 Hi, I'm Peter Burris and welcome once again to Action Item. Every week Wikibon gathers together the research team to discuss seminal issues that are facing the IT industry and this week is no different. In the next couple of weeks, somewhere near 100,000 people are going to be heading to Las Vegas for the Amazon or AWS re-invent show from all over the world. And this week what we want to do is we want to provide a preview of what we think folks are going to be talking about. And I'm joined here in our lovely Palo Alto studio, the Cube studio, by Rob Hof, who's the editor-in-chief of Silicon Angle, David Fleury, who's an analyst at Wikibon, George Gilbert, who's an analyst at Wikibon, and John Furrier, who's a Cube host and co-CEO. On the phone we have Neil Raiden, an analyst at Wikibon, and also Dave Vellante, who's co-CEO with John Furrier, an analyst at Wikibon as well. So guys, let's jump right into it. David Fleury, I want to hit you first. AWS has done a masterful job of making the whole concept of infrastructure as a service real. Nobody should downplay how hard that was and how amazing their success has been. But they're moving beyond infrastructure as a service. What are the expected, or what do we expect for how far up Amazon's likely to go up the stack this year at re-invent? Well, I can say what I'm hoping for. I agree with your premise that they have to go beyond IAS. The overall market for cloud is much bigger than just IAS with SAS and other clouds as well, both on-premise and off-premise. So I would start with what enterprise CIOs are wanting, and they are wanting to see a multi-cloud strategy, both on-premise and multiple clouds, SAS clouds, other clouds. So I'm looking for AWS to provide additional services to make that easier, in particular services for private clouds, for enterprises. I'm looking for distributed capabilities, particularly in the storage area, so they can link different clouds together. I want to see edge data management capabilities. I'd love to see that because the edge itself, especially the low-latency stuff, the real-time stuff, that needs specialist services, and I'd like to see them integrate that much better than just Snowball. I want to see more details about AI. I'd love to see what they're doing in that. There's tremendous potential for AI and operational to improve security, to improve availability, recovery. That is an area where I think they could be a leader of the IT industry. Let me stop you there. George, I want to turn to you. So AWS in AI. How do we anticipate that's going to play out at re-invent this year? I can see three things in decreasing order of likelihood. The first one is they have to do a better job of tooling, both for sort of developers who want to dabble in or get their arms around AI, but who aren't real data scientists, and then also hardcore tools for data scientists that have been well-served by recently Microsoft and IBM, among others. So this is this Iron Man initiative that we've heard about. For the hardcore tools, something from Domino Data Labs, it looks like they're going to partner with them. It's like a data science workbench. So for the collaborative data preparation, modeling, deployment, that whole life cycle. And then for the developer-ready tooling, I expect to see that they'll be working with a company called DataRobot, which has a really nifty tool where you put in a whole bunch of training data, and it trains, could be a couple dozen models that it thinks might fit, and it'll show you the best fits. It'll show you the features in the models that are most impactful. In other words, it provides a lot of transparency. So it's kind of like models for models? Yes, and it provides transparency. Now that's the highest likelihood, and we have names on who we think the likely suspects are. The next step down, I would put applying machine learning to application performance management and IT operations. So that's the whole AI for ITOM that David Fleur just mentioned. Yeah. Now presumably this is going to have to extend beyond just AI for Amazon or AWS-related ITOM. Are expectations that we're going to see a greater distribution of or Amazon take more of a leadership in establishing a framework that cuts across multi-cloud? Have I got that right, David Fleur? Absolutely. And that's an opportunity for them to provide the basics on their own platform. That's obviously the starting point. They'll have the best instrumentation for all of the components they have there, but they will need to integrate that in with their own databases, with other people's databases. The more that they can link all the units together and get real instrumentation from an application point of view of the whole of the infrastructure, the more value AI can contribute. John Fleur, the whole concept of the last few years of AWS is that all roads eventually end up at AWS. However, there's been a real challenge associated with getting this migration momentum to really start to mature. Now we saw some interesting moves that they made with VMware over the last couple of years, and it's been quite successful. And some would argue it might even have given another round of life to VMware. Are there some things we expect to see AWS do this time that are going to really energize the ecosystem to start bringing more customers higher up the stack to AWS? Yeah, I think I look at it quickly as VMware was a groundbreaking event for both companies. VMware and AWS, we talked about that research event we had with them. The issue that's happening is that