Edge to Core to Cloud - CTO Advisor Road Trip in a Telsa Model S
Transcript
(upbeat music) (engine revving) >> All right Greg, we're in a ultimate edge machine, like Elon and the Tesla group, they've optimized the experience for data collection and processing in a Tesla Model S. We talked about that in our last conversation. But when we're thinking about this edge to core to cloud theme that almost every enterprise IT vendor has, I can't help but draw parallels to what Tesla is doing with their million devices on the road. Help me kind of put some perspective and context to this for the regular enterprise.
>> Well, as I think about this, obviously Tesla is very singularly focused on full self-driving. In achieving that mission of full self-driving. And that's a real world use of AI. When we think about hybrid cloud and we think about purpose-built driven solutions that (mumbles) for the enterprise and whether we're talking about healthcare industry, financial services, automotive, you pick the vertical, they have very specific use cases and needs and what they can glean from Tesla is how they laser focus to drive value out of that data.
And it's kind of funny that we're having this conversation this week. As you'll recall, over the holiday weekend, Martin Casado and Sarah Wang came out with their analysis of the enterprise and their perspective on this, which is kind of interesting was the valuations of publicly traded software companies were being depressed by the actual cost of goods from a cloud perspective, which is the kind of the foundation for where most of those costs for those software developing companies are coming from. And basically we know the efficiencies that are generated day one when you're a new business and you're spinning up your applications or services in the cloud.
But we think about how enterprises matures or even as startups mature there's opportunities to extract efficiency and value. So you can continue to scale, continue to grow, continue to innovate but also be mindful of the actual cost of what that means. And it opens up the door for cloud optimization. And hybrid cloud to me is a critical central piece of that. Now in their analysis, they call out kind of what I think are one of the more extreme cases, Dropbox, for example.
Where Dropbox- >> That is the extreme case >> Completely repatriated everything. And there's a lot of businesses by the way and Martin calls those out. He can't call them out by name necessarily in his analysis. There's a lot of companies that are doing that repatriation. I'm not an advocate for repatriation. There's a lot of variables that go into that, but let's think of what that I wouldn't call it a half step but kind of that middle ground is hybrid cloud.
What can you manage on premises? What can you manage in the cloud? How can you extract the most value out of that? And I think, some of the digital natives will dismiss hybrid cloud. You should just go all AWS and do everything there. >> My good friend stopped the jeep. So Hall is one of those. He's like cloud is the future and there is no other way to build applications, modern applications today. And I think the research that Martin and Sarah did was pretty timely around this.
In the earlier conversation off camera, we talked about the 80s and time-sharing and how things were and how we always thought that they'll always be time-sharing as a way to administer and provide compute. You made the point that that was completely disrupted by distributed computing. >> Yeah. >> So if we were to think that five years from now your point that the way that we do cloud today is the way that we're going to do cloud five years from now. That's highly unlikely based on the history of our industry.
>> Yeah. And again you don't have to be a rocket scientist to appreciate. And as you described that pendulum swings from centralized to decentralized, decentralized back out to edge. This is a continuous momentum that is happening in the enterprise. So if you're an enterprise company what are you supposed to do? And I think part of the point that that paper that Sarah and Martin did is to understand where you're going to drive that optimization. It may not be hybrid cloud.
It may be repatriation. It may be purpose-built infrastructure. We talked about AI before and standalone storage and compute to do very specific activities. Think about different pharmaceutical company. If you're a healthcare company that is doing image diagnosis and other applications. If you're a bank that's managing risk. These are all purpose-built applications. So it depends. It goes back to the point we started with. The right tool for the right job at the right cost. And this is something I've been a big advocate for for at least a dozen years or more.
And it sounds so obvious. I think people dismiss it because it's so superficial. Of course, the right tool for the right job. Well, if you're talking about only one way to cloud you're not talking about the right tool for the right job. And this is not a critique of any of the hyperscalers. >> Right. >> This is about extracting value for what your business is trying to develop. >> So what it sounds like you just did to me, Greg is you just took a bag of a million Legos (Greg laughs) and threw out on the table.
And then just say, you know what? Build a millennium falcon. How is NetApp helping me with this big, a million Legos because as a architect who I've relied on my vendors to be very prescriptive in how I build my infrastructure, what you just told me basically is that I need to be a little bit more freewheeling and thoughtful about my approach. >> I would say freewheeling I mean, but to be prescriptive of understanding that you've got to build aglity into the architecture of what you're building and you've got to-- This doesn't mean it's the wild, wild west.
What this means, is that you're giving yourself options. Whether that application is going to live on premises, whether it's going to live in the cloud, whether it's a combination of hybrid cloud. And again, I strongly reject the notion that there's only one way to cloud and it's all in on public cloud. For a lot of applications, for a lot of business it absolutely is. >> So how does NetApp give me that flexibility? >> So if you think about the innovation that we've done around the data fabric, if you think about what we've done with hybrid cloud where it gives you the opportunity to build applications on premises, test them on premises, deploy them to public cloud.
It gives you a lot of flexibility to control how you're managing your data. And that includes simple old-school things, Keith that we talked about like backup, archive, security and how do you manage each of those elements as part of that data fabric. And again, having options, having the flexibility not being all locked in on a provider or a set of services. That you control your own destiny and you're picking the right-- Again, the right tool for the right job for the right cost.
>> So Greg, I really appreciate you giving me a ride and I'm going to call this the Big Data mobile. (Greg laughs) >> You saw the license plate? >> Yeah, I'll flash the license plate to tell people why. If you want to learn more about the CTO Advisor Road Trip you can follow us in our journey. Me and Melissa's journey. com/road-trip. You want to learn more about me or you know what communicate with me and you're vehemently opposed to this concept that cloud will change.
(upbeat music)
