 Big warm welcome, please, for Rose Schuller. Clicker is here, Rose. Wonderful introduction. OK, which way? This one? All right. Let's get the technology under control first. Good afternoon. This is more, oh, it's 11.30. That's afternoon in my time zones. Actually, it's about 2 o'clock in the morning where I'm from. But isn't this some beautiful facility? I'm not sure I've ever presented somewhere so lovely. As I was introduced, I am the Vice President of IOTG Global Sales for Intel. Now, my first warning is I'm not here to actually sell you anything. Look, I get a round of applause already for a simple statement. Why I'm here today is I've been at Intel for 28 years. And about 15 of those years were actually defining and running the P&L for networking with at Intel. Now, with that, we watched some market dynamics occur. We watched some change happening in the industry, sell some pretty interesting signals. And we adapted both our strategy and our technology to really drive a transformation in that market. When I came over to IOT, what I noticed in the industrial and energy segments in particular were some of those same market signals. So I want to spend about 20, 30 minutes talking about the learnings and networking, sharing how we think it implies to industrial and energy, and then just taking some questions from the audience. Sound OK? Let's get rolling. So I think this is data that I'm sure you have all seen at Nazium. 40% of the top 20 businesses in every industry will be disrupted in the next few years. 150 knowledge-based workers, 150 million knowledge-based workers today are going to be replaced by cognitive robots by 2025. 50% of businesses will rely on digitally enhanced products and services by 2020. The average age of a company historically is about 60 years. By 2020, they believe that that's going to be 12. So we're not going to see the pace of change like we've seen moving forward. Change is going to be part of our everyday lives and our jobs as leaders in our various industries is to how do you embrace it? How do you realign your business models? How do you realign your product strategy to take advantage of that change? So let's start by looking at the telephone. As I mentioned, I ran networking at Intel for many, many years. And I think this is a super interesting analogy. You look at the far left of this slide and it's Alexander Graham Bell's very first communication device, 1786, I think it was. And it was really not a commercial product, it was proof of the ability to communicate. And that evolved to what was called the Lucifone. You remember the curly cord with the dials on it, right? Does anybody know why it was called the Lucifone? Because it was present in a lot of I Love Lucy episodes. So a little kind of fun fact to put in your profile today. Then we saw the emergence of the Motorola Dynatech, first wireless phone. And then the BlackBerry came, where you actually had a full keyboard where you could text, right? And everybody was excited to be able to have a full keyboard and voice started to get encroached by data. And then finally was the thing that completely transformed many people's everybody's life. And that was the iPhone. And as you look at this transformation as consumers, we were very aware of the changing user experience that was occurring, the additional capabilities that were brought into our hands. But what was even more fascinating is not what was happening in our hands, but what was happening in the infrastructure behind this iPhone. As the operators looked at the surge in iPhone sales, there were a couple interesting trends and dynamics that were occurring. Number one, how many have iPhones? Okay, if you don't have an iPhone, do you have an Android device of some sort? If you have a BlackBerry, please let me know who you are afterwards. I'd like to grab it and put it in a museum somewhere. Everybody has these highly compute intensive devices in their hands today. How many of you see those little red notifications pop up on your apps? Get a lot of notifications. Each of those notifications, in essence, is the same function as a call set up and tear down on a traditional analog phone. So, do you ever do like a keynote and you sit back down and you pick up your phone and you have 50 texts that have come in and each of those creates a notification? Those notifications caused and created what was called signaling traffic. And signaling typically occurred on old five ESS switch equipment. But now, that was happening at a unprecedented pace with all of the application capability in the iPhone. Add to that what was traditionally just a voice data type was also additive with data text messaging as well as video. And it wasn't just video download anymore, it was video download and uplink. And the network wasn't architected for that type of traffic. More traffic, different data types, and uplink and downlink. So this chart has become very renowned in the telecommunications industry and that it was really their burning platform moment. Revenues and ARPUs, average revenue per user was flattening out and the traffic was going through the roof. It was an unsustainable business model. They needed a way to continue to make money, to save money on their infrastructure deployment, and not compromise performance and user experience. So what did they do? At a very interesting meeting that I attended with AT&T in about 2013 and 2014, their senior vice president looked at us and said, hey, can we use any kind of general purpose computing technology? Is maybe the $180,000 that we spend on a switch today, or $150,000 that we spend on a switch or a router? We can spend maybe $15,000 to $18,000. And if we can get that level of savings in our infrastructure deployment, we think we can weather this storm. We had a meeting with Huawei. Are you familiar with Huawei out of China? And then one of those meetings, Huawei said, Rose, we're currently