a16z Podcast - a16z Podcast: The Consumerization of IT

Episode Date: July 31, 2014

Peter Levine goes deeper into the changing nature of the data center with Yoram Novick, CEO of Maxta, and Florian Leibert, CEO of Mesosphere. Does ARM architecture, the foundation of the chips in practically every mobile phone, have what it takes to displace X86 processors in the data center? Is the private cloud --- effectively a company operating its own data center -- dead, or will a hybrid public/private cloud model dominate? Given all the upheaval in the data center how do startups fit into the equation, and how are they competing today with legacy technologies and companies? The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation. This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments and certain publicly traded cryptocurrencies/ digital assets for which the issuer has not provided permission for a16z to disclose publicly) is available at https://a16z.com/investments/. Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

The content here is for informational purposes only, should not be taken as legal business tax or investment advice or be used to evaluate any investment or security and is not directed at any investors or potential investors in any A16Z fund. For more details, please see A16Z.com slash disclosures. Hello, I'm Peter Levine here at Andresen-Harrowitz, and I'm with Yorm Novak, who's the CEO of Maxda and with Flo Liebert, who's the CEO of Mesosphere. Great to have you guys. Thanks for joining. It's a pleasure.

Thank you. So I wanted to spend a little bit of time talking about the consumerization of IT. And by consumerization, I don't mean bringing your own device, you know, mobile device to work. like that's happening and that's already done with. What I'm specifically talking about with respect to the consumerization of IT is the idea that the components of mobile devices will become the basis for hardware in the data center of the future. And so not unlike we saw the PC revolution replace mainframes in the data center,

the belief, I have this strong belief that the innovations in mobile hardware over the past five, ten years will replace sort of the PC architectures in the data center to where the mobile supply chain effectively eats the hardware of the PC world in the data center. And you can imagine, you know, a billion mobile devices being the basis of the hardware infrastructure. And then, of course, you know, the belief is that then taking all that hardware and putting it together requires some level of sophisticated software that then would aggregate and bring

kind of the enterprise class capabilities out of this underlying commodity hardware. So what do you think of that? Like, you know, how have you thought about this in your own business? Maybe start with flow. Sure. So I think Google in the past has already shown that they're using commodity off-the-shelf components in order to build their entire data center infrastructure. structure. And I think with open compute at Facebook, we're seeing a similar trend across the

industry now where everybody is using commodity components. And of course, that means that there are more failures. So software needs to be able to handle those failures. We're going to see plug-in play data centers where you can just add chips, remove chips, and they're just added to the pool. I think that's certainly a trend. Yes, I think that we see basically the similar trend in the storage side. So if you go back to the past, storage was originally proprietary system that were custom-build. And over time, people started to use commodity hardware to build new storage system and the value moved to software. And those are the X-86 standard PC hardware. And there is no reason why the next generation of hardware platform, which is

basically arm processor and stuff like flesh coming from the consumer side will not take over because any time commoditization happened in the hardware, from a cost point of view, from an advanced technology point of view, it's always winning. And what you need to make it all work together is smart software that put all the pieces together and make relatively unreliable hardware work in a better way than it would work without this sophisticated software. So do you guys, you know, we talk about arm and flash as kind of principle. There is no spinning disk in a mobile phone. There's, you know, most mobile phones don't have intel processors. They have different processors. Do you really think that these components like become the basis

of the data center? And, you know, if I'm, if I'm a existing company and the data, data center space. I'm like saying, well, that stuff will never work. Like, you know, we can't expect commodity an arm chip, really, to power the data center or for commodity flash that's going to fail all the time? Do you guys, like, what's driving that trend? And do you really think it's going to happen? I think it's definitely going to take place. And here are the reasons. If you look at the challenges in the data center today, there are around power and cooling. And arm, by its nature, provide a better footprint. So it is true that today arm processor is not as powerful as an x86, but as the PC overall, in majority of cases,

replace the mainframe, because you can always have a bunch of PC that are performing as good as a mainframe. The same thing, I believe, will happen with processors where arm processor, which is much more or much better in power and cooling over time will become powerful enough. So that's on the side of the processor. Flash is definitely going to take over. Again, power cooling and especially performance of flesh, even if reliability is not there. And software is going to bridge the gap and basically provide the performance that you need with the reliability that you need and overcome the challenges that you have today with commodity flash, which is still not as reliable as this drive, but the software and the commoditization

of hardware and the lower cost will enable it to be as reliable or even better than this drive. Yeah, I don't have too much to add to this, but I think on the mobile phones, we see that you obviously have to optimize for power consumption, right? Like you have a battery, you want to optimize for battery life. and so the effect of course of that is that there is the most innovation happening there if you can use this innovation to drive your data center where you're limited to a 65 megawatt uplink basically right

you can you can put a million cell phones into this data center or two million however many you need and just get actually much more compute out of it and I think with Flash the same right like for Flash you probably don't need much less power than for spinning disk we have we have mechanics involved maybe um you know one of the one of the questions on our minds all the time i'd love to get sort of your thinking on this is how do you think the data center

evolves and meaning is it going to be on-prem like do people take all their existing stuff and move it over to this new architecture is it sort of private clouds and public clouds off-prem what do you guys think about that and how do you you know kind of for enterprise customers which both of you are, you know, selling into, what do you think happens in the future? What does that look like? So I think it's going to start out by some companies deciding, okay, we're going to have our on-prem solution

and we're going to have some scale-out capacity in the cloud. But I think going forward, it'll be much more fluid. It'll be – I think as we are seeing already with Amazon, there's a spot market, right, for compute and for storage, for your entire data center, essentially. And I think as we see that with more and more companies evolving, we're going to be pricing jobs, individual compute jobs, rather than just providing a static pool of resources.

