 Live from San Francisco, extracting the signal from the noise, it's theCUBE. Covering Oracle OpenWorld 2015 from Studio C, brought to you by Cisco. Now your hosts, Stu Miniman and Brian Gracely. Welcome back to Oracle OpenWorld 2015. This is theCUBE, SiliconANGLE Media's flagship program. We go out to all the events, help extract the signal from the noise. Happy to have with me all day today, Brian Gracely on the program. Brian and I are also from wikibon.com. And our first guest for the day is when we talked to you yesterday and going to go into a little more detail, Eric Herzog with the IBM Storage Team. Eric, welcome back to the program. Great, Stu, Brian, thank you very much for having us. We love coming to theCUBE and talking storage and, most importantly, oceans of data. Yeah, so, you know, we've been in kind of the storage and IT industry for quite a while, Eric. One of the things I want to talk to you about is when we talk about storage, you know, it's really about not just storing my information. How do I get leverage out of it? You know, I know you probably can't really comment on it, but, you know, IBM made an acquisition of the data from the weather company. And you talk, I mean, that powers so much. I mean, we've all got it on our phones and everything like that. You know, how do you think of storage today in IBM? And, you know, how has that been changing in the last few years? Well, I think the key thing is, you know, I've been doing storage for 31 years. And while I love storage, what customers really love is the data and the business value they get from that data. So, you need to have reliable storage, available storage, and high performance storage. Short, it's the foundation, but just as in your home, the foundation is usually hidden from the rest of your house, right? You don't really know anything's wrong with it until you have a foundation problem. Storage is that foundation. But what's most critical for customers is their oceans of data, and how they analyze that data, how they can turn those oceans of data into real business value, like we've done with this acquisition that we did this morning, right? It's about that data, not about the storage or the technology, it's the data itself that has the value. Well, John Furrier would love, he's been using the term data oceans, and, you know, you kind of got the analytics pieces. We talked a little bit at IoT. There's all these new sources. There's all these new architectures. Dave Donatelli this morning talked about, you know, cloud as one of these seismic changes, and it's impacting architectures. So, you know, what do you see as some of the key trends, how's that, you know, transforming how we kind of build those foundations that you talked about in the storage stand? Well, I think the key thing is from a storage foundation technology, to be able to derive the value from this data, it's all about analytics, and those analytics engines could be on-premises. They can be, if you will, analytics software as a service, so literally through the cloud. It could be a hybrid. So if a customer of big giant global enterprise has a hybrid cloud, some of the compute could come off of a public provider, right, as extra overlap when they need more analysis of that data assets. But when you look at the information of things and really the big data revolution, which is creating these oceans of data, storage is the rock solid foundation underneath, and whether it's a cloud deployment, an on-premise deployment, or even someone who's trying to do big data with a traditional ID deployment, if the storage isn't the right thing underneath, that big data will end up with no value, and you won't be able to analyze that data effectively. So IBM's made a couple of large acquisitions, soft layer, blue box, really trying to bolster their cloud space, but then over the last six, eight months, they've also said, we're not going to just put it in the public cloud, we're going to extend that out to your data center, sort of give you a private version of that. What does that say about IBM's customer base in terms of feeling comfortable with just the public cloud versus wanting some sort of mix of on-premises and private? Well, I think what you're seeing is sort of a separation within the end user base of all types, whether they be big, small, or medium-sized companies. And some people are very sensitive to their data and how they put it out there. In certain countries, for example, in Europe, actually Canada as well, you're not allowed to have your data physically stored in a cloud that's not within the limits of that legal country, right? You can't have data in Canada in a data center in the United States, even if it's over the cloud. So these regulatory models, legal issues, financial issues for publicly traded company means that when you create a hybrid cloud, when you can use public resources, whether that be storage, compute, or any of the infrastructure that you need to keep that big data in those oceans of data active, but you also do it with some on-premises. And remember, if you create a hybrid cloud or private cloud, you're still much more agile, much more cost-effective, and much more flexible in how you deploy your entire data center resources, because it's not a traditional data center, it's a cloud data center. So even if it's private, you still save a lot of money and get a lot of extra value out of the IT resources you have, then the old model of traditional IT data centers. Yeah, but we've heard all week, security, security, security, lot about encryption. The trick with that is encryption is typically a huge bottleneck, a lot of CPU cycles to do that, especially when you're doing it in real time. What do you get a sense from your customers in terms of encrypting data? Where do they tend to make those trade-offs? Where does it make sense? Where does it, is it too costly to deal with it? Well, I think that, when you talk about not just encryption, but how do I secure the data? And so you've got perimeter security, you've got data security, the encryption capability. In fact, I'll give you a hint, next week we'll be announcing a new product with real-time encryption and we're using hardware-based encryption resources. So the impact on the storage itself is actually very small and we'll be announcing that actually next Tuesday. But you also need to have intrusion detection. And IBM now has a comprehensive model. We can encrypt the data on the disk and in fact, with our new solution, we can encrypt data that's not on IBM arrays, on competitors' arrays. Then you've got perimeter protection, which IBM offers and then you've got what I call intrusion protection. When the bad guy gets in, you've got to track them down and keep them from doing damage. I think the security community sort of recognized that people are going to get in and the way you want to do that is have a great fence, have your data encrypted, especially for the very sensitive data sets and have the capability which IBM offers to do intrusion detection. So we actually offer perimeter intrusion and in the case of the storage division, the ability to encrypt data on the disk whether it flash or traditional hard drives, we can encrypt the data AES-256. That's just a very much a defense in depth sort of approach to security and encryption. I think the approach from a security