Hello my name is Peter van Eijk and I am one of the world’s most experienced, independent cloud trainers. Let me talk to you about why cloud computing is the future of IT, and why it might well be the future of your job. IT has always been a pretty dynamic environment but cloud computing may well be the biggest change the industry has faced so far. What do you think? Is cloud computing the next wave in information technology? Is it a load of marketing hype? or is it a jump into the deep? I think there’s truth in all of these, but personally, I think that cloud computing is primarily a state of mind, a way of thinking. Let’s assume that you work with IT you’re an IT manager, a systems, or a solutions architect, or any other IT role, or you are involved in procuring IT, or as an auditor, or on the legal side of IT. I will give you a new perspective on how to organize IT, and that view will allow you to figure out which part of your job will disappear, and which part of your job will grow. Cloud computing is a revolutionary change or is it? Maybe, cloud computing is following the same rules as other maturing industries, and to understand the future we must first understand the past. The first commercial applications of IT were on the mainframe a big single computer filling a large room. There was one department running it and was serving an entire company, and it had lots of people managing it. This is the kind of hardware that I wrote my first computer program on in 1973. Now we moved to client-server in the ‘80s where basically every department had a computer, and each computer had its own system manager, so ownership changed here as well as the way in which decisions were made on computing. Then came the PC and a little later the Internet. One of the things that happened there was that companies no longer owned all the computers that users worked on. We moved stuff to customers that worked on their own computers, people that owned them. Once again a shift in the way ownership and responsibility for it was organized. Then we get the breakthrough of cloud computing, where companies no longer even owned the servers that they work on. One way of looking at this whole shift is looking at who runs it? Who pays the electricity bill? Is it the central department? Is it each business unit? Is it the user or is it totally outsourced to another company? In a few minutes we will look at how the amount of people changes that we need to run one computer. Just remember that it took a whole team to manage a single mainframe computer. Now each of these steps: mainframe, client server, PC, cloud, is a disruptive change. It is a big step change, it’s not a gradual change. It is a change that really brings us to a different level of understanding and management, and disruptive change is not a trendy marketing word. It’s a solid business innovation concept developed by Harvard Business School Professor Clayton Christensen. According to Christensen– and I’m simplifying here a disruptive innovation is about something that is not as good initially as the thing that it replaces, but it is much cheaper, so it appeals to people who are ‘over-served’ by the original product. Customers who could not afford the original product. And because of this there’s a good business model and the new provider can afford to innovate further. Which means it can be rapidly improving, and when it’s rapidly improving it will eventually overtake the original, and replace it, and drive it out of the market. Let’s look at some examples here: Wikipedia to start with. What was Wikipedia to begin with? It was nothing more than a handful of articles nowhere near the completeness of the Encyclopedia Britannica, but it was good enough for a lot of people and it was definitely a lot cheaper than the Encyclopedia Britannica. Fast forward a couple of years and which is the one that survives? Is it Wikipedia, or is it the Encyclopedia Britannica? The PC is another example, the original PC was a really simple machine. It had only 64 kilobytes of memory. I remember that and I saw one and I was thinking– hey, you cannot take this seriously. This is not a real computer. It can’t even multitask. 64 kilobytes, are you joking? My boot loader is bigger than that. But it has innovated and innovated, and innovated, and we have now come to the point that your average server is a souped-up PC, that is more powerful than the mainframe that it replaces. Now that’s what we call a disruptive innovation. Of course, then we have the Internet replacing leased lines at a fraction of the cost, and all that, and then the next thing up: cloud computing. In the beginning, it did not sound like much, but it has evolved and evolved, and right now we have to understand that cloud computing can actually be better, cheaper, and safer, than the style of computing it replaces. A few years ago for example, I was measuring the uptime of cloud services, and back then they were already better than most corporate IT in uptime. Each of these disruptive innovations has something you might call a tipping point. For a long time the world is on one side and suddenly like a seesaw it has tipped to the other side, and once it passes the threshold most of it suddenly flips. You go like, “Hey what happened?” And if you’re one of the people whose job is involved in the change you can complain about it. You can say: it’s a change for the worse. The mainframe was such a good machine. You can start