 Live from San Francisco, it's theCUBE. Covering AWS Summit 2017, brought to you by Amazon Web Services. Hi, welcome back to theCUBE. We are live in San Francisco at the AWS Summit at Moscone Center, really excited to be here. Tremendous amount of buzz going on. I'm Lisa Martin with my co-host, George Gilbert. We're very excited to have Lowell Anderson, product marketing guru at AWS. Welcome back, CUBE alumni. It's great to be here, Lisa, thank you. Great to have you here as well. The keynote this morning was so energetic with Werner, and next door is going to be on the program in a little bit. Over a thousand product launches last year. Not only are there superpowers now that AWS, I like that you don't have a T-shirt, but maybe next time. But I think the word that I heard most today so far is customer, and I think that it's such a, and as AWS really talks about it's a really differentiating way of thinking, of doing business. I'd love to understand what the products that were announced today can walk us through some of the key highlights there. Customer logos were everywhere, so talk to us about how customers are influencing the development of the new services and products coming from AWS. Yeah, well, for us, customers are always core to what drives our innovation. It's how we start. We start with what our customers want, and we work backwards from that to try to deliver a lot of the new features and services that we talked about today, and Werner covered a huge breadth of things, but they really kind of fall into maybe four or five categories. He started talking about directly for developers, talking about what we're doing with a product called CodeStar, which is designed to really help developers build and deploy software applications in the cloud. He also then went and talked about our new marketplace SaaS contracts capability, which makes it super easy for customers to sign up and purchase SaaS applications using multi-year contracts on AWS, but it also makes it easier for ISVs to make their offerings available for our customers. So, again, really trying to make that easy for customers. We talked a lot about what we're doing in artificial intelligence, with the general availability of Amazon Lex today, and then some really entertaining video with Polly, I think where we saw that avatar kind of speaking and the new whispering capability, so adding a lot more value to our suite of artificial intelligence services. Some exciting stuff in analytics where we talked about Redshift Spectrum, which is a new search capability on Amazon Redshift that allows customers to search not just the data in the Redshift database, but also search all the unstructured data they have in S3. And then some really exciting announcements are on the database space with DynamoDB DAX, which is an accelerator for DynamoDB. And we also talked about the availability of a new version of Aurora for Postgres. So, a lot of new capabilities, both in databases, big data, analytics, machine learning and artificial intelligence, and our offerings for SaaS contracts as well. And that was all before lunch. Yeah, a lot of stuff, a lot of stuff. Just a little following up on sort of in order, perhaps on let's say the comments on AI and the announcements that were made there. Microsoft, Google, Amazon all have gone beyond sort of frameworks and tools to sort of fully trained services that a normal developer can get their hands around. But in the areas of conversational user interface, natural language understanding, image recognition, why do you think that those three vendors, or the three vendors have been able to make such progress in those areas to bring them to make that capability accessible? And there's so many other areas where we're still down in the weeds. Yeah, well, I think there's, we sort of see it in sort of focusing in maybe three different areas that are really targeted at what our customers are asking for. We have some very sophisticated customers who really want to build their own deep learning and machine learning applications. And they want services like MXNet, which is a highly scalable deep learning framework that they can do and build these deep learning models. So there's a very sophisticated, kind of targeted customer who wants that. But we also have customers that want to build and train and create prediction algorithms. And they use Amazon Machine Learning, which is a managed service which allows them to look at their, maybe their past transactional data and build prediction models from it. And then the third piece is kind of what you mentioned, which is services that are really designed for the average developer so they can really easily add capabilities like chat box and speech and visual recognition to their applications with a simple API interface. And I think what you touched on is, why did we focus here? Well, I think, as Andy also talked about today, it's really early days in this space. And we're going to see a really, really strong amount of innovation here. And Amazon, which has been doing this for many, many years, thousands of developers focused on this in our retail side, we're really working hard to bring that technology out so that our customers can use it. And Lex, which is based on Alexa, which we're all familiar with from using the Echo, bringing that out and making that type of capability available for average developers to use is a piece of that. So I think you're just going to continue to see that. And over the course of the next year, you're going to see continued new services coming from us on machine learning and artificial intelligence across all those three spectrums. So let me jump to another subject, which is really a hot button for our customers, both on the vendor side and the enterprise side, which is the hybrid cloud. I don't know whether we should call it migration or journey or end point. But let's take a couple of scenarios. Let's say you're a Hadoop customer and you've got Cloud Era on-prem, you're a big bank, you put an instance of it on Amazon and on Azure so that you can move your data around and you're relatively free. Sure. Now the big use case has been like data warehouse offload. So all of a sudden, you have two really great data warehouses that are best in class on Amazon with Redshift and now the significant expansion of it and it smells like. And then you have Teradata, which now can take their