 Live from Silicon Valley, it's theCUBE covering Google Cloud Next 17. Okay, welcome back everyone. We are here live in Palo Alto for special two days of coverage of Google Next 2017, events in San Francisco, sold out 10,000 plus people. You know, really an amazing turn of events. Amazon Web Services re-invent had 36,000. Google's nipping at their heels, although different, we're going to break down the differences with Google versus Amazon because they're really two different things. And again, this is CUBE coverage here in Palo Alto Studio getting reaction sponsored by Intel. Thanks to Intel for allowing us to continue the wall-to-wall coverage of the key events in the tech industry. Our next guest is Valperitra Vici, who's the chair, board member of the Cloud Native Compute Foundation board member, right? Welcome back, you were here last week from Mobile World Conference, great to see you. SiliconANGLE contributor, what's your reaction to the Google keynote, Google News, not a lot of news. You saw the SAP, that was their biggest news and the rest were showcasing customers. Most of the customers were G Suite customers. Yeah, exactly. So I would say my first reaction is a bit of a rough keynote. You know, there's definitely not quite as much polish as Microsoft had in their head and of course, Amazon nowadays in the cloud era. But what's interesting to me is there's a whole battle around empathy right now. So the next-gen developers in the cloud erotic talk about user empathy and that means understanding the workflow of the user and getting the user to consume more of your stuff. Snapchat gets user empathy for the millennial generation, better than anybody else, Facebook as well. So you see Google, we emphasize this, even the Google Twitter account, emphasizes developer productivity and they have really strong developer empathy. But what AWS has, Amazon with AWS is enterprise empathy, right? They really understand how to package themselves and make themselves more consumable right now for a lot of mainstream enterprises. They've been doing this for three, four years that they're reinventing events now. Whereas Google was just catching up, they've got great developer empathy but they're just catching up on enterprise empathy. That was the main difference as I see. Yeah, and I think that's an important point, Val. Great, great point. I think Amazon certainly has, and I wrote this in my blog post this morning, getting a lot of reaction from that actually and some things I'm going to drill down on the network and security side, some Google folks DMing me way to do that. But really, Amazon's lead is way out front on this. But the rest, you know, call them IBM, not in any particular order, IBM, Oracle, Google, SAP, others who put Salesforce since we're talking SaaS and Adobe, they're all in this kind of pack. It's like a NASCAR, you know, pack and you don't know who's going to slingshot around and get out there. But they all have their own unique use cases. They're using their own products to differentiate. We're hearing Google, and again, this is a red flag for me because it kind of smells like they're hiding the ball. G Suite, I get the workplace productivity, it is a cloud app, but that's not pure cloud conversations. If you look at the Gartner, Gartner's recent last report, which I had a chance to get a peek at, there's no mention of SaaSifications. Google's G Suite's not in there, so the way cloud is strictly defined doesn't even include SaaS. If you're going to include SaaS, then you got to include Salesforce in that conversation or Adobe or others. Exactly. So this is kind of an optical illusion in my mind. And I think that's something that points to Google's lack of traction on customers and the enterprise. The story behind the scenes, Kubernetes is so important and why I'm involved with the CNCF. If anything, the first wave of cloud adoption, particularly by enterprise, was centered around the VM model. And infrastructure is a service based on VMs, Amazon, AWS is the king of that. What we're seeing right now is developers in particular that are developing the next generation of apps. Most of them are already on our phones and our tablets and our houses and so forth, as in all these echo devices. That is a container-based architecture that these next-gen applications are based on. And so Kubernetes, in my mind, is really nothing more than Google's attempt to create as much of a container-based ecosystem at scale so that the natural home for container-based apps will be GCP as opposed to AWS. That's the real long-term playing why Google's investing so heavily in Kubernetes. Is that counter-intuitive? Is that a good thing? I mean, it sounds like it's just they're trying to change the goalpost, if you will, to change the game because we had Joe Arnold on with the founder of Swiftstack. And ultimately, clouds or clouds, inter-clouding and multi-cloud is important. Does Kubernetes actually help the industry or is that more Google-specific in your mind? I think it will help the industry, but the industry itself is moving so rapidly. We're seeing serverless right now and functions as a service. As I think the landscape is shifting away from what we would think of as either VM or container-based infrastructure as a service towards having the right abstractions. What I'm seeing as the most innovative