(upbeat music) >> We're back at AWS re:Invent 2021. It's the biggest hybrid event of the year. One of the few physical events and we're psyched to be here. My name is Dave Vellante, and I'm really pleased to bring back the host emeritus, Stu Miniman, somebody I worked with side-by-side, Stu, for 10 years in a setting much like this, many like this. So, good to have you back. >> Dave, it's great to be here with theCUBE team, family here and re:Invent, Dave. I mean, this show, I remember back, Dave, going to you after the first re:Invent we talked, we were like, "We got to be there." Dave, remember the first year we came, the second year of re:Invent, this is the 10th year now, little card tables, gaming companies, all this stuff. You had Jerry Chen on yesterday and Jerry was comparing like, this is going to be like the next Microsoft. And we bet heavy on this ecosystem. And yeah, we all think this cloud thing, it might be real. 20,000 people here, it's not the 50 or 75,000 that we had in like 2018, 2019, but this ecosystem, what's happening in the cloud, multiple versions of hybrid going on with the event and the services, but yeah, phenomenal stuff. And yeah, it's so nice to see people. >> That's for sure. It's something that we've talked about a lot over the years is, and you remember the early days of re:Invent and to this day, just very a strong developer affinity that AWS has done a tremendous job of building that up and it's their raison d'etre, it's how they approach the market. But now you've been at Red Hat for a bit, obviously as well, developer affinity, what have you learned? Specifically as it relates to the cloud, Kubernetes, hottest thing going, you don't want to do an OpenShift commercial, but it's there, you're in the middle of that mix. What have you learned generally? >> Well, Dave, to the comment that you made about developers here, it's developers and the enterprise. We used to have a joke and say, enterprise developer is an oxymoron, but that line between developers doing stuff, early as a cloud, it was stealth computing. It's they're often doing this stuff and central IT is not managing it. So how do the pieces come together? How do apps and infrastructure, how do those pieces come together? And it's something that Red Hat has been doing a long time. Think about the Linux developer. They might've not have been the app developers, the people building Linux and everything, but they had a decent close tie to it. I'm on the OpenShift team. What we do is cloud, Dave, and we've got a partnership here with Amazon. We GAed our native cloud service earlier this year. Andy Jassy helped name it. It is the beautifully named Red Hat OpenShift Service on AWS or ROSA. But we've done OpenShift on AWS for more than five years, basically since we were doing Kubernetes, it's been here because of course customers doing cloud, where are they? A lot of them are here in Amazon. So I've been loving talking to a lot of customers, understanding how enterprise adoption is increasing, how we can enable developers and help them move faster. And yeah, I mean the quick plug on OpenShift is our service. We've got an SRE team that is going to manage all of that. A friend of the program, Corey Quinn, says, "Hey, an SRE team like that, because you don't want to manage as an enterprise." You don't want to manage Kubernetes. Yeah, you need to understand some of the pieces, but what is important to your business is the applications, your data and all those things and managing the undifferentiated heavy lifting. That's one of the reasons you went to the cloud. So therefore changing your model as to how you consume services in the cloud. And what are we seeing with Amazon, Dave? They're trying to build more solutions, simplify deployments, and offer more solutions including with their ecosystem. >> So I want to ask you. You said enterprise developer is kind of an oxymoron, and I remember, years ago I used to hang around with a lot of heads of application development and insurance companies and financial services, pharmaceutical, and they didn't wear hoodies, but they didn't wear suits either. And then when I talked to guys like Jeff Clark, for instance. He talks about we're building an abstraction layer across clouds, blah, blah, blah, which by the way, I think it is the right strategy. I'm like, "Okay, I'll drink some of that Kool-Aid." And then when I come here, we talked to Adam Selipsky. John flew out and I was on the chime. He goes, "Yeah, that's not hybrid. No, this is nothing like, it's not AWS, AWS is cloud." So, square that circle for me, 'cause you're in both worlds and certainly your strategy is to connect those words. Is that cloud? >> Yeah, right. I mean, Dave, we spent years talking about like is private cloud really a cloud? And when we started coming to the show, there is only one cloud. It is the public cloud and Amazon is the paragon of, I don't know what it was. >> Dave: Fake clouds, cloud washing. >> So today, Amazon's putting lots of things into your data center and extending the cloud out to that environment. >> So that's cloud. >> That's cloud. >> What do we call that cloud? What about the reverse? >> What's happening at the edge is that cloud is that extension of what we said from Amazon. If you look at not only Outpost, but Wavelengths and Local Zones and everything else like that. >> Let's say, yes, that's cloud. The APIs, primitives, check. >> Dave, I've