 Live from San Francisco, it's theCUBE. Covering IBM Think 2019, brought to you by IBM. Hello everyone, welcome back to theCUBE's live coverage here at IBM Think 2019 in San Francisco. Our exclusive coverage, day four, four days of coverage. Events winding down, I'm John Royce, Stu Miniman. Our next guest, Eric Herzog, CUBE alumni CMO of IBM Storage and VP of Storage Channels. Eric, great to see you, we're in the Hawaiian shirt as usual. Great, well thank you, I can't come to the Cuban, not wear the Hawaiian shirt. You guys give me too much of a heart attack. Love getting you on to get down and dirty on storage and the impact of cloud and infrastructure. First, you gave a great talk yesterday to a packed house, I saw that on social media, great response. What's going on for you at the show, tell us. So, the big focuses for us are around four key initiatives. One is multicloud, particularly from a hybrid perspective, and in fact, I had three presenters with me, handless end users, all of them were using multiple public cloud providers and all of them had a private cloud. One of them also was a software as a service vendor, so clearly they're really monetizing it. So that's one, second one is around AI, both AI that we use inside of our storage to make it more efficient and more cost effective for the end user, but also as the platform for AI workloads and applications. Cyber resiliency is our other big theme. We've got all kinds of security, yes, everyone is used to, of course, the great wall of China protecting you and then, of course, chasing the bad guy down when they breach you, but when they breach you, be sure, nice, if everything had data rest encryption. Or when you teared out to the cloud, you knew that it was being backed up or teared out, fully encrypted. Or, how about something that could help you with ransomware and malware? So we have that, and that's a storage product, not a regular, you know, what you think of from a security vendor. So those are the big things that we've been harking on at the show. One of the things I've deserved you've been very active out in the field. We've seen you at a lot of different events, Cisco Live, others. You guys have had an interesting storage product portfolio, very broad and specific leadership and categories, but you also have the ability to work with other partners. It's been a big part of your strategy, you get the channels. What is that, how would you summarize the current story around IBM storage and systems? Because it's now an ingredient part of other people's infrastructure. When cloud storage didn't become the key equation, how would you describe the IBM storage, posture, product portfolio, what are the key things? So I think the key thing from a portfolio perspective, while it looks broad, it's really four things. Software-defined storage, which we also happen to bet on an array, so theoretically that's one product line, it's the same exact software. Other vendors don't do that. They have an array pack, and you buy the array, but if you buy their software-defined storage, it's actually different software. For us it's the same software. Then we have modern data protection, and then we have management play, and that's kind of it. I do think one of the big differentiator for us is even though we're part of IBM, we have already been working with everyone anyway. So as we talked about at Cisco Live, for Spectrum Protect alone, our modern data protection platform, we have 400 small and medium cloud service providers all over the world that they're back up as a service based on it. So even though IBM cloud has their own cloud division, theoretically we're enabling the competition, but we've had that story at IBM storage now for four years. Storage anywhere, basically. It's the theme here, AI anywhere, storage anywhere. And it's not the official tagline, but that's the philosophy with software. Yeah, so even if you think, look at AI. We have an AI reference architecture with the power product line. We also have an AI reference architecture with the NVIDIA product line, and we're working on a third one right now with another major server vendor because we want our storage to be anywhere there's AI and anywhere there's a cloud, big, medium, or small. All right, Eric, let's tease that out a little bit because I had a great conversation with an IBM fellow yesterday and we think back, you know, 10 years ago, we talked about hybrid and multi-cloud. When we talked about an application, it's am I spanning between environments? Am I bursting between environments? And architectures just didn't work that way. Today, microservices are architecture. There's pieces of the solution that can live in lots of environments. Compute, I can spin up almost anywhere at any time. Data doesn't move. And I need to worry about my data. I need to worry about my security. So there's certain things that multi-cloud, like data protection, cyber resiliency, those kind of ones need to live everywhere. But when I talk about storage, I'm not moving my storage and my persistent database all over the place. So help us kind of tease out as to