September 1, 2016 | Written by: Karen Lewis
Share this post:
We sat down to chat with Andy Stanford-Clark, IBM Distinguished Engineer for the Internet of Things. Andy is also one of IBM’s Master Inventors – which means, among other things, he invents lots of cool things and gets lots of patents. Andy’s current role is Distinguished Engineer for IBM Watson Internet of Things Platform, which means he works with clients to help them use IoT, and understand where IoT comes into their businesses.
Question: Andy, what does it mean to be a Master Inventor at IBM?
Andy Stanford-Clark: A Master Inventor at IBM is somebody who first of all has above a certain number of patents. You have to have at least twelve patents in order to be considered for Master Inventor, but then it is also a peer election, so you have to be well-known to mentor other people, coach them in their innovation, and help them find the innovation in their work. It’s people who have a track record of doing that. My door is always open; my white board is always ready for half-baked ideas to be made into ideas that could be turned into patents.
Question: How many patents do you have?
Andy Stanford-Clark: I currently have forty-two patents. I usually just say more than forty but the exact number is forty-two.
Question: Do you have a favorite patent?
Andy Stanford-Clark: Yeah, my favorite patent is actually the temperature sensitive barcode. This is where one line, one black line of a barcode changes from black to transparent when scanned with a laser to give you two different codes based on whether the barcode is a certain temperature or not. This can be used to detect whether something’s frozen or not frozen. The original idea came from seeing a microwave that just had a barcode scanner and a Go button, so when you scanned the thing you were going to eat, it would go to the internet, find the cooking instructions, program the microwave for the right cooking time, and then you just hit the Go button and it would do it. Obviously, there’s the question of whether your pizza is frozen or not, and so the invention came from how the microwave would know to cook it for shorter or longer according to whether the pizza was frozen or not.
Question: Do you have a favorite patent that is in the realm of IoT?
Andy Stanford-Clark: A lot of my patents are to do with the way small devices send messages to other devices. In particular, a lot of things to do with what we call transaction processing where you’ve got to send the message reliably from one place to another, and make absolutely sure it gets there. That’s a lot more difficult if you’ve got a small device the size of your thumbnail that’s got no permanent storage or no disc drive. I solved a lot of the problems that allow us to transfer data reliably from one place to another for the Internet of Things. There’s a number of patents in that area.
Question: You talked about the messaging service that you help create. What’s it called and how did it come about?
Andy Stanford-Clark: Back in 1998, I was working with some people on oil and gas pipelines called SCADA systems, and there were lots of proprietary protocols in the world in those days, that were closely held by their inventors and by the companies that made them. We had this idea it’d be really good to have one ring to bind them all; just one more protocol that would be open source and free, and everyone could use it without a license. We invented what is now called MQTT, and seventeen years ago when this happened, I really only foresaw that it would be useful in the SCADA world, but then we saw uses for it in other industries like vehicle telematics, and agriculture, and healthcare. A whole lot of different industries. Gradually as the internet of things become more and more popular, MQTT became more and more popular, and now it’s an ISO standard and pretty much everyone’s using it. All the big manufacturers, all the big vendors that have IoT platforms, like IBM, Microsoft, Amazon, all host MQTT connections as their default option for connecting things to the Internet of Things.
Question: How did you get your start with IBM?
Andy Stanford-Clark: When I was doing my PhD, after my degree, I wrote to all the big computer companies and said “I’m doing this PhD in parallel computing, are you doing anything in parallel computing research in the UK?” People like Cray and DEC. They wrote back to me and said, nothing in the UK, but maybe there might be something in the US. IBM wrote back and said nothing in the UK, but come and talk to us anyway. I came to a graduate recruitment day, and at the end of that day, in fact it was here in Hursley, they said, “We know we want you, but we’re not quite sure what for yet, so will you accept this job offer?” “Yeah, sure.” I walked around with a smile on my face for the next week. In fact, I ended up writing a UNIX device drive as my first job because I’d done some UNIX at university, so I was doing that.
Question: You were happy?
Andy Stanford-Clark: I was very happy about that. I had actually dreamed that one day I might work for IBM. I figured I might have to work for a number of other companies first and finally get to work for IBM, but to get IBM as my first job out of university was quite a coup.
Question: How many years has it been since you were hired?
Andy Stanford-Clark: This is my twenty-fifth year at IBM. I got my pen and a certificate to say thank you for working for IBM for twenty-five years. It’s gone really quickly.