always thought cloud is an operating model, not a location. And the hybrid definition is not the old, I did an ebook on this, Dave earlier this year. It's not the decade old NIS definition of an application that spans because I don't get up in the morning as an enterprise and say, "Oh, let me look at the table of how much Google is charging me or Microsoft or Amazon," or wake up one morning and move from one cloud to the other. Portability, follow the sun type stuff, does it ever happen? Yes, but it is rare thing. Applications oftentimes get pulled apart. So we've seen if you talk about AI, training the cloud, then transact and do things at the edge. If I'm in an autonomous vehicle or in a geosynchronous satellite, I can't be going back to the cloud to process stuff. So I get what I need and I process there. The same thing hybrid, oftentimes I will do my transactional activity in the public cloud because I've got unlimited compute capability, but I might have my repository of data for many different reasons, governance or security, all these things in my own data center. So parts of an application might live there, but I don't just span to go between the public cloud in my data center or the edge, it's specific architectural decisions as to how we do this. And by the way the developer, they don't want to have to think about location. I mean, my background, servers, storage, virtualization, all that stuff, that was very much an infrastructure up look of things. Developers want to worry about their code and make sure that it works in production. >> Okay, let me test that. If it's in the AWS cloud and I think it's true for the other hyperscale clouds too, they don't have to think about location, but they still have to think about location on-prem, don't they? >> Well, Dave, even in a public cloud, you do need to worry about sometimes it's like, "Okay, do I split it between availability zones? How do I build that? How do I do that?" So there are things that we build on top of it. So we've seen Amazon. >> I think that's fair, data sovereignty, you have to think about okay. >> Absolutely, a lot of those things. >> Okay, but the experience in Germany is going to be the same as it is in DC, is it not? >> More or less? There are some differences we'll see off and Amazon will roll things out over time and what's available, you've got cloud. >> For sure, though that's definitely true. That's a maturity thing, right? You've talked a bit, but ultimately they all sort of catch up. I guess my question would be is the delta between, let's say, Fed adoption and East Coast, is that delta narrower, significantly narrow than what you might see on-prem? >> The services are the same, sometimes for financial or political things, there might be some slight differences, but yes, the cloud experience should be the same everywhere from Amazon. >> Is it from a standpoint of hybrid, on-prem to cloud, across cloud? >> Many of the things when they go outside of the Amazon data centers are limited or a little bit different or you might have latency considerations that you have to consider. >> Now it's a tug of war. >> So it's not totally seamless because, David Foyer would tell us there, "You're not going to fight physics." There are certain things that we need to have and we've changed the way we architect things because it's no longer the bottleneck of the local scuzzy connection that you have there, it is now (indistinct). >> But the point I'm making is that gets into a tug of war of "Our way is better than your way." And the answer is depends in terms of your workload and the use case. >> You've looked at some of these new databases that span globes and do things of the like. >> Another question, I don't know if you saw the Goldman Sachs deal this morning, Goldman Sachs is basically turning its business into a SaaS and pointing it to their hedge funds and allowing people to access their data, their tools, their software that they built for their own purposes. And now they're outselling it. Similar to what NASDAQ has done. I can't imagine doing that without containers. >> Yeah, so interesting point, I think. At least six years ago now, Amazon launched serverless and serverless was going to take over the world. I dug into the space for a couple of years. And you had the serverless with camp and you had the container camp. Last year at re:Invent, I really felt a shift from Amazon's positioning that many of the abstraction layers and the tools that help you support those environments will now span between Lambda and containers. The container world has been adding serverless functionality. So Amazon does Fargate. The open-source community uses something called Knative, and just breaking this week. Knative was a project that Google started and it looks like that is going to move over to the CNCF. So be part of the whole Kubernetes ecosystem and everything like that. Oracle, VMware, IBM, Red Hat, all heavily involved in Knative, and we're all excited to see that go into the CNCF. So the reason I say that, I've seen from Amazon, I actually, John and I, when we interviewed Andy Jassy back in 2017, I asked him a follow-up question because he said if he was to build AWS in 2017, "I would start with everything underneath it serverless." I would wonder if following up with Adam or Andy today, I'd said, "Would it be all serverless or would containers be a piece of it?" Because