what is the multi- everywhere and what is the, you know, the computer's going to actually move to that data. Help us squint through that a little bit. So let's do the storage part first. So most applications, workloads, and use cases that are either business critical or mission critical are going to stay on-prem. Doesn't mean you can't use a public cloud provider for overflow, whether that be IBM or Amazon or Microsoft, or like I said, the 400 cloud providers that we sell to that are not IBM. So, but you're still going to have this hybridness where the data is partially on-prem and off-prem. In that case, you're going to be using the public cloud provider. And by the way, we did a survey, IBM did. And when you're looking enterprise, so let's say companies that are three or four billion US and up anywhere in the world, you're seeing that most of them are using five or six different public clouds, whether that be Salesforce.com, which really is sales enablement, software as a service. We have a startup that we work with who uses IBM's Flash System and they do cybersecurity as a service. That's their whole business. So all of the software vendors that now deliver not on-prem but over the cloud. Then you've got regular public cloud providers for file, block, and object. For example, we not only support IBM Cloud Object Storage protocol, but S3. So we have customers that put data out in S3. We have customers that put it out in other clouds because as you know, S3 has become the de facto standard. So all the mid and small cloud providers use it. So I think what you've got is hybrid cloud is a sort of a subset of multi-cloud and then multi-cloud what you're seeing is because of software as a service. Could even be geographic issues. We have a lot of data centers at IBM Cloud, some of the other three major cloud providers, but we are not in all 212 countries. So if you have the law like in Canada where the data has to physically stay within the premises of Canada, now we all happen to have data centers that are big enough, but that doesn't mean we have data centers in every country. So you have legal issues, you have applications, what applications are good, that makes sense. What about pricing? And as you know, some big companies still buy regionally. Yeah, Erica, one of the things I'd love to get your perspective on is the SaaS providers, because if we look at the storage market, in many ways there was like the threat of public cloud, but really you got to follow the application, you got to follow the data, and as SaaS proliferation happens, your data is going to go with that. You have them as customers in a lot of environment. What are you seeing from the SaaS providers? How do they choose what offerings they have and how do they look at their data center versus public cloud mix? So when you look at a SaaS provider, they've got a couple different parameters that they've looked at, which is why we've been very successful. One is performance, they already know they're subject to vicissitudes of the cloud, so you can't have any bottleneck in your core data center because you're serving that app up. And if it's too slow or it doesn't work right, then of course the end user will go get a different piece of software from another SaaS provider. Second one is availability, because you have no idea when Wikibon theCUBE is going to turn on that service, it could be the middle of the night, right? If you guys expand to Asia, you guys will be asleep, but your guy in Australia will be using that software. So it can't ever go down, right? So availability, resiliency, can it hand up pounding if Cube Wikibon becomes ginormous and you buy all these other analyst firms, actually you know the biggest analyst firm in the world, and you have thousands of people, well guess what, now you're hammering on that software, so it's got to be able to take that workload abuse, right, and that's the kind of thing. So they look for that. That scale basically, scale is critical. At scale, right, they cannot have any issues of resiliency or availability and performance, so A, they're usually going all flash, some of them will buy like a tape or the older all hard drive arrays as a backup store, ideal for IBM Cloud Object Storage, but again, the main thing they focus on is flash, because they're serving up that software. Let me ask you a question, so I know you've been in this business for a long time, storage, you know everything about the speeds and fees, but also you've been a historian too. You're in the front end, IBM has got a killer strategy with cloud private doing very well with open shifts and red hat acquisition, you're now poised to essentially bring cloud scale across multiple clouds and with AI, really put storage at the center of the action. How is storage now positioned and how should customers think about storage because scale is table stakes, enabling developers to program the infrastructure as code, how does storage and how has it changed and how are you guys positioned to take advantage of that? How would you kind of explain that to a customer? Yeah, so I think there's a couple