Question: Can you describe your typical work day here? As a Master Inventor, you come in every day and do what, exactly?
Andy Stanford-Clark: Being a Master Inventor is kind of my part-time job. My full-time job at IBM is working on the Internet of Things with our clients. Typically, I’ll be working on a number of different projects at the same time, working with big customers, large corporations, who see the need for IoT, but they haven’t yet worked out exactly what it means for their industry and for their company in particular. Everyone worries that IoT will be a really disruptive technology – something where their competitor is going to work it out first and take out a whole market sector for them. People want to get ahead of the curve, they want to try out small prototypes, they want to do experimental projects with Raspberry Pi and Arduino, and get the Node-RED and Bluemix – to try things out so they understand the art of the possible, so that they can then look for potential projects within their organization. For example, in this environment, we might discover that a really good project to start off with IoT is one where we can augment an existing business process with IoT. When we’ve completed it, we can do a side-by-side measurement to say this has improved by xx amount because of the IoT technology. As a result of choosing something where a comparison is immediately evident, we can then build a business case or justification to help things move forward.
Question: Are you kind of leading the horse to water a little bit?
Andy Stanford-Clark: Yeah, sort of with evangelism, and I do a lot of customer briefings. I do a lot of public speaking because loads of people have heard of the Internet of Things – even random members of the public have heard of Internet of Things. Some people see it as toys that have an interactive conversation with their children, or maybe a toothbrush that measures how well you’re cleaning your teeth and gives you feedback, or it might be something like your car which is telling you how the car is performing and then sending data back to the manufacturer so that they can tell you when it needs to be brought in for maintenance. We do what’s called predictive maintenance, predicting before it happens that something’s going to break down so that we can keep things actually working.
Question: Can you talk about Hursley, the building that we’re in, your history here? What is the work that you and the rest of the IoT team that are doing? What does it mean to you to be in this building?
Andy Stanford-Clark: At IBM Hursley, in our development laboratories, we have two sets of people. One is developers who are developing, amongst other things, the core technology for IBM’s Internet of Things platform, the Watson IoT Platform. It’s a huge cloud-hosted infrastructure, which allows hundreds of millions of devices to simultaneously connect in to our cloud servers. Because of our history in messaging and transaction processing here in Hursley, which goes back fifty, sixty years, we were ideally placed to do that work when the opportunity came up.
The other group of people here at Hursley are our services people, who are working directly with the clients on projects. We’ll go to a client and they’ll say we want to do this project, we want to solve this problem, how can you help us, and people like me will draw an architecture on the white board and they go yeah, that’s really cool, let’s do that. Then we’ll engage a set of people to help them actually build it if they don’t have the skills themselves. There are people here working at the coal face directly with our clients. That’s really good working side-by-side with the developers because the problems that our services people see in real life, solving real customer problems, they can feedback and say it would be really handy if the platform also did this. Most of our customer requirements come from those interactions with our clients.
Question: Can you give us your view of how special Hursley House is?
Andy Stanford-Clark: IBM’s Hursley laboratory is a fantastic place to work. It’s out in rural Hampshire in the south of England. It’s a lovely old manor house, dating way back to 16th century. It used to be a hunting lodge for people hunting for deer, and then became a very grand, stately home for a big family. When IBM took it over after the Second World War, we set up our laboratory here, and we’ve been innovating and inventing ever since, initially in hardware, making disc drives and video monitors. Then in the mid-90s we moved more to a software portfolio, and it’s pretty much a software-only lab now. The grounds out here are wonderful to walk round for a lunch-time stroll, which might seem decadent and wasting time, but some of our best ideas happen when you’re in a very liberating environment and just chatting with people, particularly people from other departments who might be working on completely different technology. You get that cross-fertilization between ideas from one area being applied to another area. Quite a lot of innovation happens out there on the lawn.
Question: How does it feel coming into this place every morning?
Andy Stanford-Clark: IBM Hursley’s a really special place to work. There’s a real buzz around the place. You walk past people and you see pairs of people chatting animatedly over a screen, or maybe some IoT folks are looking over some little circuit or something, and there’s definitely something which keeps people coming back here. I’ve been working here twenty-five years, apart from three years working in America, and compared to the other locations, there’s a real atmosphere here. When people ask, “Why do you work for IBM?”, most people say, “Because of the people who work at Hursley; it’s such a cool place.”
Question: Can you tell me a little bit more about the Innovation Centre here at IBM Hursley?
