 Live from Las Vegas, it's theCUBE, covering AWS re-invent 2017, presented by AWS, Intel and our ecosystem of partners. Hello, welcome back. This is live coverage of theCUBE's AWS re-invent 2017. Two sets, a lot of action. Day one of three days of wall-to-wall coverage. I'm John Furrier with my co-host Keith Townsend. Our next guest, CUBE alumni is Yaron Jave, who's the founder and CTO of Guazio, hot news startup. And big news coming next week, a big announcement and following their work. Yaron, great to see you again. Thanks for coming back on. Hi, thanks. And you got a new shirt. We'll share the logo there. Nucleo, that's our new serverless framework, which is open source, really kicks ass. It's about, you know, 100 times faster than Amazon, and we're announcing it next week. It says it's 200 times faster. Yeah, we don't want to shame it. Set the bar, you kind of covered it. There's a mercy rule. Yeah, we do about, then we're doing 400,000 events per second on a single process. They do about 2,000. Like most of the open source projects are on the same ballpark. So, Yaron, I got to get this off the bat and then we can have a nice discussion afterwards. A pleasant discussion, serverless. Let's first define what that means. Because there's a bunch of, I can take Nucleo, install it in my data center, run it, am I serverless? Yeah, you know, so we, you know, I'm in the serverless working group, and- For CNCF, right. Yeah, and for CNCF, and we had a hot debate between the serverless open source startups doing what is called sort of function of the service and Amazon and Chris and others trying to push the notion of serverless, which is serverless stands for serverless, meaning you don't manage server. And the way we position Nucleo is actually both because on one end you can consume it as an open source project. Very easy to download single Docker instruction and it's up and running, unlike some other solutions. And on the other end, you could consume it as something within the Iguazi data platform that has all the, if there is a slide from Amazon which I really like, which is about serverless, and they show serverless is attached to Kinesis, DynamoDB, S3, and Athena, okay. Four services of data that attach to Lambda. Iguazio has API compatibility with Kinesis, with DynamoDB, with S3, and Presto, which is Athena as well. So exactly the same four data services, the position as part of the service ecosystem are supported on our platform. So we provide one platform, all the data services that Amazon has, or at least the interesting ones, serverless functions which are 100 times faster, a few more tricks that they don't have on that slide. So what is the definition then? Just in a pithy way for someone out there who's learning about serverless, what is it? What's the definition? So the notion is as a developer, you're sort of avoiding IT. You go, you open a nice portal, you write a function, or you write your function in a GitHub repository somewhere. You click on a button and it gets deployed somewhere. Right now you know where it's going to get deployed. In the future, you may not know. You got to set up an EC2 instance, get that prepared. It's not really an EC2 instance. No, the old way, the old way. The old way was, right? Yeah, the old way were infrastructure guys building your EC2 instance, security layers, middleware, et cetera. You go develop on your laptop and then you need to go and conform and all the continuous integration play was very complicated. Serverless comes in inherently with scale out, with auto scaling, with continuous integration. You have versioning for the code. You could downgrade a version. You could upgrade a version. So essentially it's a package version of a cloud native solution. That's a general idea. Yeah, so I can do that. If I'm doing it and managing myself, it's functions as a service. And if I'm doing it and it's provided as a cloud provider, as a server, as a service, it's serverless. There's none of my operations team is dealing with servers. It's just writing code and just go. Yeah, you're writing a function, push, commit. You know, you should play with nuclear, not just with other things. But you'll see you're writing a function, even have a built-in editor. You write, you push deploy, and it's already deployed somewhere. All right, so give us the perspective before you move on on the game, what the impact is to a developer? Apple store oranges. I mean, old way you describe it, new ways. Sounds easier. What's the impact? How much is it time, money, what? I mean, can you quantify and give some? I think the biggest challenge and you guys write about it, okay? The biggest challenge for businesses is to transform. So, an interesting sentence. It's not about digital transformation. It's about businesses that need to work in a digital world, okay? Because again, most of the communication of customers to businesses is becoming digital, okay? Whether it's today through mobile