on-prem capabilities and put them on the cloud. How does a customer sort of weigh the cost benefit of lowest common denominator versus. Yeah, sure. I think for us and for our customers, it's really, it's not a one size fits all. Every customer approaches this differently and so what our focus has been on is to give them the range of choice. So if you want to use Cloud Era, you can deploy it on EC2 and you can manage that yourself and that's going to work great for you. But if you want a more managed service where maybe you don't want to have to manage the scalability of that Cloud Era deployment, maybe you want to use EMR and deploy your Hadoop applications on EMR, which is manages that scalability for you. And so you make those trade-offs and each of our customers makes those trade-offs for different reasons and in different ways and at different times. And so our focus has always been to really try to give them that flexibility to give them services where they can make the choice themselves about which direction they want to go for their individual applications and maybe mix it up and try different ways of running these types of applications. And so we have a full range of those types from the ability to deploy those directly onto EC2 and manage it themselves all the way to fully manage services that we maintain all the scalability and management and monitoring ourselves. One of the interesting things that Andy Jassy said in his fireside chat just in the last hour or so about HyperCloud was that most enterprises are going to operate in HyperCloud for the next several years. And there are those customers that are going to have to or want to have their own data centers for whatever type of applications. But something also that he brought up in that context and I know you know a lot about this George is the EMR. So when I was looking at the announcement that was made in the last six months or so about VMware vSphere based cloud services, VMware's just sold off their vCloud Air kind of competing product. Wondering with the VMware cloud on Amazon, how does that, what are really kind of the primary drivers there? Is that sort of a way to take those VMware customers eventually towards HyperCloud? Or is that an opportunity to maybe compete with some of the other guys who might have more traction in the legacy application migration space? Yeah, I think for us, it's again, it comes back to our customers saying, you know, some of our workloads that maybe for a long period of time have been deployed on VMware and we've been using VMware ESX for many, many years on premise. And we have these applications that employed for many years there and they're highly integrated. They use specific features of VMware. And maybe we also like using VMware's management tools. We'd like using vCloud to manage all of these different instances of our VMware virtualization platform. But we just want to run it in the cloud because we want that scalability. We, you know, when you deploy that stuff on premise, you're still kind of locked in. Every time you want to expand, you've got to go out and you've got to buy more hardware. You really don't have the agility to expand that business, both as it grows or as it declines. So you're paying for that hardware to power it and run it no matter what. And so they're telling us, we'd like to get some of this up into the cloud, but we don't want necessarily to have to, we've built these apps, we're comfortable with how they're running them, but we want to run them up in the cloud and we want to do it with low risk. And that's what this VMware relationship is about, is letting those enterprises that have spent years building and maintaining and using VMware and their various management tools to do that up in the cloud. That's really what it's about. So let's switch gears to another topic that Andy talked about since all his topics were topical. Edge computing and IIoT. That's another big shift that's coming along and changing the architecture. So we have more computing at the edge again and huge amounts of data. Obviously there's many scenarios, but how do you think customers will basically think through this, or how should they think through how much analytics and capability is at the edge, that issue of should it look like what it's in the cloud or should it be really tight and light and embedded? Yeah, I think we're seeing just an increasing range and also a really interesting mix where you have some very intelligent devices, your laptop and so on that is connected to the cloud and has this pretty significant amount of processing power. And so there can be applications that run on that machine that are very sophisticated. But if we're going to start to expand that universe of edge devices out to simple sensors for pipelines and simple ways to monitor the thermostat in your home and simple ways to measure and monitor and track all sorts of automobiles and so on, that there's going to be a range of different on-premise or edge types of compute that we need to support in the cloud. And so I think what Andy's saying is that we want to build the cloud to be the system that can act as the, has the analytics power to ingest data from these maybe tens of millions of different devices which will have a range of different compute power and support those applications as, on a case by case basis. We've got to kind of wrap things up here. And I know this conversation could continue for many hours. I think what we've heard here today is a tremendous amount of innovation. I made the joke all announced before lunch, but really it was. We're seeing the flexibility, we're seeing the customers really drive the innovation. Also the fact that, you know, AWS starting in the startup space with the developers, that's still a very key target market for you even as things go up to the enterprise. So continued best luck with everything going forward. We're excited to be at re-invent in just what five or six months from now and with many, many more thousands of people and hearing the great things that continue to come from the leader in public cloud. Well thanks for joining us, Lowell. We appreciate it. Next time I want the super power T-shirt. Okay, I'll take you up on that. All right, I'm Lisa Martin from my co-host George Gilbert. Thanks so much for watching, stick around. We are live at the AWS Summit in San Francisco and we will be right back.