sometimes underneath it doesn't matter or sometimes it can be containers and serverless. It's a single unit in Amazon and when they position things, it's now that spectrum of unit, everything from the serverless through the containers, through... James Hamilton wrote a blog post today about running Xen-on-Nitro and they have a migration service for a mainframe. So what do we know? That one of the only things about IT is almost nothing ever goes away. I mean, it sounded like Amazon declared coming soon the end of life of mainframe. My friends over at IBM might not be quite ready to call that era over but we shall see. All these things take time. Everything in IT is additive. I'm happy to see. It is very much usually an end world when I look at the container and Kubernetes space. That is something that you can have a broad spectrum of applications. So some of my more monolithic applications can move over, my cool new data, AI things, I can build on it, microservices in between. And so, it's a broad platform that spans the cloud, the edge, the data center. So that cloud operating model is easier to have consistency all the places that I go. >> Mainframe is in the cloud. Well, we'll see. Big banks by the next site unseen. So I think Amazon will be able to eat away at the edges of that, but I don't think there's going to be a major migration. They claim it. Their big thing is that you can't get COBOL programmers. So I'm like, "Yeah, call DXC, you'll get plenty." Let's talk about something more interesting. (Stu laughs softly) So the last 10 years was a lot of, a lot about IT transformation and there was a lot more room to grow there. I mean, the four big hyperscalers are going to do 120 billion this year. They're growing at 35%. Maybe it's not a trillion, but there's a $500 billion market that they're going after, maybe more. It looks like there's a real move. You saw that with NASDAQ, the Goldman deal, to really drive into business, deeper business integration in addition to IT transformation. So how do you see the next decade of cloud? What should we be watching? >> So, one of the interesting trends, I mean, Dave, for years we covered big data and big data felt very horizontal in it's approach thing. Hadoop take over the world. When I look at AI solutions, when I look at the edge computing technologies that happen, they're very vertically driven. So, our early customers in edge adoption tend to be like telco with the 5G rollout manufacturing in some of their environments. AI, every single industry has a whole set of use cases that they're using that go very deep. So I think cloud computing goes from, we talked about infrastructure as a service to it needs to be more, it is solution, some of these pieces go together. When Adam got up on stage and talked about how many instance types they have on Amazon, Dave, it's got to be 2X or 4X more different instant types than if I went to go to HPE or Dell and buy a physical server for my environment. So we need to have areas and guidance and blueprints and heck, use some of that ML and AI to help drive people to the right solutions because we definitely have the paradox of choice today. So I think you will find some gravity moving towards some of these environments. Gravatar has been really interesting to watch. Obviously that Annapurna acquisition should be down as one of the biggest ones in the cloud era. >> No lack of optionality to your point. So I guess to the point of deeper business integration, that's the big question, will Amazon provide more solution abstractions? They certainly do with Connect. We didn't hear a ton of that this show. >> Interestingly. (Dave speaking indistinctly) So the article that you and John Furrier wrote after meeting with Adam, the thing that caught my eye is discussion of community and ecosystems. And one of the things coming after, some, big communities out there like, you and I lived through the VMware ecosystem in that very tight community. There are forming little areas of community here in this group, but it's not a single cloud community. There are those focus areas that they have. And I do love to see, I mean, obviously working for Red Hat, talking about the ecosystem support. I was very happy to hear Adam mention Red Hat in the keynote as one of the key hybrid partners there. So, for Amazon to get from the 60 million, the 60 billion to the trillion dollar mark down the road, it's going to take a village and we're happy to be a part of it. >> Hey, great to have you back, enjoy the rest of the show. This is, let's see, day three, we're wrapping up. We're here again tomorrow so check it out. Special thanks to obviously AWS is our anchor sponsor and of course, AMD for sponsoring the editorial segments of our event. You're watching theCUBE, the leader in tech coverage. See you tomorrow. (bright upbeat music)