AWS has continued to have a run on the marketplace. They've been the leader in cloud every year. It's been a slew of announcements. This year is no different. They're going to have more and more announcements. In fact, they had to release some announcements early before the show because they have, again, more and more announcements. So they have the under the hood stuff going on that David Floyer and George were pointing out. So the classic build strategy is to continue to be competitive by having more services layered on top of each other, upgrading those services. That's a competitive strategy for Amazon under the hood. On the business side, you're seeing more competition this year than ever before. Amazon now is highly contested, certainly in the marketplace with competitors. You're seeing FUD, fear and certainty endowed from other people, how they're bundling. But it's clear. The cloud visibility is clear to customers. The numbers are coming in multiple years of financial performance. But now the ecosystem play is really the interesting one. I think the VMware move is going to be a tell side for other companies that haven't won that top three position. Example. I will say SAP. Oh, really? You think SAP is going to have a major play this year? We might see some more stuff about AWS and SAP. I'm hearing rumblings that SAP is going to be expanding their relationship. I don't have the facts yet on the ground. But from what I'm sensing, this is consistent with what they've been doing. We've seen them at Google Cloud Platform. We've talked to them specifically about how they're dealing with cloud and that their strategy is clear. They want to be on Azure, Google and Amazon. They want to provide that database functionality and their client base in from HANA and roll that in. So it's clear that SAP wants to be multi-cloud. We've seen Oracle over the past couple of years where our research has suggested, I would say, that there's been kind of two broad strategies. The application-oriented strategy that goes down to IAAS aggressively. That'd be Oracle and Microsoft. And then the IAAS strategy that's trying to move up through an ecosystem play, which is more where AWS, David Floyer and I have been writing a lot of that research. So it sounds like AWS is really going to start doubling down in the ecosystem and making strategic bets on software providers that can bring those large enterprise-installed bases with them. And the thing that you pointed out is migration. That's a huge issue. Now, you can get technical and say, what does that mean? But Andy Jassy has been clear and the whole Amazon Web Services team has been clear from day one. They're customer-centric. They listen to the customers. So if they're doing more migration this year, and we'll see, I think they will be. I think that's a good tell sign and good prediction. That means the customers want to use Amazon more. And VMware was the same way. Their customers say, hey, we're ops guys. We want to have a cloud strategy. It was such a great move for VMware. I think that's going to lift the fog, if you will, pun intended, between what cloud computing is and other alternatives. And I think companies are going to be clear that I can partner with Amazon Web Services and still run my business in a way that's going to help customers. So I think that's the number one thing that I'm looking for, is what is the customers looking for in multi-cloud or if it's serverless or other things? Well, or, yeah, I agree. Let me ask one this by you guys. It sounds though multi-cloud increasingly is going to be associated with an application set. So, for example, it's very difficult to migrate a database manager from one place to another as a snowflake. The cost to the customer is extremely high. The cost to the migration team is extremely high, a lot of risk. But if you can get an application provider to step up and start migrating elements of the database interface, then you dramatically reduce the overall cost of what that migration might look like. Have I got that right, David Fleur? Yeah, absolutely. And I think that's what AWS will, I'm expecting them to focus on, is more integration with more SaaS vendors, making it a better place. Or just software vendors. Software vendors. Well, SaaS vendors in particular, but software vendors in particular. Well, SAP is not a SaaS player, right? Well, they are. They are a little bit. But most of their installations are still SAP on Oracle. And moving them more over the last is going to require a significant amount of SAP help. And one of the things I would love to see them have is a proper tier one database as a service. That's something that's hugely missing at the moment. And using HANA, for example, on SAP, it's a tier one database in a particular area. But that would be a good move and help a lot of enterprises to move stuff into AWS. Is that going to be sufficient, though, given how good Oracle is? No. I mean, they need something, general purpose, which can compete with Oracle, or come to some agreement with Oracle, who knows what's going to happen in the future. Yeah, I don't know. We're all kind of here going here. It will be interesting to see. But at the end of the day, Oracle has an incentive also to render more of what it has as a service at some level. And it's going to be very difficult to say, we're going to render this as a service to a customer, but Amazon can't play, or AWS can't play. That's going to be a real challenge. The Oracle thing is interesting, and I bring this up because Oracle has been