of changes. First of all, you're looking for a storage vendor, which should be us, but you're looking for a storage vendor that is always making sure, for example, when microservices first came out and containers. Okay, great, except when containers came out and it's still a problem, you don't have storage consistency, whereas in a VMware, Hyper-V or KVM environment, you do, so when you move things around, you don't lose the dataset. Well, we have persistency storage. So the key thing you want to look for is a storage vendor that's going to stay on that leading edge as you move. Our copy data manager has an API so the developers could spit up their own environments but use real data. So as you guys know well from your past, that the last thing you want to do is have the DevOps guy be developing things on foe datasets, try to put it in production and then the real dataset doesn't work. At the same time, if they put it out to a public cloud provider, you could have a legal or security breach, right? So by being able to take modern data protection as an example and not just to have grandfather, father, son backup, we all remember that. I remember it better than you guys since I'm older but that's backup, right? It's not backup anymore, it's modern data protection. You need to be able to take the snapshot, the replica or the backup dataset and use it for development. So you want a storage vendor that's going to be on the leading edge of that and we've done that at IBM on the Caner side, the modern data protection side and we'll continue to do the whole multi-cloud thing. IBM, as you know, is now all about multi-cloud, what Red Hat's been doing. The storage division of IBM's been working with Red Hat for 15 years. Going to the Red Hat summit every year, I know you guys do the cube from there sometimes. You're on. But this is software defined. So at the end of the day, the software defined that with arrays have paid off. Yes. You'd say that would be kind of a key linchpin. I would argue that while there's some hardware aspects to it, so for example, our flash core modules give us a big differentiator from a flash perspective. In general, the number one differentiators for a strong, powerful array vendor is actually the underlying software code. The RAID stack, what you can wrap around it, file block and object support, what could you enhance, our spectrum discover, allowing you to use metadata about unstructured data, whether that be in the file space or the object store. That allows the data scientist to dramatically reduce the time it takes to prep the data when they're doing either AI or an analytic workload. So we just saved them money, but we're really a storage company that came up with something that a data scientist could use because we understand how storage is that essential foundation and how you could literally use the metadata for something actually valuable, not to a storage person, because a data scientist is not the storage guy, of course. And Erica, I'd love to get your feedback. What are some of those key discussions you're having with customers here at the show? We've been talking a lot this week, digital transformation, AI into everything there. Are those some of the themes? What are the struggles that really the enterprise today are facing and how your group's helping them? So one of the big things is understanding that it's going to be multi-cloud. And so because we've already been the Switzerland to the storage industry and working with every cloud provider, all the big ones, including ones that compete with our own sister division, but all the little small ones too, right? And all the softwares of service vendors we work with that were the safe bet, you don't have to worry about it because whoever you pick or for a big enterprise, in fact, I had Etna on stage with me. And he said, he's using seven different clouds, one of which is their private cloud and then six different cloud providers they use. And he said, not counting salesforce.com and I forgot the other name. So really if you count the softwares there, she really got like nine clouds. She said, I use IBM because I know it's going to work with whoever I, and you're not going to say, oh, I don't work with this one or that one. So that's been obviously making sure everyone realizes that whole company is embracing it as you saw. And what we're going to do obviously with Red Hat and continue for them to participate with all of their existing customer base that they've been doing for years. So you see multi-cloud and sweet spot that highlights your value proposition. Would you say that? I would say that. And then the second one is around AI. All the storage vendors including us have had AI sort of what I'll call inside of the box, inside of the array, and use that to make the array better. But now with AI being ubiquitous from a workload perspective, you have to have the right foundation underneath that. Again, performance resiliency availability. You're going to use AI in a giant car factory, and it's going to run all of those machines. You better make sure the thing never fails because