I think that's where we're going clearly. I have no doubt that over time the cloud model will take share away from the private cloud. The private cloud will be there for a very long time, especially for larger enterprises because the performance and the control is required in a lot of cases and the public cloud doesn't provide all of those. I think that the model that will take places probably in most cases, a hybrid model where some applications, some use cases are going to be on-premise and many of them will be in a cloud.

But nevertheless, the data center, whether it's on-premise or a data. data center of a cloud provider is going to follow the same theme, which is commoditization of hardware, the best available hardware, the best in terms of price and performance. And I've no doubt in my mind that because of the numbers of mobile devices, the components coming from the model will eventually take over in the same way that the, you know, 10, 15 years ago, the stuff coming from the PC took over the mainframe and the mini-computer. But in terms of whether it's going to be cloud or on-premise, I think over time more and more we'll move to the cloud,

but I don't think the private cloud is going away anytime soon. Both of you are CEOs of startup companies, and startups typically sort of exist in the cracks of where incumbents can't get to, and then hopefully startups like, oh, I'm hoping that both you guys become really big, and you become the next franchise in your spaces. But tell me, like, you know, why can't incumbents make it over? And are you guys afraid of incumbents sort of, you know, coming into this new data center? How do you think about that?

And, you know, are there examples maybe where, hey, you say, like, there's a new space here and we're going to go dominate that or incumbents kind of, you know, have a good shot? I'm just sort of wondering what happens to all the existing. companies in the space, given this transformation into the new data center design. Yes. So I think the way we see it is that the legacy that was built over time created all kind of boundaries. So, for example, there are server companies, and then there are storage companies,

the networking company. And all those boundaries made sense in the past, and a lot of companies build their entire business model based on those boundaries. And this is what really make the... them thrive. And I think that what we will see going forward, the commoditization of hardware will create a lot of convergence where a lot of the software component will be converged to run on the same single platform. And this really create a new model. And many of the existing companies will have to cannibalize themselves if they want to get in, which is always very difficult. So this

is a typical innovator's dilemma. They may know that they have a problem, but in order to solve the problem, they need to give up most of their business, which most companies just cannot do because of their cost structure, because of all kinds of other reasons. So I think there is a huge opportunity for software company, the commoditization of hardware, the fact that everybody realized that in the future, data center will be built with commodity hardware and all of the value will be in software, and converging the software to run on the same platform, get rid of a lot of the legacy, opens up the opportunity for startup, but it's not easy for an existing company that build their entire model on those boundaries to basically compete on the same

level. Yeah, I think if we look at companies like VMware, for example, VMware started out with providing VMs for Windows, essentially, and we're seeing this tectonic shift right now in data centers from Windows over to Linux. And obviously, Linux can actually deal much better with, with these components like arm processors and so forth. And we're seeing this on one end. So, of course, VMware has to actually get into this market

and capture more and more of the Linux workloads. At the same time, VMware is also moving, or I think they will be moving across different types of applications. So some grid engine, for example, doesn't generally run on VMware. I mean, of course it could, but most people don't. HPC generally doesn't HPC workloads generally don't run on VMware and what VMR has done

they have really built a control panel for your data center right with a V-sphere product line and they've been trying to integrate more and more areas of your data center into their control panel I think with the acquisition of Nazira they're obviously trying to get into the SDN space

so and yeah I think we're going to see a bunch of these big companies really going into this. But at the same time, I think there's a lot of opportunity in the smaller space cloud. I think is still pretty small, actually. And I think that's where startups can really execute much better, I guess. And I think you're bringing a very good example. So if you take VMware as an example, VMware actually build their business by taking it away from a company like Microsoft.

Okay. And it was a great business model, but even VMware, as a legacy company those days, if we think about Linux container, they may take over because all of VMware business was built around the idea of virtual machine and virtualization and all the stuff around it. You move to a new model, which in a lot of ways is much more efficient. How does VMware going to compete against Linux container? Are they going to give up their existing virtualization business in order to build a Linux container? it's a very difficult decision for a company like Vimmer, but a new startup that has no legacy, has no problem ditching all the old stuff

and saying, okay, Linux container is the way to go. Maybe they'll just start giving away ESX for free, right? That'd be quite traumatic on their bottom line, I'm sure. Speaking of Linux and containers and kind of the kind of trends there, flow maybe talk about how open source is impacting your you know the the data center and some of the innovations occurring yeah yeah so i think if we look at containers right now and docker specifically um we're just gaining a ton of traction the the roots there are actually