Andy Stanford-Clark: We’re in our IoT laboratory, which is part of the Innovation Center at Hursley. We set this up to really show the spectrum of technologies that are involved in IoT. So everything from industrial automation, through to home automation, vehicle IoT, healthcare, assisted living – we’ve got examples of all of those here. We’ve got this really cool demo system, which is driven off a touchscreen, where you can select the particular demo you want and the whole room changes. The lighting theme changes and then all the screens fill up with different pictures to demonstrate that particular story. You can change the whole mood of the room just to give a particular anecdote. It’s essentially a performance when a customer comes in here, and you can talk to them about the different technologies. People really love it because in many ways it brings IoT to life. It just shows how easy it is to get modern generation devices connected to an IoT service (like the Watson IoT Platform) within seconds and you can see data coming from a temperature sensor or a motion sensor appearing on a screen behind you that’s showing data in the cloud.
Question: There are a lot of examples of IoT devices behind us and on the shelves. Do you have a favorite piece of technology or IoT technology?
Andy Stanford-Clark: I’m particularly excited at the moment by robotics and its use in industrial automation for cognitive manufacturing. Playing with little robot arms – actually toy robot arms you get from an electronic store – but those are really exciting to play with. The principles that we use when we animate a small toy robot actually apply to big industrial robots as well.
Question: What’s that process like? How do you really sell them on the fact that this technology can very easily scale to a big, industrial robotic system?
Andy Stanford-Clark: It’s quite easy to show clients the movement from a single prototype through to a big operating-at-scale thing. That’s because everything in this room is all driven by MQTT, and we can actually put up the Node-RED flows on the screen behind people and say, actually, that’s linked to that.
For example, when the weather changes, the weather forecast comes in, it turns the room lights from green to magenta to red to show it’s going to rain in a few minutes’ time. Or we can say we’ll re-route that message so that it turns the room lights a particular color, or we can say it sends me a text when it’s about to rain. You can link input sources to output sources very easily and very flexibly, and people suddenly see that, although it’s only a few devices in this room, nonetheless it’s trivial to link them all to each other. They can just see how, when that then operates at scale, when you’ve got ten thousand or a million vehicles out there, or a whole load of washing machines sending back data about how they’re being used and how they’re performing, they can make the mental leap because everybody understands that cloud-server technology has this elastic scaling. It elastically expands and contracts to meet that load. That’s really the promise of cloud-hosting, and that’s what we deliver.
Question: Do you have a favorite example of when that scale happens in the real world? For example, MQTT. You helped develop that language, and now it’s being used everywhere. Can you talk about how that felt?
Andy Stanford-Clark: When I developed MQTT, with my collaborator Arlen Nipper, seventeen years ago, I had what I used to call my ‘modest plan for world domination’, which was that one day all devices would connect to TCP/IP, and all of them would talk MQTT. At the time, it was a whimsical thing to say because I imagined it would only be the oil and gas industry. As things started to grow, I became a one-man-band teaching both IBM internally and customers externally what this whole thing was all about. Then about six or seven years ago, we crossed a tipping point when the open source MQTT broker, called Mosquitto, was written and got out there, people realized this wasn’t a little thing that IBM had dreamed up. It was actually a lot more useful than that, and more broadly applicable. That opened the floodgates.
When Internet of Things began its exponential growth, and IBM, Microsoft, and Amazon adopted MQTT as their core protocol and it became in international standard, suddenly I realized that that modest dream for world domination maybe wasn’t too unlikely after all. There are now, I can safely say, hundreds of millions of devices connecting to the internet through MQTT. I feel so proud about that.
Question: When did you realize that this was going to be big?
Andy Stanford-Clark: Actually, the tipping point for me was when one of the big social media sites adopted MQTT for their smartphone app. All the people who were doing messaging on that application were using MQTT under the covers. Right there, that was 750 million clients overnight. Wow, that’s just amazing. In fact, that was the night before one of the big conferences where we were going to be talking about MQTT, so I was able to stand up and say … People say, “How many people are using MQTT then?” At least seven hundred and fifty million because they just turned this app on last night.
Question: When you heard MQTT was ISO certified, how did you receive that news?
Andy Stanford-Clark: We were just high-fiving, whooping for joy. It was just amazing. It’s funny because it had been a 15-year journey for me to get from the point where we’d designed this protocol thinking that one day it might be useful to quite a lot of people, to seeing this was something that was actually going to become known worldwide. I know various other inventors in IBM who’ve done amazing things, like Mike Cowlishaw with the Rexx programming language, which became a standard. People write books about it and have conferences about it. When I first met him, I thought I just couldn’t imagine that someone could invent something that would be used worldwide. Actually people are having conferences and writing books about MQTT now, so that’s pretty cool.
