The Mechanics of Disruption: Adrian Cockcroft on Open Source, SSDs, and DevOps

Sure, thanks for having me on this. The talk really came from there's two pieces to it. The first part was about how to speed things up and the speed of innovation being very important and how to think about what is innovation and what is disruptive rather than what is kind of the incumbent sort of existing companies and the way they do things. And so there's this phrase that I kind of hinged it around which is "it ain't what you know that messes you up, it's what you know that ain't so. " It's a Will Rogers quote and a lot of it's... I have to untie that one. Yeah, so a lot of the disruptions that happen in this industry happen because an incumbent figures out, "Okay, we have an idea, we know how to optimize for that idea. " You build a business model around it and then you go off and build your business and then somebody else notices that one of your assumptions is no longer true and you're optimizing for something that used to be really expensive but now it isn't. And then you... The disruptor basically takes advantage of this thing that everyone else has assumed is a big expensive hard to do thing that has now become cheap and easy.

But you end up disrupting the industries that way and there's a number of examples of that that I went through. Yeah, because the first thing I think about is databases. You know that we have Oracle and MS SQL, that they're... Microsoft, they have these big monolithic products. But then along comes open source products like MySQL and Postgres, which are now running huge operations. And before... I mean, there's so incredibly intensely difficult problems they're solving, but they're able to take in and disrupt the business of Postgres. I mean of Oracle and... So, yeah, so really the open source itself is one of those... The concept of open source, the concept that the best software you could buy is the stuff you get for free that's built by a community is something that is sort of alien to the the traditional enterprise software businesses where you have a product manager who comes up with an idea and then they go away and work on it for a few years and then they pop out with the product and try and sell it to you for a lot of money. And historically, that was how you bought stuff because it was the best stuff you could get. And these companies knew... were the places that knew best how to build things. And I worked for Sun Microsystems for a long time and you know, part of that process. But what really happened was the best things you can get now are free. And they are built by communities and the best engineers don't work at those enterprise companies. They work for end users. So, you know, the most interesting, scalable, reliable software now are coming from open source projects. And I highlighted a few of those in the talk. So that's one disruption. Another one I mentioned was solid state disk replacing, you know, spinning rust to traditional disks and that the way that that has completely taken what used to be a it's a multi-trillion dollar business or whatever it is in storage. And if you look at the way that a Cassandra-based deployment on AWS is, there are no storage admins. There's no SAM. There's some SSD inside an instance. And that's all there is. So the person managing it is probably a Java distributed systems programmer, developer type of person with a little bit of operational experience to automate how to keep that thing up and running. But there's no, the whole layers and layers of sand fabrics and storage arrays have completely disappeared. So it's not just that you disrupt the product, you make the entire product category disappear. That's one of the most disruptive things that's going on right now. I have to think about when you're talking about trying to change the mindset of people who've built their business, their entire business on a certain set of assumptions and looking at more about just a human quirk there. Just reading an article about how people who have deeply held beliefs that they've internalized into their own persona, their own personality. And that's the only way to get them to change their thoughts. You can't just say here's a fact that what you know is wrong. You have to just change literally the way they think and address their personality. Is that something that when you think about when we're dealing with these companies that have these deeply embedded, entrenched, I should say, not embedded, but deeply entrenched processes like this is how we build software because this is how things are and this is how we develop products. And how we can approach, do you talk about a little bit about how we can approach those kind of thought processes and address them and work within them?

So I talked about a few other things. I have time, you know, I had a whole hour, so there was a lot of things I covered. One of them was sort of continuous delivery as another way of disruptively speeding things up. And the assumption you're making there is that you can move a lot quicker. But what it really comes down to is the DevOps kind of approach, where you're actually taking away the handoffs between teams. So everyone knows that developers have to stay in development and operations. People run their code and they're different kinds of people and they optimize for different things. But if you reorganize so that you put them in the same organization, you blend those responsibilities and you automate everything. It's not just that you create a DevOps team, it's that you actually have to reorganize your company and have a cultural change. So DevOps is in many ways a cultural problem. And companies that get over that are the ones that are doing well.

" That was the question. My response was it's the other way around. The enterprises that are ready for DevOps are the ones that are going to disrupt the enterprises that aren't. Oh, okay. So they're going to just knock out the ones that aren't. Yeah, it's much harder to compete if every time you do a product release, it takes you a year to get it to customers and your competitors doing it in a few weeks and learning and they'll run rings around you. So that is the sort of existential challenge that a lot of companies are getting through right now. And the other big change is that software used to be a little thing on the side for a lot of companies and now it's central. It's central to everything they do. It's central to the way they do marketing and sales and build things and the way to support customers and the way they deliver products. Internet of things is just one aspect of that. But what you're seeing is this whole software is the world driving people and now it's central they have to get good at it. And so they're looking to the big enterprise companies, the big web companies as examples of how to do that. Netflix is one example. It's a very, very rapid development process and Google, Twitter, Facebook, all of those companies are leading in this space and the enterprises around the world are trying to figure out how do we learn, what do we learn from them and how do we adopt it. So that was kind of the first part of the talk. The second part was some more practical things about what that actually means in practice for a developer. There's some techniques and I was trying to make it a little bit more hands-on and a bit more detail. So the talk was kind of a little in two sections. The reason that what I was about to say is that I interviewed Bill Scott who's with PayPal and he talked about how it was fresh off of the acquisition of Braintree and how PayPal is trying to rethink their entire business process and they're looking at companies like Braintree who were able to deliver these things. A lot of people in the press that I heard were looking very skeptically on PayPal acquiring Braintrust and thinking, "Okay, well, I'm Braintree. Well, that's going to bust the Braintrust at Braintree. " Say that 10 times. But when he described it, they said, "No, we're looking at how they do things and really trying to figure out how to bring those into our operations in a nutshell. " Sure. Yeah, I worked with Bill Scott when he was at Netflix and I actually worked at eBay, PayPal back before I joined Netflix. So I get some of the challenges there. You've got a successful business and it's working and then you have to, you know, PayPal has this big challenge from other financial processing stuff like Square and Stripe and whatever. Those companies are, you know, we're more Agile and PayPal's trying to respond to that and they're doing a pretty good job and what Bill's doing there is good work to move them strongly in that direction. It's good to see. If you were to go and just flip a bit in everybody's head that was in your talk and have them take home at least one factoid, what would, or one idea, what would you like to have impressed upon them the most?

Look at, look at the way you're operating, what you're doing and what assumptions you're making. And if you're assuming something and optimizing for it and then you look at the trend for that thing and you know, it's either going to suddenly stop being true or over time it's going to stop being true. Make those explicit and get out ahead of those examples. And sort of what the one of the key things that Netflix does, it tries to lean into the future. If there's a trend towards cloud, they'll get in early and they'll dive in and they'll learn how to get good at it before it's really ready for mass consumption. And that is a very powerful technique for learning how to take full advantage. So once it becomes mass adoption, you're already well optimized for it. So you have that advantage.