using nine different architectures across our network. What challenges does that create? Number one, for each one of those architectures, I need a unique team with a unique skill set. I need a unique set of tools and capabilities for that architecture. I would really love to reduce that to two or three. What can you do? Well, at the time, our specific architecture wasn't terribly good for some of the workloads that were pervasive in networking, and specifically that's packet processing and signal processing. So we had to invest a tremendous amount in our technologies to support those various workloads, so x86 as an architecture could be used in the network. The other thing that we noticed was, hmm, if we moved in that direction, there's this other interesting trend in enterprise IT. The starting to gain momentum, it's called cloud. Can cloud architectures be used within these real-time workload-intensive infrastructure implementations within the carrier's networks? Again, investment was required to take something like virtualization, which wasn't real-time, but when you're an operator and you're gonna use virtualization in a radio network controller, which is hosting and supporting calls on a wireless network, your latency has to be real-time, 40 microseconds or less. So when you move a call for one VM to the other, you don't drop the call. We wanted to maintain user experience. And then finally, the third major transformation from a technology perspective was really getting scale in the management and provisioning of these networks. How do we get really efficient in managing the traffic? And by supporting technologies like software-defined networking, we were able to get that scale across provisioning management of both the network and the traffic on the network. Three big shifts. Workload consolidation, virtualization of applications, and software-defined provided the technology base that allowed us to support the needs of the infrastructure. So, what did we go from and to? We went from fixed boxes like firewalls, VPM, intrusion detection, routers, switchers, radio network controllers, base stations, you name it. And we moved it from a hardware implementation to that application running virtualized on a standard compute platform based on open standards and open architectures. And I think both of those phrases ring very loud and very true to this audience. The timeline, it's still in progress. It was a 10-year labor of love for me running that P&L. But as we look, what were the big things, the major initiatives? I had mentioned the architectural requirements. We had to learn how to move packets really efficiently on a standards-based architecture. And an innovation called Data Plane Development Kit allowed us to do that. And the evolution of Moore's law allows us to continue the performance per watt per dollar economies of scale that comes with using an open architecture. Open standards that were sometimes driven out of industry. We'll talk more about that. Sometimes driven out of academia. Involvement from the end user community. The likes of AT&T standing up and their chief operating officer saying, we are going to move to domain 2.0, really solidified the industry and pulled through the technologies. We saw some small startups, act as disruptors that were then acquired, like we had NCME at Cisco that was a spit out and then spun back in. We saw acquisitions at the likes of VMware because VMware had to augment their technology for applicability in this market segment. Technology, open standards, academia, industry alignment, new emerging players, which I'll talk about, all came together to really drive this transformation. Who were some of those critical players? We saw the likes of AT&T, John Donovan, making basically a gigantic statement of we will move to standards-based architectures and it will be called domain 2.0 and we will transform AT&T around this initiative. British Telecom, Verizon, China Mobile, driving research and product opportunity around virtualization of mobile networks. All allowed the supply chain to then follow suit so they could bid on the RFPs that were requiring this type of approach and technology. Suppliers that traditionally hadn't had a focus in networking, whether it be HP or Dell, because of the move to this standards architecture and leveraging information technology aligned with operational technology, the blend of IT and OT really created, I'd call some good tension in the supply chain. The likes of Cisco who had a traditional networking business as well as an emerging server business had a very interesting role in that they were able to kind of watch the gameplay out from both sides and align according to the needs of their end user community. We had mentioned VMware, Ericsson, Huawei's interest in aggregating their architectural approach and then the standards efforts around use cases out of Etsy and then an organization very similar to the one we're sitting at today, OPNFV, allowed for the open source capabilities to be productized and integrated and market and definition set forth. Standards, disruptors, new players and users with a passion for change all pulled through this type of technology. I was looking up, I did a little research this morning in preparation and I think it was SDN Central noted that now by 2020 they're expecting NFV, network function virtualization which is kind of the tagline of this whole transformation to be a $20 billion business, $15 billion business by 2020, excuse me, 15 billion by 2020. So about a decade of transformation leading to a $15 billion industry and growing. So let's look at what's happening and on the industrial side with the Industrial Revolution. You look back to the late 70s, you saw things like steam and machines and transportation really take hold in industries like textiles that drove that first phase of the Industrial Revolution that was followed by