 Live from New York, it's theCUBE, covering AWS Summit, New York, 2018. Brought to you by Amazon Web Services and its ecosystem partners. Welcome back everyone. It's theCUBE live in New York City for Amazon Web Services Summit 2018 in Manhattan. I'm John Furrier with my co-host Jeff Frick, all day coverage here in New York. Our next guest is Erica Windish, founder and CEO of IOPype, CUBE alumni. Welcome back, great to see you. Thanks for coming on. Thank you. All right, so you are a superhero, Amazon hero, superhero, what do they call it? Amazon community hero. Community hero. Specifically of the first serverless, I'm the serverless community hero. The serverless community hero, and of course that was their theme all last year, be a superhero, burner on stage. What's new? What's new since last year? With, well, a lot of things are new since last year. IOP, IOP, are you a company? Yeah, IOPype, maybe we're not new, we've been around, we've made a lot of changes. We've been working not just on the infrastructure observability, but also helping developers understand how the application is being used by the users. So if you build an application, let's say, filters data, you can understand how many records were filtered or how many were processed. You can understand really more insights into things, like for instance, if you have an Alexa scale, you can automatically, without any changes to the app, besides adding IOPype, filter down the conversations to users and really get insight into each indication, the whole conversation that occurred for that Alexa scale. And that's just not for Alexa, that's for any serverless app. The key is the data pipe. So the question that we heard earlier, I asked an early Amazon guy, came on and said, hey, what's going on with the adoption of serverless and Lambda? And she said surprisingly, she said, enterprises are adopting more than they thought. Do you see the same thing with Lambda and serverless? What's going on there? What are some of the trends that you see happening? Yeah, I mean, yes, there's lots of adoption. And I think it's, what I'm actually seeing is a lot of the smaller companies, really, like individual developers that a year ago, were saying, oh, serverless is a fad, serverless is silly, whatever. Those people are now turning around a lot. And some of those people even joined Amazon, right? They were like, yeah, they were like skeptics and now they're like, oh, we're going to work for Amazon. But for those large enterprises, Amazon solved a lot of challenges that they had in the early days, things like PCI compliance, the HIPAA compliance. Like these are problems that Amazon's been solving in the last year or two. And now enterprises are able to adopt. There are still challenges for some of those larger enterprises, especially around things like the development model. It's a bigger paradigm shift to go to serverless than it was to go to Docker. Because with Docker, you already moved things to VMs. You just have to shrink that image and fit it into containers. And now with serverless, it's easier, but different, right? So I would say it's easier that containers were, but also more different than they were. So mindset's shifting, easier. So you kind of got to take a leap of faith if you're from the old VM world. You're like, okay, I've got to kick the tires. I got to go play with it. Is that some of the adoption challenges? Is it, I mean, because that seems to be the issue. It's like, once you believe it's like, okay, once you take it with Amazon, you're like, okay, that seems to be the trend. Do you see it that way? A little bit. I think that Amazon's improving their messaging. They were really pushing for, let's go into an enterprise and tell them, you know, rewrite your entire application or your whole business around serverless, right? Go all in on serverless. And you know, I think they've gotten better with saying- Don't let the door hit you on the way out. I think they've gotten better at telling the story of, it's okay to put your feet in the water. You run it for a little project, do some operational things. You don't have to build large, complex applications with it immediately, right? You're going to get there. I think that there was definitely a lot of push to telling corporations and our large enterprises to go all in, which is not realistic, right? It's a transformation. And that transformation takes time. So when you say different, that needs to be more of like, makes sense now because if it's different, you can't hard sell it. You got to get people to ingratiate into it, feel comfortable because it's going to be the fear of will the parachute open if I jump out of the plane? And then once it works, they kind of get it. I guess the question I would have is, if I'm a developer and I'm like, okay, I function as a service, you got to land all this cool stuff with microservices right around the corner. You get containers, you see service meshes and you see Kubernetes. Certainly the cloud native community has been going crazy over Kubernetes. But then the eye on service meshes, what's your take on that? How does that fit into serverless trend? Is it a FogUnl? Does it connect? Is it a migration path? So I look at serverless applications or stateless applications that run on stateless architecture. You cannot, serverless is a stateless architecture. That requires it to be a stateless application. If you cannot describe your application as stateless, then it's not going to be a good fit for serverless today. So if for some reason you can't define your application as stateless, you're going to have to run on something with Kubernetes. Databases are really straightforward. Example of something that's very, very staple. It's very hard to