enterprises today don't really care about their per minute or per hour cost for a cycle of computer, a byte of network transferred or stored. They care about big table, big query, the natural language processing, visual search, and a whole category of these new AI-based applications that they want to base their own revenue-generating products and services based on. So it's abstraction now. Is that new battlefield? AWS brings that cult of modularity to it. They're delivering a lot of cool services at a very high level, Lambda centered based on really cool modularity, whereas Google's doing it with just very, very elegant abstraction. So at the developer level, at the technical level, that's what the landscape is right now. Are you happy with Google's approach because I think Google actually doesn't want to be compared to AWS in a way. From what I can see from the keynote now. Only by revenue. Well, certainly they're going to win that by throwing G Suite on it, but this is, again, a philosophy game, right? I mean, Andy Jassy is very customer focused, but they don't have their own SaaS apps, except for Amazon, which they don't count in the cloud. So their success is all about customers, building on Amazon. Google actually is its own customer, and they actually include that in, as does Microsoft with Office 365. Yeah, that's the irony is if we go back to enterprise empathy, I think it's Microsoft has that legacy of understanding the enterprise better than all the others. And they're beginning to leverage that we're definitely seeing Azure slotting comfortably to a number two position behind AWS. But it really does come back to, are you going to lead with a propeller head lead in technology, which Google clearly has? They've got some of the most superior technology we were rattling off some of the speeds and feeds that one of their product managers shared with you this morning. They've got amazing technology, that's unquestioned. But what they do have also is this reputation of almost flying in rarefied air when it comes to enterprises. What do you mean by that? What I mean by that is that most enterprise IT organizations, even the progressive ones, have a hard time relating to Google technology. It's too far out there. It's too advanced in some cases. They just can't understand it. They've never been trained in college courses on it or even post-grad courses on it. MBA is older than three years old. Don't even reference the cloud. So there's a lot of training and a lot of knowledge it has to be conducted on the enterprise side. AWS has packaged that technology there as the modularity in such a way that's more consumable. Not perfect, but more consumable than any other cloud vendor. And that's why with the early head start they've got the biggest enterprise traction today. Yeah, I mean, I'm really bullish on Google. I love the company, I've been following them since 98. A lot of friends here at Palo Alto, a lot of Google was living in my neighborhood. They're all around us. Larry Page, seeing them around town. Great, great company. And they've always been kind of like an academic, speed of academic, very strong technically. And that is clearly they're playing that card. We have the technology. So I would just say that to counter that argument would be if Google, if I'm Google, if I'm on the team with Guy in green and you know, look at what I would do is we want to be the intel for the cloud. So the hard and top is we don't really care if people are trained. Should be so easy to use, training doesn't matter. So I mean, that's a little bit more of an arrogant approach, man. I don't think Google is being arrogant in the cloud. I think that ship is sail. I think Google has kind of been humbled in the sense in recognizing that the enterprise hard, they're checking the boxes. They have a partner program. You're right. I mean, if you take a look at their customers today you've got Spotify, it's Snap and Evernote and you know, Pokemon Go and Niantic. All the leading edge technology companies that have gone mainstream, they're, you know, startup oriented Snap. Of course, they're on Google Cloud, but that's not enough. You know, the enterprise, I did a seminar just last week promoting Container World with Jim Ford from ADP. The enterprise is not homogeneous. The enterprise is complicated. The L word legacy is all over what they have to budget and plan for. So the enterprise is just a lot more complicated than Google will acknowledge right now. And I believe if they were to humanize some of their advanced technology and package it and price it in such a way that AWS, you know, where they're seeing success, they'll accelerate their inevitable sort of leap to being one of those top three contenders. So I'm just reading some of my, I'm putting together because for the Google folks I'm in an interview, I'm just prepping for this. Just networking alone, isolating cloud resources. That's hard, right? So, you know, virtual network in the cloud, Google's got the virtual network. You got multiple IP addresses per instance, ability to move network interfaces and IPs between instances, enhanced networking support, network traffic logging, virtual network peering, managed NAT gateways, subnet level filtering, IPv6 support, use any CIDR including RFC, 1918, multiple network interfaces instance. I mean, this is complicated, it's not