apps, tomorrow through Alexa, okay? As Luke Sturney said, it's all software. Your business is the software. But it's all about interactive, really, okay? So, as a business, I always position there are two things you need to take care of as a business. One is increasing the revenue, not by engaging more customers and increasing the revenue per customer, okay? How do you engage more customers through digital services, okay? Whether it's Twitter's or providing a new service for your web portal, et cetera. And the next thing is how do you generate more revenue from a customer is by showing recommendations on, you know, what he wants to buy. Providing more value. Right. And the other aspect is operational efficiency. How do you automate your operations to reduce the cost? So, you know, Amazon's use robots to do the shipping and packing so their margins can now, you know, be lowered. So, the general idea is both those things. Reducing cost is becoming more and more dependent on automation, which is digital, and reducing increasing revenue become more about customer engagement, which is digital, okay? Okay, so now you're a traditional enterprise. You have your exchange to worry about and all the legal stuff and the mainframes, but if you're not going to work on the transformation piece, you're going to die, because some other startup is going to build insurance company, which is sort of, you know, agile and all that. So, you made an interesting comment earlier when you were talking about Nucleo and integrating the functions that really matter, the services that matter. Amazon releases 800 new services a year. Actually, 1300. I'm sorry, 1300. This time, less, no? Right now, they're at 1130 and they expect to be 1500, 1700 by the end of the year. Two years ago, it was like 750, and then a year before, that was like 600. So, is that an indicator as to Amazon's leading this race between the big, I don't know, three, four cloud providers, rack and stack them for us? What, how do we assess the capability? It's a matter of mentality, okay? Bezos thinks like a supermarket, okay? Just like an Amazon market. I could say I need a, you know, cover for my iPad, I'm going to get 100 covers for my iPad. No one really, you know, I need to now choose, okay? So, their strategy will put dozens of services that do similar things. One is better at this, one is better at that. We control the market, we'll sell more. We have a different approach. We do fewer services, but each one sort of kicks ass, okay? Each one is much better, much faster, much better engineered, okay? This is also why we have one data platform that provides 10 different data APIs and not 10 different individual data platforms. All right, so let's talk about the scoreboard. So, okay, even though they might be thinking of supermarket, you got Amazon, Azure, Microsoft, and Google. I look at some of the data. I mean, Microsoft's been international for a while from their MSN business. They now have Skype, they got data centers. They know a little bit about cloud. Amazon's got a lot of more services. They support multiple versions of things. Google is kind of non-existent on the scale of comprehensiveness. So, isn't it a game? And then you look at their serverless functions, by the way. I don't know, in terms of new stuff, TensorFlow and serverless. No, but serverless, they only support Node.js. They have very few triggers and it's still defined as beta. Yeah, well, I mean, this is the point. I mean, so people are touting, I wrote this in my Forbes article, they're touting like a feature and there's a lot more that needs to get done. So, the question I have for you is, there's a level of comprehensiveness that you need now. And I know you guys spent a lot of time building your solution. We've talked about this under our last CUBE interview. So, the question is, the whole MVP cost of minimum viable product is great when you're building a consumer app for an iPhone, but when you start talking about a platform and now cloud, question to you is, is there a level of completeness bar to be hurdled over for a legit cloud or cloud player, what's your thoughts? I think you need a thousand services to build a good cloud, but you do need a bunch of services, okay? Now, the way that we see the world, and like Satya, okay, which is there is the core cloud, but there is sort of the belt around it, which is what he calls intelligent cloud. We would define ourselves as the intelligent cloud. So, if someone is building a machine learning model and needs a five-year worth of data, I need to just do crawling on top of it. It's not really an interesting problem. It's commoditized, lots of CPU power, you know, object storage, but the bigger challenge is doing the inferencing close to the edge. This is what needs to happen in real time. So there, you need fewer services, but there need to be more real time. Smart integration, to do that, right? You need, you have density problems. You don't have a