electricity, steel, assembly lines and auto manufacturing where you could increase productivity, improve yields, really get that economy as a scale which led to the 90s when, the late 90s when you saw computers and IT and automation move from things like what was analog type implementations to digital implementations and things like programmable logic controllers and such. Now we're facing what is the internet of things where machines will interact with machines, where intelligence will sit, whether it be in the robot or the edge on prem cloud or in a public cloud. You'll see machines interact with people, change their behaviors based on the implementation and the environment in which they reside. So how do our learnings from that networking experience really play themselves forward into this fourth Industrial Revolution? Well first, like we had in networking, we have to start with some type of business problem. People don't invest in change just to invest. There has to be almost a burning platform moment and that often centers around the need to lower CAPEX and OPEX. If you're in energy and in the oil industry, maybe it's the price of oil and the forecast of that price of oil and how do we reduce the infrastructure cost to maintain our margin position? In some cases it could be like faster time to revenue. If you're in pharmaceuticals, you can't be late to market with a new drug because it'll cost you billions of dollars in profit. On the customer driver side, people, if you look at some of the most pervasive use cases in IoT today, you often see predictive maintenance at the top of the list. Why is predictive maintenance so interesting? One, lower downtime, optimizing yields, being able to predict that downtime, get ahead, do predictive maintenance on the equipment. How do you get to a set of factory networks where you have the flexibility to move manufacturing seamlessly without being burned with a proprietary implementation that doesn't allow for that flexibility? And finally, we know within this fourth evolution of industrial networks that security, whether it be an industrial or transportation or pick another market, will be the foundation of the internet of things. So let's talk about how we believe this is gonna happen in industrial first. I think industrial is a fairly unique scenario and that there's a number of things like pumps and water pumps and air conditioners and you name it that aren't connected today. One set of data that we've seen is there's over a thousand manufacturing facilities, a thousand factories worldwide that have the ability to produce up to a petabyte of data a day. So how do we get that knowledge, that data off of that equipment so we can make more proactive business intelligent decisions? Then what we'll see is that these things start to get smart, whether it be the robot or the train, I think the most real time example that we see all the time in the news are cars, right? Not an industrial example, but we see the car getting more intelligent. When you connect it just like the phone, you start running applications on it, you secure it, you manage it, it continues to drive up performance and make that device, that edge of that network and the compute required to get the business intelligent out of it, smarter. One more click, there we go. And then finally the last phase, how does it become autonomous? How does it use the information from itself and the environment within it resides to make better business intelligence and decisions for itself? Connect it, it gets smarter, it will become autonomous. So what's that model look like? If we take what we saw in networking from these proprietary boxes and we play it forward to an industrial environment, some with similar characteristics, we see things that are very proprietary, specialized, monolithic, not being able to be integrated across multiple suppliers because of their proprietary nature. That could be a DCS, a programmable logic controller, human machine interface, machine controls. If you think about it, those hardware implementations are just like those network boxes. And what can we do as an industry to take those hardware functions and again make them a software implementation that can reside in a container on a standard compute platform? What are the requirements of that new platform for things like time synchronized networking, real time attributes, virtualization, security? How do we need to amend the technology to support the workloads that you see in the upper left-hand corner of this chart? That's what we as an industry need to define and work on together. Then you see an implementation where the software implementations run on a standard compute node. You have sensors and connectivity on the thing itself. You can run analytics on an on-prem private data center or if depending on the application burst it to a public cloud. Then you have the opportunity to achieve that flexibility, leverage the economy's scale from an IT environment while accounting for the needs of your OT world. So as we look across where these transformational opportunities are in industrial or energy, we can start in places like discrete manufacturing where efficiency and throughput are critical. If you talk to a car manufacturer, there are over 600 welds in a single automobile that are inspected after the car is manufactured. Is there a way by monitoring the equipment to predict where those bad welds may occur? As I mentioned, in something like continuous manufacturing, pharmaceuticals, if you're trying to get a manufacturing lineup quickly to get a time to market advantage, how can we leverage these standard architectures that provide that interoperability and flexibility to achieve those goals? In the domain of energy, the smart grid, predicting use cases, I'll create a quick analogy. When we were looking at leveraging this technology and wireless networks, we found that there was a bimodal