment is another area where they're getting more and more information about us as consumers. Which again helps lead into the airport ecosystem itself. And equipment telematics, when you land and your flight is early, and you sit there for 30 minutes cuz the gate's not ready. It's rarely that there isn't an open gate. There's almost always an open gate. It's that they can't get, they don't know where the baggage carts are. They can't get the staff coordinated to be able to get the equipment they need, the gate handling, the ramps. They can't coordinate that in real time. So knowing where, what's available, what's scheduled, and being able to do that dynamically is very important. And in fact, it's the turn at the gate that determines whether the next flight is gonna be on time or not. So we believe that in order to keep all of these airport ecosystems, to be able to provide control to the different stakeholders who need control of their assets, but still be able to securely share information and insight level information amongst those ecosystem players. We really need a multi-cloud solution. And that again provides each airline has a good sense of their own IT. They've been doing it for a long time. These are very well developed systems. So we're not trying to replace that. We're trying to add a layer on top that provides a microservices architecture. And frankly, gives all of the information and insight and allows you to combine it in a meaningful way. So in the fifth element, when you needed to go anywhere, whether a cab or a flight, you had your multi-pass. So we think there needs to be a multi-pass, just spelled a little bit differently. And this is actually, we talked to players across the field. The M2MI came to this space because we had a lot of device vendors, the guys who sell the baggage guns. And things that you drop into your bag to know where it is and you can text something and find out your bag tells you it's been mishandled. We also have a bunch of folks that have GPS on the telematics on those, they're called LE3 carts. They've got special carts that will tell you if the bag was put on the right cart or not and some self-dropped baggage devices. Obviously GE comes in from initially the jet engine. And then of course, a lot of predictive maintenance and airport operations standpoint. So being able to combine that type of information, it's at the interfaces of those traditional silos that, frankly, there's new revenue and there's new efficiencies to be gained. So we got together and said, what can we do to help the industry? And we created this airport ecosystem testbed. We are initially, as I mentioned, focused on smart baggage. But these are our players. It's GE, M2MI, and Oracle. Oracle's got an aviation data model. They've been playing with IATA for a very long time on what is a standard way of presenting frequent flyer information. It's really impressive when you can take passenger information and combine it with IoT information. Things like, hey, Shyam's bag hasn't made the flight yet. How valuable a frequent flyer is he? Is he worth holding the plane by five minutes in order to get that bag on? Or are we just going to take the $250 charge and have it delivered to him at his final location? I'm sure he's worth the five minute wait. But it's Oracle's data model that helps them identify that. Again, the other piece of this, and although we will talk quite a bit about baggage, because there have been some recent regulations. And it's an excellent use case where you've got all of us as consumers adding devices into the ecosystem, which are untrusted. Plus you've got a bunch of things sitting on the tarmac which are controlled by the airport and or by the airlines and their operators. So those are generally a trusted environment. And so that's a very interesting ecosystem, if you can address a mixed trust environment in a meaningful way. So, taking a quick look at self-dropped bags. And I promise we're getting on to the Cloud Foundry bits and why Cloud Foundry. But again, just to sort of set this up, you'll see there's a lot of messaging that goes on between, you've got new bags which have built-in IoT components. Samsung's done one, BlueSmart, Regan, they're frankly all over the place at the moment. We also have tags that you can drop into a traditional bag. It gives essentially the same information, very different format. Many of these guys have been building these for years. They wrote their own TCPIP stacks. Thank you very much. This is another industrial component where everybody wrote components. They wrote their own stacks, they wrote their own protocols. Nothing is standardized and it tends to be very limited resource devices. They're also a bunch of self-drop components. So one of the key pieces of IoT and any ecosystem, but particularly in airports, is that real-time availability of information. So although we talk a lot in cloud systems about fresh data and stale data and historical data, this turns very stale very fast. You need to be able to share information and you need to be able to do it consistently and quickly with the players who can consume it while it's fresh. Doesn't matter where your bag was, if you're already home. So again, these are just sort of who the players are and what I wanted to point out here, we're going to talk a little bit about GE's Predix Cloud, which is obviously a Cloud Foundry Environment. M2MI runs on Predix. Oracle is at this point in Oracle's Cloud, but we'll also be running on Predix as a microservices architecture. The M2MI components are composable IoT elements, so they do a bunch of functionality, again, it's