complaining about it, you can keep complaining about it, or you can try to find out what you’ve learned in your last job and carry that over to the next generation of technology. I think you can do that, but that’s a different story. The end result for this – for IT – is that it’s not your daddy’s data center anymore – you know, the big fortress on the hill protected by our large moat and big walls – it is no longer there. This is no longer the model for IT. Digital data is no longer confined to the walls of the data center. It’s the whole world that’s being connected, and your systems are basically plugged into that giant world wide grid of interconnected machines. Now what makes these developments happen? Well, the biggest driver for information technology has always been Moore’s law, by Gordon Moore. Where we can see that every two years, the number of components in the chip doubles. As a result the capacity in terms of processing and storage increases by more than an order of magnitude every decade. In my lifetime I’ve seen the price performance ratio of computing, storage and bandwidth improve by a factor of 1 million. And that drop in price goes hand-in-hand with an increase in the amount of IT stuff that we have to manage. And you can imagine that’s also like orders of magnitude more. So what does that mean for the number of people that have to manage this? Is that increasing by orders of magnitude as well? No, it’s not because we don’t have that many people. The productivity of managing that IT really has to go up and it actually does go up. We started out with 10 people managing one computer. A little later it was one person managing one computer, then one person managing 10 computers, and now we see for specific functions people manage thousands or even tens of thousands of computers, so that productivity has to go up. At the same time we need a lot more skills to run that stuff. We have more technology, we have more components, more programming languages, more interfaces, more different devices, more existing software, and our IT is having more of an impact on business and society. From the back office it expanded into the front office, and from the front office it moved to being the product for some companies, like Netflix or Uber or Airbnb. But just about every company these days cannot go back to manual order taking, or manual production control. And then, the threat landscape in terms of the number of people that are trying to hack and abuse that technology is also increasing rapidly – not to mention their skills level. And that means that we need a lot more brain power to manage all that IT. A lot more brain power. And it’s no longer feasible anymore to run that IT entirely in your own shop. We need a way to organize IT that does not force us to deploy 100 skilled people for a 100 person company. Doesn’t work anymore. Now let’s look at how that problem has been addressed historically. Henry Ford, allegedly invented or at least made big the assembly line method of automobile production where a lot of people were working on very specialist jobs in the assembly line, and that made productivity really go up massively. In IT, it’s no different we have specialists for just about everything; network engineers, server admins, storage admins, database administrators, front-end and back-end software developers, and so on. But the parallels go much deeper. In this picture here you can see how Ford’s assembly line was laid out in its River Rouge factory. The factory was build out between 1910 and 1920 but it’s still there. The River Rouge plant was the first full assembly line cranking out cars. On one end of the factory through the River Rouge ships would come in with iron ore, on the other end T-Fords would come out, in between was the assembly line, and everything needed to produce that car. The Ford company even had its own steel mill, its own electricity plant, its own glass factory; everything was there on the River Rouge plant. At the time it was the largest industrial estate in the world, but after the assembly line was introduced other changes started happening. Fast-forward a couple of decades later, and this plant is no longer exclusively owned by the Ford Motor Company. What happened was that that division of labor was extended beyond the company. Components of the whole factory were outsourced. The Ford Company no longer owns a steel mill, it is outsourced to another company with more steel mills. It no longer owns the electricity plant, that’s now part of a large electricity company. Again, we see specialization and this time beyond the borders of a specific company. The assembly line has now become part of a supply chain of companies each of which delivers to the next company downstream in the flow of goods – specialization of labor above the level of an individual company. And currently only 20% of the value of the car is produced by the car company. The rest is outsourced. The same thing has happened with Apple, the original Apple I computer was developed around 1976 by Steve Jobs and Steve Wozniak, so the first order was 50 of these by an electronics shop. And they soldered these together on the kitchen table in Steve Job’s parents house. Again, fast forward a few decades and this is what the iPhone supply chain looks like now. It is produced around the entire world. Apple no longer manufactures these it’s all outsourced, and the only thing that Apple does these days