then the assembly line goes down and those things are hundreds of millions of dollars of build every day. So that's the kind of thing you got to look for. So AI has got to have the right platform underneath as well. All right, give us some reporting from the field as you're out and doing a lot of talks, talking to customers. Give it a couple of anecdotal examples of where the leading edge is in storage and where are use cases that would be a good tell sign of where this kind of multi-cloud is going. Can you just give some examples of the use cases situation and kind of why is that relevant for where everyone will be going? Where's the puck going to be so I can skate to where the puck is, as they say. So from a multi-cloud perspective, A, you've got to deal with how your company is structured. If you have a divisionalized company or one that really lets the regions make their own by decisions, then you may have NTT cloud in Japan. You may have Alibaba in China. You may have IBM cloud Australia and then you might have Amazon in Latin America. And as the IT guys, you got to make sure you're dealing with that and embrace it. One of the things I think from an IT perspective is why I'm wearing the Hawaiian shirt. You don't fight the wave, you ride the wave and that's what everyone's got to realize. So you're going to use multi-cloud for letters and remember, the cloud was the web, was the internet. It's actually all the same stuff from a long time ago, the mid 90s, which also means now procurement's involved. And when procurement's involved, what are they going to say to you? Did you get a bid from IBM cloud in addition to that bid from Amazon and Microsoft? So it's changed the whole thing up. I can just go to any cloud I want to. Now procurement's involved that even mid-sized companies, procurement says, you did get another bid, right? Did you not? Which for server, storage and network vendors, that's been the way it's been for 35, 40 years. Well the bids are changing too. So what are the requirements now? Amazon has a cloud, they have storage, you have storage, but people have on-premise, they have multiple environments. I mean, if the world is one big data center with multiple regions and locations, this is the resiliency piece. But what's the new requirements as procurement gets involved? Because procurement isn't dictating the requirements. They're getting the requirements from the application workloads and the infrastructure. So what are the new requirements that you see? So I think the thing you're seeing is if you take cloud just a couple of years ago, I'm going to put my storage out there. Okay, great. I need this kind of availability. Oh, that's extra money. Sorry, Mr. Wikibon, Mr. Cube. We got to charge a little extra for that. Oh, we need a certain amount of performance. Oh, that's a little extra. And then for heavy transactional workloads, the data's constantly moving back and forth. Oh, we forgot to tell you that we're charging you every time you move the data in and every time you move the data out. So as you're putting in these RFPs, you need to be aware of that. Anything you got to do? Those are hidden costs. Those are hidden costs that are, I think the reason you're seeing such as the ride of the hybrid is people went to the public cloud and then someone in finance or maybe even in the IT group sat down with a spreadsheet and said, oh my God, we could have just bought an IBM Array or someone else's Array and actually had less money, even counting support money because every time we move it, but for archive, for backup, where you don't move the data around a lot, it's a great solution for anything. Then you have the whole factor in of software as a service. So part of that is the software itself. If you're going to go up against Salesforce.com, then whoever does, they better make sure the software's good. Then on top of that, again, you negotiate with the software vendor, I need it globally. Okay, what's the fee for that? So I think the IT guys need to understand that with the ubiquity of the cloud, you've got to ask way more questions. In the storage array business, everyone's got five nines and almost everybody's got six nines. Well way back when it was four nines, then it was five, now it's six. So you don't ask anymore because it just changes, right? And in the cloud is still new enough and the whole software as a service is a different angle and a lot of people don't even realize software as a service is cloud. When you say that, they go, what are you talking about? It's just I'm getting it over with service. Where do you think it comes from? A cloud data center. Well, the trend is software defined. You guys were on that early. Congratulations. And don't forget the hardware. You need high performance hardware as well, arrays and whatnot. So great, great job. Eric, thanks for coming on. Appreciate it. Thank you very much. Cube coverage here. I'm John Furrier, Stu Miniman, day four of our live coverage here in Moscone North in San Francisco for IBM Think 2019. Great packed house here at IBM Think. Back for more coverage after this short break.