Question: That is pretty cool. Where were you when you got the news?
Andy Stanford-Clark: We were here in Hursley in the Briefing Center, and somebody came in pretty excited waving their phone around, going “look what’s just been announced.” Wow, so cool. Quite often on Twitter you see these things overnight, so it was already on Twitter, but I’d missed it. People were actually talking to me on Twitter saying it’d be so cool if MQTT became an ISO standard and stuff like that. People saying 20922. I was replying, “What? 20922? I have no idea what you’re talking about.” Okay, I finally Googled 20922 and MQTT and realized that was the ISO number that had been assigned to it.
Question: What does it mean to be ISO certified?
Andy Stanford-Clark: ISO standard is just something that everyone’s heard of, MQTT was already an OASIS standard, but to also be an ISO standard as well, means you’ve really made it. When you tell people it’s an ISO standard, people know that’s something important.
Question: They deemed it fit for widespread use because of how stable it is, is that right?
Andy Stanford-Clark: It’s partly the stability. It’s also the wide adoption of it. What tends to happen with standards bodies is that if just one company comes along and says: “Can I make this a standard, please,” they want to make sure this isn’t just your thing you’re trying to foist on the world. If half a dozen companies come along and say we’ve got this protocol we like, and by the way, we’ve all implemented it and it’s been around for ten years and loads of people are using it, that’s pretty much what gets you to the level of entry for a normal standard, like an OASIS standard.
Then it’s the worldwide adoption; it’s no longer controlled by IBM. I don’t control MQTT. Although I’m on the technical committee that shapes MQTT, I personally don’t have the controlling vote or anything like that. Although I may be the father or the grandfather of MQTT, it’s in other people’s hands now, which is great. MQTT was always designed to be very simple, very short specification, very easy to implement on small devices. I feel I have to keep the team honest and say, (if they say it’d be really cool if we had this or that feature in), yes, it’d be really cool, but it’s going to move it towards one of the existing enterprise messaging standards, and then it’ll be a 300 page spec, and really complicated to implement on a tiny device. I keep people honest and say we’ve got to go back to our founding principles and keep it easy to implement, small, compact, and that keeps people focused.
Question: In a way, are you still carrying the torch?
Andy Stanford-Clark: Yes, obviously I’m still carrying the torch, and at the technical council meetings, I still say a lot because I have a lot of anecdotes about uses of MQTT in the past, which have led to certain design features being made. People say there was this feature in MQTT we never quite understood, what were the designers thinking when they put this in. I can tell them, and that really helps people understand why we made certain design decisions, and why certain things were left out, or put in. In fact, after we launched the MQTT protocol back in 1998, we didn’t make any changes at all to it for ten years, so it stood the test of time for that long before we made any changes at all to it. Now we’re going through some revisions that make it more cloud ready and more ready for the next twenty years, but it really has stood the test of time very well. It’s always that simplicity. Everyone can understand it really easily, what it does, how you implement it and get going.
Question: What is the next standard? What is something that you’re working on now that you know will reach that same level of success?
Andy Stanford-Clark: I think my role has changed since the MQTT days. I was quite focused on one particular technology then, which is an underpinning technology for the IoT, but now the IoT is here and we’ve got all the building blocks. My interest is more in helping customers create solutions, and they vary a lot. There isn’t really one technology which will take over the world. We’ve got a lot of IoT technologies. It’s really how do you apply those technologies, how to use those building blocks to create solutions to solve things as widely varied as vehicles, or healthcare, or agriculture, or helping people go into space, or manufacturing on factory floors. It’s such a wide range of things. In fact, I took MQTT home with me about fifteen years ago, and turned my house into a living laboratory so I could try out a whole lot of ideas, which means that I can talk about home automation and home energy saving in the context of MQTT and IoT, and people can easily make the leap to their particular industry. Everybody’s got a house, everybody understands about the need to save energy, people have mice problems in their houses so they understand the need for an electronic mouse trap that sends out an MQTT message, which goes to a tweet when you catch a mouse, which I have in my house of course.
Learn more about MQTT as defacto standard for ISO messaging, read the press release.
Discover what Andy’s been up to on Dinosaur Island, and get started using Watson IoT and Node-RED (Recipe).