is design them, and market them. They are produced in global supply chains, but this also creates its own new problems. In a supply chain we can no longer solve production glitches by having one department head talk to another department head. We need to make the control over that supply chain more explicit. I ran into an example of that last year when I was delivering a course in Tunisia. At night in the hotel I chatted up with a German guy. Turns out he was there to audit the supply chain of a big German car manufacturer, and the company in Tunisia was producing a part of the electrical system of the car – not even the whole electrical system. And the guy was out there to audit their process. He didn’t even look much at the product only at the process. If the process is good the product will be good, and he was coming back every couple of months or so to audit that part of the supply chain. He also explained that if this company stopped delivering the assembly line in Germany would stop in 48 hours, so they had strict delivery agreements and fines when they failed. The upshot of all this is that today’s car is not only cheaper comparatively, it is also more fuel efficient, more functional, safer, it is better in all dimensions, and you can even get it in another color than black. And this is how specialization of labor in a supply chain works. Now with cloud computing it is a similar thing. Cloud computing, basically creates an IT services supply chain. Here’s a diagram that shows how that might look like – simplified. As a cloud customer you consume services from a Software as a Service provider – and a lot of these SaaS providers, actually, have their own infrastructure sourced upstream. Of course, the reality is a little bit more complicated than this. It’s more looks like this picture, where you have a ton of SaaS providers – the average company has a few hundred. Similarly a lot of platform providers upstream, and then it’s all hosted on Amazon– no, no, I’m just kidding . There’s also hosting on Microsoft, Google and a few others and some SaaS providers actually host themselves, so there you have it. The IT services supply chain. The new model for delivering IT services; better, cheaper, safer – potentially – more complicated, more moving parts, more people to manage it, different ways to control it. And now the fun starts. Once we have this, what happens? Once we have this new things become possible, new things that might include your future job role. Once we have programmable infrastructure we can do automated integration, automated testing, automated deployment, we can do DevOps. Once we have flexible capacity at scale we can do big data, once we have that stuff out there on the network, it makes more sense to develop social and mobile applications for individuals, as well as for companies, and as a consequence, it has never been cheaper or easier to start a software company, or set up any company’s IT, for that matter. I reckon that if you put your mind to it even with a team of just two or three people you can start a company, including its entire development pipeline in less than a day. Just by procuring the right cloud services. Let’s focus on one of these. The one that’s really driving this IT transformation from the business perspective is the DevOps revolution. Back in the old days, a new version of the software was released once a year and that took a while to implement. It had to be installed on all the machines, databases migrated, and so on – very complicated. With DevOps you can automate a lot of the entire software development and deployment pipeline. This is called continuous integration or even continuous delivery, and it allows companies such as Amazon, Netflix, banks, dotcoms, to deploy new services or bits of services literally thousands of times a day. And that creates for their business enormous agility, enormous feature velocity, which allows them to innovate quicker, and better, and adapt quicker to the market. And as you remember from Darwin it’s not the strongest of species that survive, but the ones that can adapt to changing circumstances quicker, and it’s the same for companies. Now if you are in IT or related to IT, what are you going to do with this? What should you? This is a new world; this is inevitable, it is already there. This is the new way that IT will be ran. Of course, it will take another ten years for this to reach 95% saturation, but there’s only one way, there’s only one direction that this is going. We still have mainframes but there’s not a lot of work on mainframes, we still have mini computers, but there’s not a lot of work in mini computers anymore, we still have PCs, but there’s not so much work in managing these anymore, in fact, we’re often leaving that to the users. Cloud computing is where the new action is. If you want to succeed in that world you have to know about the nuts and bolts, how it is pieced together, you have to know how to control suppliers, and everybody in that ecosystem – everybody upstream from you – and you have to demonstrate to the people downstream that you are in control. You have to know how to control all kinds of risks and you need to know where you fit in that ecosystem. Don’t be stuck, and once you get it there’s a great new world out there. Do you want to know more? Subscribe to my YouTube channel, or sign up for my e-book. My name is Peter van Eijk, and this is the inevitable coming of cloud computing.