services. But the real interesting thing here is all of the different components. So in this one data flow, we've got Bluetooth, we've got MQTT. And these are from the device vendors. This isn't me just putting up the IoT primordial soup. We've got the proprietary TCP IP I mentioned. And the radar guns can actually be switched between HTTPS and MQTT. But they give you different outputs depending on the protocol you select. So that's always fun. So why Cloud Foundry? Especially given the complexity and the heterogeneity of IoT, talking about all of those transport protocols, all of the different data structures. That's a lot of processing to be able to put elements into Cloud Foundry. And certainly, Diego helps a lot with some of the routing capabilities around TCP, can't wait for some UDP stuff too cuz everybody has built something. But essentially what we've done is we have moved some of those components which don't currently conform to Cloud Foundry into a distributed node. M2MI's components are distributed. So we can move some of the data handling and the device management, which is over the air configurations. We can move that out to a distributed element usually sitting at the airport, although it can also sit at M2MI's environment. It's really just essentially outside of Cloud Foundry at this moment. We pass information securely onto the M2MI central platform, which is sitting on GE Predix. We are leveraging a number of the services from Predix around their MQTT support as well as some of their data services. The data comes in and other things people always talk about in IoT is if you've got a line of, let's say, a conveyor belt. And you wanna be able to figure out where the bag was at a specific time. You'll get a notification from the bag, and you'll get a notification from the conveyor belt itself. They are time stamped. They don't happen magically simultaneously. So if you're building a record for something like the Oracle Data Model or other analytics packages, you've gotta have a consistent record of the information of that system. It's sort of a snapshot, but all you're really given is pixels. So a big piece of the data components here are actually putting together those pixels, those bits of information from all the different devices and building out that snapshot that then the analytics packages can go and do. That's a challenge in and of itself, but it is the sort of first pass of how you get to the in-stream analytics, which are the good parts here. The whole point of this is the services layer. So the sort of castle top there is us serving information and accepting commands down from higher level applications. So this environment allows airlines, allows device vendors, allows service operators to be able to get the coherent information they need and have access to. The privacy module is contextually based. So that means that it has a model of the state of the system and it provides guidance on whether you get access to the data you want. For instance, if a bag has been identified as a security risk, you may not get access to that information because it's been removed from essentially your whitelist of shared information. Unless you're TSA, in which case you may be the only one who gets access to it. That ability, again, to do real time responsive security around information is absolutely critical to not just the airport, but industrial ecosystems at large. And it's, again, it's not something that you want to build policies around. You want something that can keep track of the system itself because you will never have enough policies to cover enough scenarios to get the system locked down. This also provides a complete audit trail for everything that happens in the platform as well as in the airport. We are providing application development, again, on Cloud Foundry, which gives the airlines as well as others who are writing applications to this, the flexibility of writing on Predix and then moving to whichever cloud they want. Again, this is not about driving down some kind of vendor lock-in. This is about opening up the data in a controllable way that compliance can measure. And that's it for me. Shyam's going to talk to us a little bit about particularly how Predix. Thank you, Sarah. So GE has a long history in aviation. We make about two-thirds of the world's commercial jet engines. And historically, we have learned how engines operate. So we know how to make them more efficient from fuel perspective or reduce unplanned downtime using industrial internet. The number one passenger complaint will always be that flight is on time or at least it takes them to the destination and everything else follows. So for a long time, we have embedded sensors in jet engines, used that data through Predix platform, built industrial analytics to analyze that information, help in flight efficiency for our operators, which are typically the airlines. And the domain knowledge we have acquired around aircrafts, airports, we are now bringing into the airport domain and focusing on the inefficiencies that are in and around the airport. An airport is exactly where the passengers meet the airlines, and airlines meet the airports, which are essentially extensions of cities. So one way to look at airport is like a mini city. It has all those problems like traffic congestion, movement, security, and so on. So in order to power this application for aviation, we are leveraging Predix, some of the capabilities of Predix are being able to map the, acquire the information from assets, build the model of assets. So that you know what does the sensor data coming from a different sensor mean. And