struggling as a company with cloud-native messaging. In other words, they're putting out, they have a lot of open source, we know what they have for tooling, but they own IT. I mean, you talk about Oracle, they get the databases, David pointed out tier one, but they know the IT guys. They've been doing business in IT for years as a legacy vendor. Now they're transforming, and they're trying hard to be the cloud-native path, and they're not making it. They're not getting the credit. And I don't know if that's a cultural issue with Oracle, but Amazon has that positioning from a developer cloud DNA now winning real enterprise deals. So the question that I'm looking for is, can Amazon continue to knock down these enterprise deals in lieu of these incumbent or legacy players in IT? So if IT continues to transform more towards cloud-native Docker containers or containers in Kubernetes, these kinds of microservices, I would give the advantage to Amazon over Oracle, even though Oracle has the database, because ultimately the developers are driving the behavior. So the question here was, would you disagree with that? The trouble though is the cost of migrating the applications and the data, that is huge. The systems of record are there for a reason. So there are two fundamental strategies for Oracle. If they can get their developers to add the AI, add the systems of intelligence, make them systems of intelligence, then they can win in that strategy. Or the alternative is that they move it to AWS and do that movement in AWS. That's a much more risky strategy. Right, but I think our kind of concluding point here is that ultimately if AWS can get big application players to participate and assist and invest in and move customers along with some of these big application migrations, it's good for AWS. And to your point, John, it's probably good for the customers too. I don't think it's mutually exclusive. David makes a point about migrating for Oracle. I don't see a lot of migration coming off of Oracle. I look at overall database growth is the issue. So Oracle will have that position, but it's kind of like when we argued about the internet growth back in 1997. Just internet users growing was so great, the rising tide floated. So I believe that the database growth is going to happen so fast that Amazon is not necessarily targeting Oracle's market share. They're going after the overall database market, which might be a smaller tier two kind of configuration or new architectures that are developing. So I think it's interesting dynamic and Oracle certainly can play there and lock in the database. Here's what I would say. I would say that they're going after the new workload world and a lot of that new work is going to involve database as it always has. I don't think there's anything. The notion that we have solved or that database is 90% penetrated for the applications that are going to be dominant matter in 2025 is ridiculous. There's a lot of new database that's going to be solved. I think you're absolutely right. Rob Wolf, what's the general scubble butt that you're hearing? You know, as editor-in-chief of Silicon Angle, what is the journalist world buzzing about for a reinvent this year? Well, I guess, you know, my question is because of the challenges they're facing, like we just talked about with the migrating difficulty in migrating some of these applications, we also see, you know, very fast-growing rivals like Google, you know, still small, but, you know, growing fast. And then there's China. That's a big one where, you know, is there a natural limit there that they're going to have? So, I mean, you put these things together and I guess, you know, we see Amazon Web Services still growing at 42% a year or whatever. It's great. But is it going to start to go down because of all these challenges? Because some of the constraints that may start to assert themselves. Exactly, exactly. So that's kind of the journalism world is kind of saying, you know, are there some speed bumps up ahead for AWS? Exactly. We saw one just a couple, well, just this week with China, for example. You know, they sold off $300 million worth of data centers, equipment and such to their partner in China, Beijing Senate. And they say this is a way to, you know, comply with Chinese law. Now we're going to start expanding. But, you know, expanding while you're selling off $300 million worth of equipment, you know, it begs a question. So I'm curious how they're going to get past that. That does raise an interesting question. I think it might go back to some of the AI on ITOM, AI on IT Operations Management, is that do you need control of the physical assets in China to nonetheless sell great service for accessing assets in China? Right. My guess is that if they're successful with AI for ITOM and some of these other initiatives we're talking about, in fact, may be very possible for them to offer a great service in China, but not actually own the physical assets. And that's a, it's an interesting question for some of the Chinese law issues. Dave Vellante, anything you want to jump in on and add to the conversation? For example, if we look at some of the ecosystem and some of the new technologies and some of the new investments being made around new technologies, what are some of your thoughts about some of the new stuff that we might hear about at AWS this year? Well, so a couple things. Just to comment on some of the things you guys were saying about Oracle and migration, to me it comes down to three things, growth, which is clearly