>> Announcer: Live from Manhattan, it's theCUBE covering AWS Summit New York City 2017. Brought to you by Amazon Web Services. >> Welcome back live to Midtown Manhattan, along with Stu Miniman, I am John Walls. We're here on theCUBE and we're wrapping up our coverage here at AWS Summit. Again, kind of tough to get a feeling for just how many folks were here. But some were in that seven, eight, $9,000 range and most of them are still here I think, out on the show floor here behind us. Good keynotes this morning. Good programming throughout the day as well and then really good buzz here on the show floor. So, good day I think, for AWS Stu, and we've talked about it, it is kind of remarkable to see the number of people who turned out for a regional show. >> Yeah John, you know I've been to some shows in the Javits Center where people wander in, they get some swag, they look for a free beer and a t-shirt and then that's kind of their... These people are, you know, kind of diggin' in. I know there's a bunch of sessions been going on. The pavilion here has had all these little breakout sessions. There was one on, you know, VMware and VMware and AWS and it was, you know, not only the seats, which usually it was like oh come on in, you know, come get a prize and things like that. >> John: Right, right. >> There was five rows of people standing pressed in and asking questions like, "How do I set up "the networking on this, how does this work?" Things like this, so it's like a mini AWS re:Invent, so their big show, one we've done theCUBE at a number of years, I've been there a number of years. I commented on our intro that this is larger than the first Amazon re:Invent that I went to like four years ago. >> How about that, in that short of period of time? >> Yeah and that's one of the things about Amazon and public Cloud in general and all of these technologies, the growth and the speed of change is just amazing. It used to be we talked from a software standpoint, it was like okay, I'm tied to that Intel release of every 18 months that I'm going to click out, then it was like okay, we kind of go to a yearly cycle. Now it was more like well not only is a lot of software released, you know, continuous integration and continuous deployment CICD, which sometimes it's every six weeks, sometimes it's daily, but Amazon's releasing new features every day. We talked in the intro, oh there were three major releases and we had the guy I'm talking about, the machine learning stuff and he's like oh you mean the three announcements that we had in machine learning? And we're like oh, we only heard about one of those. Wait, you had a couple others underneath there? Oh, let's talk about the F1 compute instance and the FPGAs. There's always so much in Amazon and when you go into any environment in the little boxes that they put in there and you start peeling the onion, it's impressive. >> It is. >> And there's just depth and customers are interested in it and people are using it. You know, I was used to so much in my career where something gets announced and a year later it's like hello, is anybody using this? As opposed to at this show, a bunch of the announcements, I already talked to a bunch of people that have been in private beta, they've been testing this out, they're excited about it and because it's just so easy to get on all of these new features. >> Right, and I mean, we've seen it here, we've heard from many people here from a lot of different walks of life. You mentioned some of the past shows, AWS Public Sector. I was at that not too long ago in Washington, D.C. and you see a company that has its units very focused and very driven and doing very well and the right relationships. Buzzword, serverless, right? We heard it a lot today. Serverless applications, serverless computing. From more than one source, we heard it from several folks and so obviously this is not just a popular piece of nomenclature for the day, this is a trend, a theme that's going to be evolving and maturing over the next year or two. >> Yeah I mean everybody for the last couple years they've kind of been looking at it with their head sideways. I'm not sure that I understand it. We talked to two companies today, it was IOPipe and A Cloud Guru that their company, their IT infrastructure was all built on serverless, and they both got funding recently, so this isn't just oh yeah, some developer does some cool stuff on the side, microservices, buzz buzz, things like that. We talked to FICO is using serverless for their admin functions, certain areas they're not ready to roll it out across the board, governance compliance, things like that, I need to understand it. It is still very early, but that being said, there's a lot of usage in it. Last year it was oh, if you want to develop for the Alexa platform, the Amazon Echo type thing, that uses serverless, so we're seeing lots and lots of cases. That really is a new way of architecting the way to roll out really microservices driven applications and when we talk about the big challenge of our time, it's distributed architectures and how do I have new applications? We talked to a number of companies moving from the old way of doing my application to building new application, that's the long hole in the 10. This is not something that happens overnight, but I can start playing with it in a much smaller form factor and do it for pennies not years and millions of dollars so there is really serverless has really in many ways eclipsed kind of the container's discussion for the hot buzz in the industry. Kubernetes fits into that whole picture, but not just serverless in general, but AWS Lambda is the leader of the pack out there and you know, yet another reason why Amazon just going strong, their revenue still doing well, keeps adding to what they're doing and you don't hear many people griping when you walk around the show floor as to what they wish they had. It's a very positive experience. >> And you hear criticisms saying, "They only had 42% growth year to year." It's not what it used to be. But 42 as you know, most people would gladly be in that position. What about your thoughts about the maturation of the Cloud? You mentioned transformative and things are evolving and growing, where do you put it now? Is this second phase, next phase, late phase? Where are we in terms of what's happening and what AWS is making happen? >> So a couple years ago we know that Cloud is here to stay. There's still the joke a friend of friend of mine in the keynote. 