easy. So, you know, I think, you know, the strategy is going to be interesting to see how does Google go into the point-to-point solutions that, or they just say, this is what we got, take it or leave it and try to change the game. That's what they've been up until now. I don't think it's working because they have very formidable competitors that are not standing still. So, I think they're going to have to keep upping their game again, not in terms of better technology, but in terms of better packaging, better accessibility to their technology, better trust, if you will, overseas. Cloud is a global game, it's not US only, and trust is so critical. There's a lot of skepticism in Europe today with the latest WikiLeaks announcements or Asia today around any American-based cloud provider truly being able to isolate and protect my citizen's data, you know, within my borders. I think Google Cloud is one fatal flaw that I'm looking at. All the data is that, in the analysis that we've been looking at with Wikibon team and our research is that, there's one thing that jumps out at me. I mean, the rest of the role I look at is, you know, Google's got such great technologies, they can move up fast, they can scale up code, but the one thing that's interesting is their architecture, the way they handle their architecture is, they can't let customers dictate data where data is stored. That is a huge issue for them. And to your point, if a user in Germany is using an app and it's got to stay in Germany. This is back to the empathy disconnect, right? As an abstraction layer for a developer, what I want is exactly what Google offers. I don't want to care as a developer where the bits and bytes are stored. I want this consistent uniform API. I want to do cool stuff with the data. The operation side, particularly within legal parameters, regulatory parameters, you know, all sorts of other costs and quality assurance parameters, they really care about where that data is stored. And that's where having more enterprise empathy in their thinking and their offerings and their pricing and their packaging will leapfrog Google to where they want to be today. Of Alper Gevici, great analysis. I mean, I would totally agree just to lock that in. Their developer empathy is so strong and their operational one needs to be, they need a blind spot there or they've got to work on that. And this is interesting because people who don't know Google are very strong operations. It's not like they don't have any ops chops. They are absolutely in the five nines. They're awesome operations, but they've been operations for themselves. Exactly. So that's the distinction you're getting at, right? Absolutely. Okay, so the next question I got to ask you is back to the developer empathy because this, I think, is a really big opportunity for Google. So I pointed out the fatal flaw in my opinion was the data locality thing. But I think the opportunity for Google is to change the game using the developer community opportunity because you mentioned Kubernetes. There is a huge open source, I won't say transformation, but an evolution to the next generation. You're starting to see machine learning and AI start to tease out the leverage of not just data now. Data's become so massive, now you have data sets that can be addressable and be treated like software programs. So data as code becomes a new dynamic with AI. So with AI, with open source, you're seeing a lot of activity, CNCF, the Cloud Native, Compute Foundation. Folks should check that out. That's an amazing group in the Linux Foundation. This is an awesome opportunity for Google to use Kubernetes as saying, hey, we will make orchestration of application workloads. Absolutely. This is something that Amazon's been great with open source, but they don't get a lot of love. Amazon has a blind spot on containers. Let's call it spade to spade, let's not be around the bush. They do have a blind spot around containers. It is something they strategically have to get a hold of. They've got some really interesting proprietary offerings, but it's not a natural home for a Docker workflow. It's not a natural home for a Kubernetes workflow yet. And it's something they have to work on. AI as a use case couldn't be more pertinent to business today because it's that quote, you know, the future is here, but unevenly distributed. That's exactly where AI is today. The businesses that are figuring it out are really leaping ahead of their competitors. We're getting some great tweets. My phone's blowing up, Val, you got great commentary. I want to bring up some, I've been kind of over the top with the comment that I've been making. It's maybe mischaracterized, but I'll say it again. There seems to be a cold war going on inside the communities between what Kubernetes has done. We've seen Docker, we've seen Docker containers be so successful in this serverless vision, which is absolutely what we're cloud-native is needs to be and that notion of separating out physical gear and addressability, making it completely transparent, full DevOps, if you will, to who's going to own the orchestration and where does it sit on the stack? And what Kubernetes to me is interesting is that it tugs at some sacred cows in the container world. And opens up the notion of multi-cloud. I mean, assume latency can be solved at some point. It was actually