there. We've talked about 40% plus growth. Momentum, you know, the flywheel effect that Amazon has been talking about for years and something that really hasn't been discussed much was economics, and this is something that we've talked about a lot. And Amazon is bringing a software like marginal economics model to infrastructure services. And as it potentially slows down its growth, it needs to find new areas and it will expand its time by gobbling up parts of the ecosystem. So, you know, there's so much white space, but partners got to be careful about where they're adding value because ultimately Amazon is going to target those, much in the same way in my view anyway, that Microsoft and Intel have in the past. And so I think you've got to tread very carefully there and watch where Amazon is going and they're going into the big areas of, you know, AI, trying to do more stuff at the edge, and anyway there's automation, they are going to grab that piece of value in the value chain. So one of the things that we've been, we've talked about two main things. We've talked about a lot of investments, a lot of expectations about AI and how AI is going to show up in a variety of different ways at re-invent. And we've talked about how they're likely to make some of these migration issues even that much more tangible than they have been. So by putting some real operational clarity as to how they intend to bring enterprises into AWS. We haven't talked about IoT. Dave just mentioned it. What's happening with the edge? How's the edge going to work? Now historically what we've seen is we've seen a lot of promises that the edge was all going to end up in the cloud from a data standpoint. That's where everything was going to be processed. We started seeing the first indications that that's not necessarily how AWS is going to move last year with Snowball and serverless computing and some of those initiatives. We have anticipated a real honest to goodness true private cloud AWS stack with a partnership. Hasn't happened yet. Dave, what are we looking for this year? Are we going to see that this year or are we going to see more kind of circumvalidating the issue and doing the best that they can? Well, my prediction last year was that they would come out with some sort of data service that you could install on your on-premise machine as a starting point for this communication across a multi-cloud environment. I'm still expecting that, whether it happens this year or early next year, I think they have to. The pressure from enterprises, and they are a customer-driven organization, the pressure from enterprises, is going to mandate that they have some sort of solution on-premise. It's a requirement in many countries, especially in Europe. They're going to have to do that, I think, without doubt. So they can do it in multiple ways. They can do it as they have done with the U.S. government by putting in particular data centers, whole data centers within the U.S. government, or they can do it with small services or they can take the Microsoft approach of having an AWS service on-site as well. I think with pressure from Microsoft, the pressure from Europe in particular is going to make this an essential requirement of the whole strategy. I remember a number of years going back a couple of decades when Dell made big moves because to win the business of a very large manufacturer that had 50,000 workstations, mainly engineers returning over every year to get that business, Dell literally put a distribution point right next to that manufacturer. And we expect to see something similar here, I would presume, when we start talking about this. Yeah, I mean, I want to make a comment on the IOT. First of all, I agree with David said, I don't want to make a prediction, but I'm kind of taking a contrarian view on this and I'm watching a few things at Amazon. Amazon always takes an approach of getting into new markets either with a big idea and small teams to figure it out or building blocks and they listen to the customers. So IOT is interesting because IOT is hard. It's really a fundamental important infrastructure, architecture that's not going away. I mean, it has to be nailed down. It's obvious. Just like blockchain kind of is obvious when you talk about decentralization. It does on those two fronts. But what's interesting to note is, Amazon always becomes their first customer. When their retail business, AWS, was powering retail. With Whole Foods and stuff they're doing on the physical side, it'd be very interesting to see what their IOT strategy is from a technology standpoint with what they're doing internally. We get food delivered to our house now from Amazon Fresh and they got Whole Foods and all the retail. So it'd be interesting to see that. They're buying a lot of real estate and I thought about this as well, John. They're buying a lot of real estate and how much processing can they put in there. And the only limit is that I don't think Whole Foods would qualify as a particularly secure location when we start talking about this. But I think you're absolutely right. It brings the question, how will they roll out IOT because he's like, okay, roll on appliance. That's more of an infrastructure thing. Is that their first move? So the question that I'm looking for is just to kind of read the tea leaves and saying, what is really they're doing? So they have the tech. And it's going to be interesting to see. I mean, it's more of a high level kind of business conversation. But IOT is a really big challenging area. We're hearing