20,000 people registered for this event and it was like well, I guess this Cloud thing might have legs, so we are still early in the overall wave of this. I've been in a number of conferences this year that we've done theCUBE on. You talk about the infrastructure companies and companies that have built on virtualization. They said, "We went through a decade "of tremendous growth with virtualization." Virtualization is still very important. Amazon builds their infrastructure not on VMWare, but they leverage virtualization technologies, but the next 10 years will be this huge wave of really that going up the uptake of the S curve so we're past really the classic crossing the chasm. We're in the early majority going to mid majority of people using it and there's just no shortage of new use cases that people can use it for. We've talked to lots of companies that start up and say, "I'm just leveraging Cloud because it's easy." THere's VCs that look at that as how to get involved and as I've just mentioned before, there's companies now that are building themselves on serverless so this is even kind of the next piece that follows these waves we are early in Cloud if you look at kind of overall ham of IT, public Cloud is still a very small piece. At Wikiban we've been talking for the last I think two years about what we really the multi Cloud environment. There's true private Cloud and there's public Cloud and how do I get that operational model that I can scale, I can build really a distributed architecture? I shift more to an operational expense rather than a capital expense, so it's flexibility, it's agility, it's speed, and it's very interesting, exciting times. There's no more exciting time to be in tech than today, maybe tomorrow, because we know the only thing constant is that the pace of change keeps increasing. >> It does increase and two big drivers of that, we heard again today, artificial intelligence, machine learning. How would you rate or how would you characterize the impotence that they're providing in terms of pushing the envelope? >> Absolutely there was some good announcements today, I don't know that there's any today that you'd say, "I'm going to look back five years from now and be like, 'Wow, I was in New York City when that was announced.'" >> John: Right, but just in general? >> But in general, let me say one of the things that I didn't hear today, I was was little bit disappointed, I mentioned it in the open, we talked to a couple of the partners here, you know the Kubernetes option. Adrian Kovrov got up on stage. He had written a blog post there was an announcement last week, no mention of where Kubernetes is going to fit in here. Definitely they're committed to it, they're making developments, but maybe something will come out in beta soon. I would expect by the time we get to the re:Invent show in November that we will have more clarity here. I was hoping to hear that more and that was something that didn't come out of Amazon, but they're embracing it. Customers are asking for it, developers, there's a ground swell on that, so they're involved with it. Lambda and serverless absolutely. Amazon is at the vanguard, they're pushing things forward. Machine learning and IoT, Amazon is at the table. It is still very early, they're driving a lot of things forward. Yeah, you know, you get enough, it's like come on, there's no BitCoin discussed today, why is that? So some of the other vendors there, but Amazon is in all the appropriate conversations. There's not any wide gaps that you'd say customers like hate these. Amazon's not in this base and I expect them to and therefore I'm going to choose another platform provider. That being said, it's not a winner-take-all, it is a multi Cloud world, most of these environments, we talked about even if I do serverless, if I architect them a certain way I can move them and make changes, Kubernetes the same way. So Amazon, one of the things that they pride themselves on is they need to keep proving to their customers every month that they are the ones that they fuse on because otherwise it is relatively easy to make a change, but they're the big dog, they got the leadership position, and it's always impressive to watch them. >> It is and you speak of impressive. re:Invent, is just what, two and a half months away, three months away, we'll be out there as well. Huge show, probably one of the largest shows by far that we attend and looking forward to that and seeing you down the road. Always a pleasure to be with you. >> Thanks so much. >> And great job as always. Stu Miniman does an outstanding job providing analysis for Wikiban, so on behalf of Stu and all the crew here at theCUBE, we thank you for joining us here at the AWS Summit in Midtown. We've been live at the Javits Center. Have a good week and we'll see you down the road here on theCUBE. (light electronic music)