lot of room to put a hundred servers. It needs to be a lot more integrated. You know, look at Azure Stack, okay? You know, the slogan is consistency, okay? Look at a slide that shows which Azure services are part of Azure Stack, less than 20%, okay? Because it's a lot more complicated to take technologies that were designed as hyperscale and put them on few servers. So how- And the customer's figured it out. What does the customer do? It's all mind-boggling. I look at that and you, I love the concept of core services and then value around those core services. What are those core services that a cloud must have before it starts to invest in that cloud provider's strategy? Yeah, so the point again, there's a lot of legacy that you need to grab with you, especially someone like Amazon, okay? So they have to have VMs and migration services from Oracle, et cetera, okay? But let's assume I'm a startup and building a new cloud native applications. Do I need any of that? No, I can probably do with containers. I don't really need VMs. I can use something like Kubernetes. I can use no SQL databases. You know, maybe some MySQL for, you know, so I could redesign my application differently with a lot fewer services. The problem for someone like Amazon in order to grow and be a supermarket, you have to have 10 of everything, okay? But again, if I'm someone that focus on new applications, I don't need so many services and so much legacy. Well, I will say one thing. You can call them a supermarket, use that retail analogy. I buy that analogy only in the extent that you used it. But if that's the case, everyone's hungry for food and they're the only supermarket in town. Yeah, but hopefully maybe have less stuff on the shelf. I mean, everyone else is like a little hot dog stand compared to the supermarket. I mean, Amazon is crushing it. Your thoughts. Yeah, but again, I think- I say that, but are they kicking ass? I mean, 40% growth- Obviously Amazon is kicking ass, but I think Azure is ramping up faster. Amazon is generating more alienation among people that is starting to compete with, you know, like one- They're just copying Amazon. Right, but they have a different angle. They know how to sell to enterprises, okay? They already have the foot in the door for Office 365. I've talked to a customer, okay? And I thought, we're going Azure. They say, why? They say- Office 365. They said, we already certified the security for the 365 for us to use Azure, it's a good. So the- That's another service. Yeah, right up until that next breach. Right, so the guy's owning IT, it's easier for them to go to Azure. The developers want Amazon. Because Amazon is sexier. That's the question, I know we got a break, but I want to get asked this to you. We debated this on the intro segment with the analyst. Question, IT buyers have been driven by a top-down, CIO driven, CXO driven, waterfall, whatever you want to call it, old way. With developers now at the driver's seat, with all this kind of serverless function as a service coming around the corner, very fast, are developers driving the buying decisions or not, or is it still IT? The budget's still there, they want to eliminate labor, they want more efficiencies. Are you seeing it yet? Will it happen? Yeah, because we're just in the middle, okay? Because on one end we're an infrastructure, but we're an infrastructure consumed by developers, okay? So we keep on having those challenges within the accounts themselves, you know? IT doesn't get what we're doing, you know? Serverless, you know, and databases is a service. You're speaking a foreign language. And because they like to build stuff, you know, they want to take the new TANIX and take a hundred services on top of it, and it will take them two years to integrate, by that time sort of the business already moved somewhere else. So IT could be a dinosaur like the mainframe. Right, but I think the smart ITs understand that they need to adopt cloud instead of fight it and move the line further up the stack, okay? And that's sort of the thing that we're trying to provide to them, you know, when you're building stuff you're buying EMC storage, okay? You're not just taking disks. So why do you focus on this low level of block storage when you're buying infrastructure? Why not buy databases as a service, you know? And then you don't need all the hassle, you know? Streaming as a service, serverless as a service, and then you don't need all that stack. Yaron, you should be our guest analyst. We're too busy building a company. We're going to see you next week in Austin for KubeCon, congratulations on your success. I know you guys have worked hard. Yaron Havie, founder and CTO in Aguazio, you're going to hear a lot about these guys, smart team. Let's see if they don't need, they either going to go big or go home. I think they're going to go big. Congratulations, more coverage here at AWS ReInvent after this short break. I'm John Furrier with Keith Townsend.