that all over the place from CIO. It's like, what's the architecture? What's the playbook? And it's different per company. So it's challenging. Although one of the reasons why it looks different per company is because it is so uncertain as to how it's going to play out. There's not a lot of knowledge diffused. My guess is that in 10 years, we're going to look back and see that there was a lot more commonality and patterns of work that were in IOT that many people expected. So I'll tell you one of the things that I saw last year that particularly impressed me at AWS re-invent was the scale at which the network was being built out. And it raised for me an interesting question. If, in fact, one of the chief challenges at IOT, there's two, there's multiple challenges that every company faces with IOT. One is latency. One is intellectual property control. One is legal ramifications like GDPR, which is one of the reasons why the whole Europe play is going to be so interesting because GDPR is going to have a major impact on a global basis. It's not just Europe. Bandwidth, however, is an area that is not necessarily given. It's partly a function of cost. So what happens if AWS blankets the world with network and customers to get access to at least some degree of edge no longer have to worry about a telco? What happens to the telco business, at least from a data communication standpoint? Anybody want to jump in on that one? Well, yeah. I mean, I've actually talked to a couple of folks like Ericsson and I think AT&T, and they're actually talking about taking their central offices and even the base stations and sort of outfitting them as many data centers. There's Pops. Yeah. But I think we've been hearing now for about 12 months that maybe Edge is going to take over before we actually even finish getting to the cloud. And I think that's about as sort of ill-considered as the notion that PCs were going to put mainframes out of business. And the reason I used that as an analogy, at one point, IPM was going to put all their mainframe-based databases and communication protocol on the PC. That was called OS2 extended edition and it failed spectacularly. For a lot of reasons. But the idea is you have a separation of concerns. Presentation on one side in that case and data management communications on the other. Here in this, in what we're doing here, we're definitely going to have the low-latency inferencing on the Edge. And then the question is, what data goes back up into the cloud for training and retraining and even simulation? And we've already got, having talked to Microsoft's Azure CTO this week, they see it the same way. They see the compute-intensive modeling work and even simulation work done in the cloud and the sort of automated decisioning happening on the Edge. All right, so I'm going to make one point and then I want to hit the action item around here. The one point I want to make is, I have a feeling that over, and I don't know if it's going to happen to reinvent this year, but I have a feeling that over the course of the next six to nine months, there's going to be a major initiative on the part of Amazon to start bringing down the costs of data communications and use their power to start hitting the telecos on a global basis. And what's going to be very, very interesting is whether Amazon starts selling services to its network independent of its other cloud services because that could have global implications for who wins and who loses. Well, that's a good point. I always add color on that. Just anecdotally, from my perspective, you asked the question, I haven't talked to anyone, but knowing the telco business, I think they're going to have that VMware moment because they've been struggling with over the top for so long, the rapid pace of innovation going on that I don't think Amazon is going to go after the telcos. I think it's just an evolutionary steamroller. It's an inevitability. It's an inevitability that the steamroller is coming. Users don't sign long-term data communications deals right now. Why wouldn't you do a deal with Amazon if you're a telco? You get relevance, you have stability, lock in your cash flows, cut your deal, and stay alive. That's an interesting thought. All right, so let's hit the action item around here. So really quickly, as a preface for this, the way we want to do this, guys, is that John Furrier is going to have a couple of hours, one on one with Andy Jassy sometime in the next few days. And so if you were to... Well, tell us a little bit about that first, John. Well, every reinvent... We've been doing reinvent for multiple years, and I think it's our sixth year. We do all the events. We cover it as the media part, as you know. And I'm going to have a one-on-one sit-down every year prior to reinvent to get his view, exclusive interview for two hours, talk about the future. We broke the first Amazon story years ago on the building blocks and how they... Okay, but now they're winning. So it's the time for me to sit down and get his insight and continue to tell the story and document the growth of this amazing success story. So I'm going to ask him specific questions, and I want to love to know what he's thinking. All right, guys. So I want each of you to pretend that you are, so representing your community, what would your community... What's the one question your community would like answered by Andy Jassy? George, let's start with you. So my question would be, are you going to take IT operations management, machine learn, enable it, and then as part of