how do you add value or metadata to the sensor information, so that it can be analyzed in context of the enterprise data. And once you have figured that out, you can create industrial data analytics to help optimize. So for instance, collecting information about passenger flows to the airport or baggage flows, we can learn over time what is the best way to optimize the operations. For instance, one of the airports looks at the number of people or bags transferring between the two flights that are landing close to each other in time, and they can dynamically put them in gates next to each other. So that passengers don't have to walk across terminals and the bags can easily move from one aircraft to the other. And typically happens in airports which are typically used for flying through. So the Middle Eastern airports are typically where people fly through them or London, UK, even though it's a small country, has second highest number of air travel after US because a lot of people fly through those airports. And optimizing the operations based on information, whether it's passenger, the bags, or other things, or availability of ground equipment, as Sarah mentioned, can help to add efficiency at the airport. So next we look at some of the capabilities of critics platform in details, it's obviously built on Cloud Foundry, provides data infrastructure that can be used for persisting the data. And our goal is to insulate the developer from the actual technologies used for Datastore, rather expose them as APIs, be it relational, object store, time series, key value pair, and so on. And then you can use asset services, analytics, and other kinds of services that we make available to quickly help to build business applications. So as we start our journey of building applications for aviation industry, we have a sort of roadmap where our first application is around baggage handling efficiency. And there are business reasons for doing that. IATA, which is International Air Travel Association, has come up with a regulation, it's called Resolution 753, which will require the airlines to have the end to end custody information of bags by June 2018. Now, it might sound funny that we give our bags to airline, we expect them to have our bag information with them. But the reality is, once the bag leaves the airline chaos, it is in conveyor belts or other equipments which are operated typically by third parties, airports, baggage handling companies. And then if there is a connection, again, the airline may have no control in that airport where the bag transfers. And then at the destination airport, again, the ground services which may or may not be airlines own have the bag. So today, different parties use different information, some newspaper manifest, handguns and so on. It's actually very difficult for airline to have the baggage information end to end. And as a result, 7 out of 1000 bags are mishandled, which includes delay, damage, and loss. And when that happens, airlines incur on an average $100 to repatriate a delayed bag. And if a bag is lost in US, the liability is as much as $3,300. So, and we saw how thin the operating margins for airlines are. So every bag they can prevent mishandling adds to the bottom line. And adding, sensing, so today, typically bags have a printed two dimensional barcode, which can be only scanned by line of sight. So the scanning hits are very poor. If we make it three dimensional, it can be an active sensor or 3D printing barcodes as possible now. So that can improve the efficiency of the baggage handling system. Once we are done with the baggage handling, we have a roadmap of taking this to other areas such as connecting passengers which on permission. So your smartphone can be used for tracking your location that can help the airlines tell where the passenger is, how long should flight be held, and so on. Crew, often airlines have no clue where the crew is. They only know where they landed at the last gate, but don't know how long it will take them to get to the aircraft. Likewise, for ground handling resources, vendors and retail. Retail is a big source of non-arrodotical revenue for the airports and airlines, so there is a lot of efficiency we can drive by connecting the passengers and the retail related services. And finally, the number one, in a recent survey, the number one desire of airline passengers has been guided navigation through the airport. There was recently a New York Times article which said, airports are designed for everyone except the passenger. Airlines have been focused on guiding the aircrafts from gate to gate very efficiently. But it seems nobody has really paid attention on the passengers. So guided navigation at the airport, especially for first, if you're first time in an airport, you're not a frequent traveler, can be of much value. So our goal is to partner because we can innovate faster, we can solve real world problems faster. It's too big of a space for any one company or airline or airport to try to handle it. And that's why we leaned on the Industrial Internet Consortium, which has over 250 companies. All these companies like GE, Oracle, M2MI are all members of that, and we decided to collaborate to innovate. And we are always open to have new partners that could be sensor providers, telecom providers, or different services that can be used to solve this problem. So that brings an end to this. I don't know if we have time for questions, but we will be around and we'll be happy to take your questions. There's also a GE booth. So feel free to connect with us. And if you have any questions, we can help you. Thank you so much.