well the roots lie actually at google so four or five years ago google contributed c groups to the to the Linux kernel and C groups are the foundation for what is now Docker right it provides the isolation the really strong isolation in the Linux kernel without the need for spinning up a VM and Docker has built really fantastic tools around C groups allowing you to really well essentially have the same benefits of a VM where you can just package your application and just run it anywhere. But on top of these lightweight abstractions without requiring the full VM image to be shipped. And I think that's, I think that this, these building blocks that enable you to build really large, complex systems that are, that essentially can solve big problems. That's really

powerful. And we are leveraging that as well. We are leveraging C groups and containers and Docker. For us, this is a really pluggable model. Yeah, isolation is really important when you're trying to get the most out of your data center. You want to run multiple workloads on the same machine, and you need something to isolate. Yes. I mean, the way when we talk to people in the data center, and yes, Linux is open source and it is taking a lot of the operating system market share. Now we see the same trend in virtualization. VMware and HyperVe definitely are there, but KVM, which is open source, in many cases, especially

with cloud provider, is the main virtualization scheme. So basically, the software company that tried to build their business on proprietary infrastructure are seeing some, you know, competition coming from open source. When the open source is good enough and pervasive enough, it does take over. It may not be as good. For example, if you look at KVM and VMware and you look at the features and function, but if it's free, it has a lot more value and people are building a lot of the software stack on top of open source rather than on top of proprietary infrastructure. So it's interesting, Yoram, that you talk that way about open source. Your company is mostly as proprietary. You build on some of these building blocks and flow. Your company

is open source and building some proprietary stuff. So I'm just sort of curious as to how you rationalize that thinking, Yoram, like, hey, if open source is going to eat the stack, why doesn't it eat you guys, one? And then Flo, when Yoram responds how he's got his thing going, how do you guys actually see open source being a viable business model for you if everyone uses this stuff for free? So maybe let's just talk about that.

No, it's a very good question. So I believe that open source, makes a lot of sense at some level of the stack. At the very basic level. Just not a year level. Well, even look at operating system, look at what happened with Linux. So Reddit makes a lot of money.

So the reason is the very basic operating system is something that open source makes sense, but all of the services around it, people are paying money for it. So at the end of the day, the customization and make it work in a certain environment is something that people value. So

I believe that open source makes sense where you are creating some platform that can be distributed and make sense at a certain level, but at some level, you know, the proprietary makes sense now over time things go up. So

whatever used to be proprietary, over time may become open source. You just need to keep up one level on top of it and providing more value than anybody else can at your level. So it just doesn't make any sense for us to ignore open source. It's there. It's going to be there.

And what we view as our value is providing better customization and solution on top of open source and being one step ahead of everybody else, knowing that one day, even our piece may be eaten by open source and we need to step one, you know, go one step further. Yeah. Yeah. So as Peter pointed out, we affirm. relievers in open source, most of what we have released to date is open source.

And I think it really allows us to get traction. It allows people to try out our software without having to sign a contract. And I think where we can provide additional value that we can charge for us in making this open source easier to use. And coming from companies like Airbnb and Twitter where UX is really essential, I think we can apply the same to open source and I think traditionally open source hasn't been very good about UX

right like I mean if we think about the first distributions of Linux for example I mean the interfaces were just horrible and nobody wanted to use it and Microsoft had this finished product and I think that's really where we can actually build something of value

for the user and I think for that you can charge even if the core is open and I think it's important that the core is open, because as we're seeing now with a lot of security breaches, for example, it's great if everybody can look at the source. If people can discover which algorithm is used for encryption, that's great. It'll lead to much less disastrous. But you don't need this visibility necessarily in the UX layers. And I think the UX is actually a lot of work.

I mean, getting UX right is really, it's an art. And typically system companies have notoriously done a poor job in UX. It's always been one of these things where, like, you're a system engineer and somebody else is going to worry about the UI because I got system problems to deal with. And so I think you're probably right, like combining some of those elements together and thinking about the design and form and function of how some of this operates in the next generation. of data center companies, I think that's going to be ever, ever more important. Absolutely. I think when we started,

we talked about the consumer at the hardware level. I think we need to talk about consumer at the software level. I think that part of being a consumer company is providing very interactive and very efficient and easy to use user interface. And I think that over time, the data center is going to move into the same level of sophistication at the user interface.

In the past, It was all command line. Nobody cared about it. But as the same user that are used to a very nice GUI on the mobile devices are not going to accept a user interface that is not on par. So I think that on the software side, delivering software, the software needs to be good from a system level point of view. But at the same time, it has to provide the same user interface. It needs to work on every device that is out there. And you cannot ignore the user interface, even if you are a system.

company, even at the end of the day, you're providing service to an enterprise. Yeah, those are great points. So maybe to just wrap up here, not only will mobile hardware become the basis of the next generation data center, but it feels like you guys also agree that the interfaces and design elements that we've all come to know and love on our mobile devices may also be a pretty critical part of what happens in the future. So Flo and Yoram, thanks again. And we'll talk with you all soon. Thank you. Thank you.