offering a hybrid cloud solution, do you extend that capability on-prem and maybe even to other vendor clouds? That's a good one. David Floyer. I've got two, if I may. I'll say them very quickly. The first one, John, is you being a AWS that developed a great international network, a fantastic performance. How is AWS going to avoid conflicts with the EU, China, Japan, and particularly about their reticence about using any US-based nodes from an in-country and from in-country telecommunication vendor? So that's my first. And the second is, again on AI, what's going to be the focus of AWS in applying the value of AI? Where are you going to focus first and to give value to your customers? Rob Wolf, do you want to ask a question? Yeah, I'd like to... One thing I didn't raise in terms of the challenges is Amazon overall is expanding so fast into all kinds of areas. Whole Foods, we saw this. I did ask Jassy, how do you contend with reality that a lot of these companies that you're now bumping up against as an overall company now don't necessarily want to depend on AWS for their critical infrastructure because they're competitors? How do you deal with that? Great question. Yeah. David Vellante. Yeah, my question is, would be, you know, as an ecosystem partner, what advice would you give, because I'm really nervous that as you grow and you use the mantra, well, we do what customers want, that you are going to eat into my innovation. So what advice would you give to your ecosystem partners about places that they could play and a framework that they should think about where they should invest in that value without the fear of, you know, you consuming their value proposition? So it's kind of the ecosystem analog to the customer question that Rob asked. So the one that I would have for you, John, is at what... the promise is all about scale. And they've talked a lot about how software at scale has to turn into hardware. What will Amazon be in five years? Are they going to be a hardware player on a global basis? Following this China question, are they going to be a software management player on a global basis? And they're not going to worry as much about who owns the underlying hardware because that opens up a lot of questions about maybe there is going to be a true private cloud option and AWS will just try to run on everything and really be the multi-cloud administrator across the board. The Cisco as opposed to the IBM in the internet transformation. All right, so let me summarize very quickly. Thank you very much, all of you guys, once again for joining us in our action item. So this week we talked about AWS re-invent. We've done this for a couple of years now. theCUBE has gone up and done 30, 35, 40 interviews. We're really expanding our presence at AWS re-invent this year. So our expectation is that Amazon has been a major player in the industry for quite some time. They have spearheaded the whole concept of infrastructure as a service in a way that in many respects nobody ever expected. And they've done it so well and so successfully that they are having an enormous impact way beyond just infrastructure in the marketplace today. Our expectations that this year at AWS re-invent, we're going to hear a lot about three things. Here's what we're looking for. First is AWS as a provider of advanced artificial intelligence technologies that then get rendered in services for application developers but also for infrastructure managers. AI for ITOM being, for example, a very practical way of envisioning how AI gets instantiated within the enterprise. The second one is AWS has had a significant migration as a service initiative underway for quite some time. But as we've argued in Wikibon research that's very nice but the reality is nobody wants to bond the database manager. They don't want a promise that the database manager is going to come over. It's interesting to conceive of AWS starting to work with application players as a way of facilitating the process of bringing database interfaces over to AWS more successfully as an onboarding roadmap for enterprises that want to move some of their enterprise applications into the AWS domain. We mentioned one in particular, SAP, that has an interesting potential here. The final one is we don't expect to see the kind of comprehensive edge answers at this year's re-invent. Instead, what our expectation is is that we're going to continue to see AWS provide services and capabilities through serverless, through other partnerships that allow AWS or the cloud to be able to extend out to the edge without necessarily putting out that comprehensive software that's comprehensive software stack as an appliance being moved through some technology suppliers. The green grass, certainly serverless, Lambda and other technologies are going to continue to be important. If we finalize overall what we think one of the biggest plays is, we are especially intrigued by Amazon's continuing build out of what appears to be one of the world's fastest, most comprehensive networks and their commitment to continue to do that. We think that this is going to have implications far beyond just how AWS addresses the edge to overall how the industry ends up getting organized. So with that, once again, thank you very much for enjoying Action Item and participating and we'll talk next week as we review some of the things that we heard at AWS and we look forward to those further conversations with you. So for Peter Burris, the Wikibon team, SiliconANGLE, thank you very much and this has been Action Item.