a core religion. What impressed me about the whole Kubernetes community and the communities that's great as strengths, by the way, is the fact that they had religion around multi-cloud from day one. It wasn't about, we'll add it later because we know it's important. It's about portability and even Docker lent that to the community. Portability is just the number one priority and now portability at scale across multiple clouds dynamically orchestrated, not through potential for human error, human intervention, as we saw last week. That's the secret sauce there today. I think not only is a Cold War is a negative connotation but I think it's an opportunity to be sitting in the sun if you will, on the beach with a Pina Colada because if you take the Kubernetes trend, that's got developer empathy with portability, that speaks to what developers want. I want to have the ability to write code, ship it onto the network, and have it integrate in nicely and seamlessly. Things can software can do all that and AI can help and all those things. Connecting with operational challenges. So what is in your mind that intersection because let's just say that Kubernetes is going to develop a nice trajectory which it has now and continues to be a nice way to galvanize the community around orchestration, portability, et cetera. Where does that intersect with some of the challenges and needs for operational effectiveness and efficiency? So the dirtiest secret in that world is data gravity. It's all well and fine to have workload portability across multiple instances in the cluster across multiple clouds, so to speak. But data has weight, data has mass and gravity and it's very hard to move, particularly at scale. Kubernetes only in the last few releases with a furious pace of evolution, 1.4, 1.5 has a notion of provisioning persistent volumes, this thing they affectionately called pet sets that are not a stateful set, I love that name. Taddles. Exactly. So Google is waking up and Kubernetes I should say in particular is waking up to the whole notion of managing data is really that last mile problem of cloud portability and operational maturity and planning around data gravity and overcoming where you can data gravity through better operational procedures is where this thing is really going to take off. I think this is where, I like Google's messaging and I like their posture on machine learning and AI, I think that's key. But Amazon has been doing AI, they got machine learning as a service, they've had Kinesis for a while. In fact, Redshift and Kinesis were their fastest growing services before Aurora became the big thing that they had. So I think they're addressing the ingest with the trucks and the snowmobile stuff. So I think certainly Amazon's been doing that data and then rolling it into some sort of AI. And they've been humanizing it better, right? I can relate to some of Amazon's offering and sometimes I have it in the house. So the packaging and just the consumability of these Amazon services today is ahead of where Google is and Google arguably has a superior technology. And I think, I was laying out my analysis of Google versus Amazon but I think it's not fair to try to compare them too much because Google is just making their opening moves on the chessboard. Because Diane Green got to give her credit, she's really starting behind. And that's been talked about, but they are serious. They're going to get there. The question is, what does an enterprise need to do? So your advice to an enterprise would be what? Stick with the use cases that are either Google specific apps or cloud native, where do you go? How do you? I would say remember the lock-in days of the Unix vendors and even Microsoft that they're in their heyday and definitely think multi-cloud. Cloud first is fine. But think really data first in the cloud before you think a particular cloud first. Always keep your options open. Seek the highest levels of abstraction, particularly as you're innovating early on and fast failing in the cloud. Don't go low right away. Go low later on when your operation lies in its scale and look into squeeze efficiencies out of a new product or service. Don't go low, you mean don't go low on the stack? Don't go low on the stack, exactly. Start very high on the stack. What would be an example? So Lambda, you know, taking advantage of, if we're bringing Kinesis IoT workflows, all sorts of sensor data coming in from the edge. Don't code that for efficiency day one and switch to Kafka or something else, it's more sophisticated. But keep it really high level as events triggering off, whether it's the IoT SDK and sensor inputs, whether it's S3 events, DynamoDB events, write your functions at a very, very high level. Get the workflows right. Pay a bit more money up front. Pay premium for the fast-fail. Also bootstraps for the training challenges you mentioned. So with Google, pick some things that are known out there. You mentioned IoT and one of the things that I was kind of disappointed in the keynote today there wasn't much talk about IoT. You're not seeing IoT in the Google story. That may come up in tomorrow, Urza's keynote, that may come up tomorrow in a more technical context. But you're right, that's an area both Azure and AWS have a demonstrable lead right now as they've had really good SDKs out there to be able to create