build a database that is stateless. So databases are a place where serverless is not going to be a good fit. And those workloads of staple applications are going to continue running for now on things like Kubernetes. If your application is stateless, you're like, you know, we can just kill these servers, you know, shut them down, and we can do chaos engineering, and we're all stateless, then serverless is fantastic, right? Cause you're building a stateless application. It makes sense to run it on stateless architecture. Makes sense. What's your biggest observation the past year? What's your, when you look at the landscape, obviously you got other clouds in there. You got Google and Amazon, both consumer companies that build clouds for their scale. And then you got Microsoft, they're not a consumer company. Oracle, IBM, kind of business to business companies. Is there a trend? I mean, we've been kind of poke at this. Stu and I and the team have been like, okay, is there a pattern? Is there going to be two sides of the street? Because I think Google's got a great cloud, a great tech, TensorFlow, right? Not as robust as, say, Amazon, and not as deep and further along, but Azure is essentially a bolt on. And so it's Oracle, IBM's IBM, those are business companies. Is there a pattern? Do you have any insight into that? What's your perspective on the makeup of the cloud providers in general? I mean, is consumer a good thing? I think it is. But then there's other things don't have a consumer background. See the pattern? What do you think? Yeah, I don't know. I guess I haven't thought too much about this. I mean, I thought about it a little bit more recently because I know some people like, I mean, this morning there was a protester that came to the keynote. There's not everybody loves Amazon. Not everybody loves Google. Not everybody likes Microsoft, right? There were a lot of the tech community just like Microsoft for a long time. A lot of the non-tech community dislikes Amazon's practices around just like their factories and so forth, right? Like you have a big corporation like this, you're going to have people that don't like you. And I'm sure that bleeds over a little bit into the cloud. Probably not so much at the enterprises, but I mean, like the fact that these are large enterprises with multiple business segments and subsidiaries and whatever, that all affects the end product. I was talking with a person at Google, it was a great interview I had with her and she was talking about like, okay, they did a study around what makes developers happy. And interestingly enough, wait time was critical between commits. They don't want to wait more than a few minutes before they move on to something else. And the other thing that was interesting was they want to work in an environment that's going to be safe, but they really wanted to have the collaboration. And so the question is, is that, what makes developers happy these days? Is it the time? Is it the collaboration? What are some of the things that you see are making this new modern developer happy these days? What do you see out there in the marketplace? So, I mean, developers generally don't want to be operating things. You know, there's this whole idea of like DevRel, or sorry, DevApps and how DevApps is a culture. It's a way of development, it's a way of approaching your thinking. And I think that really the thing that they were selling there was, developers don't want to operate, right? Or operators don't want to deal. I mean, DevRel was in a way kind of a pushback from operators saying, please, we're going to give you the tooling so you can do this yourself. But then there were a lot of developers that pushed back and said, but I don't want to do operations. I don't understand it. And I think things like Serverless is really helpful for those developers who don't have to do operations anymore. They like the cloud is now handling the upside of DevApps. And it's just depth again to a large degree. These things have kind of like operations are still there. It's just much different. And it's much more about application operations instead of infrastructure operations. What's next for you guys in the company? What are you guys working on now? Any big news you guys hiring? Give me an update on the company that you've found. Really, we've just been kicking ass. Can I say that word? Of course, yeah. Damn straight, you can't. But we've been making sales, improving a product. We're about to release what we call auto tracing, which is going to give users, developers insight into API calls and their responses and their timings and so forth. So you get more of a deeper tracing of an invocation. We released the GoLang agent, Java. I mean, since last year when I spoke to you at least at this conference, we made a lot like, we support it four different languages, three different languages since then. Cause I think back then we're just node. Now it's Python, Java, GoLang. We have profiling since then for all of the languages we support. Wow. You guys been busy? We've been very busy. So we've just been building and building and building and selling. Awesome. How many people in the company now? Still lean and mean? Yeah, just shy of a dozen, I think. Cool. Yeah. Awesome. Well, congratulations, Erica. Great to see you. Thanks for coming on theCUBE. Thank you. Serverless Hero here at AWS on theCUBE. Bringing it down. A stateful and stateless application whenever we want to be in theCUBE. Of course, we're bringing it without any service because we are in Amazon. Thanks for sharing the comment there. I'm John Furrier with Jeff Frick. Stay with us for more coverage here for all day here in Manhattan for CUBE coverage of AWS in New York City. We'll be right back.