distribution of traffic on a base station. I'm guessing you can figure out when that was. In the morning when people were driving to work and at night when they were driving home, the traffic on that base station would peak to max out utilization. And when you deployed a network, what did you have to do? You had to deploy based on the peak, not on the mean or the average, because you couldn't move the traffic on that base station or within compute platforms within that base station. So the ability to predict and proactively manage and distribute that traffic over an open platform and a standards-based architecture, Joe, I believe it was a 60% savings and CapEx and OpEx. Create that same analogy for a smart grid. If you could predict when your peak times were coming, distribute that energy appropriately, what type of infrastructure savings could you realize? And then for our last example, oil and gas, I would like to bring up Don Bartuzia, Don. Don is Chief Engineer of Process Control at ExxonMobil and Don's gonna take a couple of minutes to talk about the problem statement that they're looking at and how open architectures can help solve that. Thanks, Don, it's good to see you. Good to see you. All right, good morning, everybody. So in addition to Chief Engineer for ExxonMobil Research and Engineering, I'm also Co-Chair of the Open Process Automation Forum of the Open Group to put a plug-in for the home team here. I love it. So the business challenge that has brought me here today is ExxonMobil, we have to replace the majority of our process control systems that we use in our refinery and petrochemical plants over the next 15 years. And this represents a once-a-generation opportunity for us to achieve a step change improvement. Step change reduction in cost of these systems, step change reduction in the value generation that we're capable of getting from these computing systems. And in our search for a solution to that business problem, we looked around the world and looked at adjacent industries. And the experiences and transformations that we've seen in adjacent industries and I'll give you two specific examples in a second. We think are showing us the path to success to solve our business challenge. So the first adjacent industry that we saw that has technologies and business model transformation opportunities that we thought was a very useful analogy was avionics. And in fact, it was the future airborne capability environment face consortium of the Open Group and our connection to Lockheed Martin, which is really the reason we're here today. The second of the adjacent industry experiences that we saw is the innovations that have occurred in telecommunications, which Rose really elaborated on. But we first perceived the opportunity by seeing bits of technical solutions. And that opened the door. But as we got into understanding what happened in avionics and what happened in telecommunications, we realized it's more than the technology. It's a business transformation that has to occur. So our vision for the solution of our business challenge is to move from the proprietary closed systems that Rose described towards a standards-based, open, secure and interoperable process automation architecture. Against which instances can be built that are fit for purpose for the end users. And we, ExxonMobil, humbly acknowledge that we alone cannot affect this transformation. The essential thing that has to happen is other end users have to see this vision and join in this ask for change. And so that's really why we're here. We're well aligned. I mean, we're really excited to work, again, Intel was the company that first showed us the technology that got us excited about what Telecom had done. And now that we've understood network function virtualization and we've recognized the business model transformation that has occurred, we just want a piece of the action just to emulate it. And we think it's a fantastic example. Thank you, Rose, for giving me this opportunity. Thank you for your experiences, thank you. So one of my favorite sayings that one of my SVPs taught me a long time ago was, are you gonna be the disruptor or the disrupted, Rose? And it was a statement that stuck with me through my entire career. Because you can't sit back and watch it. You end up being the disrupted at that point. So as you sit in the audience, think about from your vantage point in the supply chain, are you a disruptor or are you going to be disrupted? This convergence of IT and OT is beginning to happen today. What role can you play? How can you connect the unconnected? How can you leverage technologies like security and virtualization to really take advantage of the economies of scale that comes in our cloud and our compute environments? And what are you doing to leverage the data that's created once you drive the connectivity to create business intelligence, leverage analytics, artificial intelligence to improve your current capabilities and drive your business models of the future? I love this quote by someone that I think is pretty darn brilliant. I think he most would agree with me that the ones who are crazy enough to think that they can change the world are the ones that actually do. What are you guys gonna do and Gal's gonna do in the next couple of years to change your individual worlds? Thank you. Now I think, Steve, you got a couple questions for maybe Don and I, yeah. Take the seat and Don, if you wanna join us as well. I'm gonna give you this because I'm gonna be playing with this clicker if I keep it in my hand right there. Join us too, Don. Thank you very much for that, Rose, and for the very articulate piece at the end, Don. So first question we've had from the audience. There's a lot of focus at the moment on security of IoT systems. Do you see this as a fundamental challenge or is it just part of the normal adoption cycle for this new digital environment? Well, certainly