workflows without even being an expert in some of the devices that you might own and maintain. Google's got some differentiation. They've got some things I'll highlight, one that I like I think is really compelling. TensorFlow, TensorFlow's got a lot of great traction and then Intel is writing chips with their Skylake product that actually runs much faster silicon. So is NVIDIA, it's a GPU game, it's a CPU game when it comes to machine learning. And it's a really fast thing to see. What does that mean for users? That's exciting, you're smiling. I get geeked out on that because if you think about that, if you can have a relationship between the silicon and the software, what does it mean from an impact standpoint? Do you think that's going to be a good accelerant for the game? Massive accelerant, and this is where we get into sort of more rarefied air with Elon Musk's quota on the fact we'll need universal income for society. There are a lot of static tasks that are automated today. There's more and more dynamic tasks now that these AI algorithms through machine learning can be trained to conduct in a very intelligent manner. So more and more task-based work all over the world, including in robotics context, but also call centers, stock brokerage, for example, has been demonstrated that AI algorithms are superior to humans nine times out of 10 in terms of recommending stocks. So there's a lot of white-collar, as well as blue-collar work, that's just going to be augmented and then eliminated with these technologies. And the fact that you have major players that's economies of scale, such as Intel, Nvidia, and so forth, accelerating that, making it affordable fast, low power in a certain edge context, that's really good for the industry. So day one of two days of coverage here with Google, just thoughts real quick on what Google needs to do to really conquer the enterprise and really be credible, viable, successful, number two or leader in the enterprise. I'm a big fan, you know, I've had personal experience with fast following as opposed to leading or innovating sometimes in terms of getting market traction. I think they should unabashedly and unshamely examine what Microsoft or what Amazon are doing right in the cloud. Because, you know, simple things like conducting a bit more of a smooth keynote, Google doesn't seem to have mastered yet right now in the cloud space. And it's not rocket science, but shamelessly copying what works, shamelessly copying the packaging and the humanization of some of the advanced technologies that Amazon and Microsoft have done in particular, and then applying their technical superiority, you know, their uptime availability advantages, their faster networks, their strong consistency which is a big deal for developers across their regions, emphasizing their strengths after they package and make their technology more consumable as opposed to leading where the tech specs. And you have a lot of experience in the enterprise. Table stakes out there that are pretty obvious that they need to check the boxes on what it would be, what? A really good question. I would say first and foremost, you really have to focus on more, you know, transparent pricing. I think something, there's a whole black art in terms of optimizing your AWS usage and there's an industry that's formed around that. I think Google has, and they have had blogs advertising a lot of advantages they have in the granularity and the efficiency of their auto scaling up and down, but businesses don't really map that. They don't think of that first, even though it can save them millions of dollars as they do move to cloud first approaches. Yeah, and I think Google's got to shake that academic arrogance in a way that they've had a reputation for. Not that that's a bad thing, I'll give you an example. I love the fact that Google leads a lot of price performance at many levels in the cloud, yet their SLAs are kind of wonky here and there. So it's like, okay, enterprises like SLAs, so you got to nail that. And then maybe keep the price a little higher and can make more money, but so you're saying is that enterprise might not get the fact that it's such a good deal. It's like enterprise sales one-to-one. You talk about the operational benefits, but you also talk about financial benefits and business benefits. Catching it in those three contexts in terms of their technical superiority would do them a world of good as they seek more and more enterprise adoption. All right, Balford Trevici, CTO, also a CTO, and also on the board of the Cloud Native Compute Foundation known as CNCF, a newly formed organization, part of the Linux Foundation, really looking at the orchestration, looking at the containers, looking at Kubernetes, looking at a whole new world of app enablement. Falf, thanks for the company, great to see you. Turning it to be a guest contributor here on theCUBE Studio, appreciate his time. This is theCUBE, two days of live coverage. Hope to have someone from Google on the security and network side coming in and calling in, we're going to try to set that up. A lot of conversations happening around that. A lot of great stuff happening at Google Next. We got all the wall-to-wall coverage. Reporters on the ground in San Francisco as well as analysts, and of course in studio reaction here in Palo Alto. We'll be right back.