security is an absolute requirement, and we've seen really rapid innovations in security technologies that we think can be applied in process control, operations, technology, IoT services. So we're optimistic that this problem can be solved. Yeah, from our perspective as a technology company, we're looking at a few things. We believe that it's foundational. We won't be successful without a very capable and pervasive security implementation. We believe it needs to go from the thing to the network to the cloud, not only in the device, but in the transmission of data. And we also believe in our driving alignment around not only making sure that the hardware is secure, but also the software capability as well. And we just, as an example, used a very open approach. We licensed a technology recently called Epic, such that we can get a pervasive architecture in the thing for secure device onboarding at scale, which we think is a pretty significant problem in the industry that we were trying to help to solve. Okay, thank you. So you talked about the fact that end users making a stand or making demands, effectively end user requirements, pulled through effectively in open-grip language, pulled through the technological change in telecoms? Absolutely. And that's something we've seen as well ourselves. If the users are demanding it, then the vendors will see the value in producing the product. So do you see that being directly repeatable? Are there changes in the types of standards that are gonna be utilized or caused that pull through for the new age, either for you or in Don's area? Go first this time. Well, so I'm personally involved in standards development organizations as well. I mean, most specifically ISA. And one characteristic of the successful standards is when there's really the end users demanding it. When the standard activity gets to the point where it's just the suppliers debating on mine versus yours, it tends to be the characteristic of an unsuccessful standard. So Steve, your point is absolutely correct. I mean, the demand for the standards really have to come from the end users. And we like the open groups process. The concept of standard of standards resonates with us. I mean, we think in our world in open process automation, in our forum, we are gonna be doing a lot of selection of existing standards. There are some white space areas where we think we're gonna have to spin up a request to an SDO. But in general, standards are the way to win. Standard, successful standards come from the aligned voices of the end customers. And I'll just share my experience and I agree 100% with Don. We had, you saw the logos of a few carriers up there. The very first step that was taken was we embraced the needs and the use cases of those carriers. We had a very small group of folks that got together and defined what the top use cases were. There were nine. And then by the time we got to the standards bodies, like you saw Etsy mentioned, a lot of the groundwork and the alignment of the carrier requirements had been achieved. So then we were able to move very, very quickly in getting those use cases approved. So but that alignment and the needs of that end user community, I think was the tipping point moment for the adoption of the technology. And I know in your case, Don, we started with oil and gas with the open process automation very quickly identified as a similar interest level in many other industries too. Can you say a little more about that? Absolutely, so it's all the industry verticals that are using control systems. And there are differences in the use case requirements for discrete manufacturing versus continuous processes. You have to recognize that continuum. It's not a one size fits all situation but many industry verticals need this solution ranging from oil and gas, pharmaceuticals, electric power generation, minerals, mining, industrial gas as there's a large number of industry vertical sectors who need to align to call for the standards required to sustain what we're looking for. The other thing that I would add is we're a technology provider. We're not providing end solutions. We're not providing the software itself. So we very much rely on the end user requirements to drive technology modifications required to successfully support the use cases as defined by our end user community. So we really look to that dialogue with end users to say, hey, your virtualization technology as the example I gave doesn't provide enough real time support. You need some type of time synchronized networking capability on this platform. It doesn't come in a standard server configuration. How do you support that Intel? So in many cases, we will look to that community and then drive technology modifications back on our side to support those, the needs of these case. Last question. We have a lot of interest in enterprise architecture in our group. How to get involved in the business area on the need of the design for the new architecture? Which deliverable could show the value of an EA. So I guess the role, my interpretation of that is the role of enterprise architecture with the business in this new world. So in making the step change. Well, I'll answer that question very personally. In the open process automation forum, one of our top level working groups is enterprise architecture. And it's another fantastic story of cross fertilization. So I'm gonna give a shout out, Mark Bush from Shell, who's really coming from the IT side into this process control environment is bringing a lot of that skill and expertise in enterprise architecture. And we are learning rapidly as we go and it's proving to be tremendously valuable to us. And I just like thank my lucky stars that this happened to us. I'm a go EA. Good for you. All right. Good place to leave that. Rose, Don, thank you both very much. Thank you. Thank you very much. It was good to see you. Thank you.