perspective needs to be holistic. If you only encrypt the data on the disk, that's not the right solution. All you do is have the perimeter fence, that's not the right solution. I think end users, even small end users, are realizing security is a holistic approach, not a piecemeal approach. Just encrypting data on disk would be piecemeal. Just having perimeter speech. The good thing is IBM, because of our multiple divisions, are able to craft a security solution that has all elements and gives them a real holistic solution to enhance them and minimize any badness that would happen from, quote, the bad guy getting into the data set. So, Eric, we were talking about cloud. One of the big changes in cloud is how I pay for it and how I think about consuming it. Rather than capital expense, it's more of an operational expense. And even in the storage industry, we've been talking about, what's the right way to talk about storage? I don't talk about price per gigabyte, when I'm talking about Flash, because there's performance, there's more, I get out of it. And even at Wikibon, we've talked, okay, if what's really important is the data and what I can do with it and business value, it's time to value its productivity enhancements. What are the conversations you're having with customers? How do they think about kind of the financial side of how they choose and differentiate between the solutions out there? Well, I think what you're seeing is, when you talk to the storage admin, he still talks about cost per gigabyte. When we are talking to the application owners, the database guys, the guys that own the logistics software, the manufacturing software, they don't think that way. They think, as you talked about, let's do the value of the data. So for example, we have several customers documented public references where to run a report on traditional storage took them 30 hours. Now they can run a report on our Flash systems in two. Well, that's time to money, as you mentioned. And so when the CIO looks at it, the CFO, and definitely the line of business owners, and usually the software app owners, whether it's an Oracle guy, or the SAP guy, or the DB2 guy, those guys look at it, I've got an SLA, and if I can do that report, or we can execute 20 million transactions in an hour instead of two million, that is time to money and time to value. And I think that that is changing the discussion around storage and how storage in this era of Flash in particular can derive real business value versus saying, hey, I just need a solid, reliable storage piece, yeah, yeah, yeah, yeah. Now not only do you need that, but actually by leveraging Flash technologies like our Flash systems, you actually can see things like 30 hour reports running in two. And then you never could have done that with regular storage in the old days. No way in hell could you have ever gotten there. Yeah, so you've been a product manager for a long time, you've seen a lot of development cycles. A lot of talk this week about you've got to have the exact same hardware system, everything on-premises as well as in your cloud. Talk about the complexities of, you know, if you're building a storage array, how complicated is it to go, I'm going to optimize for enterprise versus optimize for service provider. Is it realistic to be able to build this exact same system for both environments or is that changing? So actually with the exception of possibly multi-tenancy, it's the same because so many companies, if they're not doing a public cloud or doing either a hybrid or a private cloud, right? So if I was a giant bank and the chief legal officer came and said, you know what, Stu, there's no way you're going to be able to run in a public cloud because of all the regulation, blah, blah, blah, blah, blah. If I'm the sea, I'm so say, okay, I get it, Stu, but I'm going to create a private cloud that's going to be secure, that's just for us, blah, blah, blah, blah, blah. So when you create your storage for that environment, other than multi-tenancy, it's exactly what you would do for a cloud service provider or a managed service provider. It's the same model, except that it's private versus a public or a hybrid cloud. So I think that one of the key things is there are ways, you don't always have to have the same storage on both sides, right? You don't always have to have the same exact infrastructure. One of the powers of things like storage virtualization like our spectrum virtualize or server virtualization like VMware, Hyper-V, Oracle virtualization manager is that you can abstract the hardware. So the hardware does not have to be the same. Now, the real issue is when you have the same SLA. So then you've got to meet the SLA. So if that report I mentioned earlier really needs to run in Singapore and in the US in two hours, you've got to figure out a way to do that regardless of location. Then maybe you might need exactly the same, but otherwise I would disagree with Donatelli on that issue. Yeah, so, Eric, one of the biggest strengths of IBM is you guys have a huge application portfolio. You know, when we talk here at the event, it's like, oh, it's the red stack. They're going to optimize from the hardware all the way up to the software. Well, you know, there's a blue stack. And you guys have kind of the pure application family. There's a lot that IBM does to be able to really put together an environment, especially analytics, we've looked at that space a lot. Can you kind of give us from your hat on the storage side? Where is it that the single vendor stack is awesome and where are there places where IBM would say, no, no, the red stack, this is where we're going to help you? Well, so at IBM what we do is it's all about customer choice. And we've got, if you will, a blue stack. Our flash systems and our spectrum scale are often used in Watson environments, in big insights environments, and in other analytic workloads that we have. We also offer, with our power division, a pure power, which is a vertically integrated stack that includes compute, networking, IBM stored solutions, all in integrated stack with a hypervisor that they can be wheeled in, rack and stack and installed, you know, in a matter of hours versus a matter of days. So whether that be software integration with the hardware, such as we do with Watson and with big insights, or would that be more physical integration like we do with pure power? At IBM it really depends on your application workload and use case. And once we understand that application workload and use case we can say, look, the fully integrated way is the way to go and we could load the software on and that's right, or you know what, we can put together, even if you will, a heterogeneous configuration and that will work perfect. So it's really about understanding the number one thing that I think companies don't do is try to force feed. And what it really is, is what's your application, what's your SLA, what's the workload like, what do you need to see from that workload and what are you trying to drive, as you mentioned earlier, the business value. And once you know that, you know, which is the right recipe, if you will, the solid blue recipe or some sort of heterogeneous recipe. And you need to be flexible in that to meet the needs of the customer. That's the number one thing is meet the needs of that customer. Yeah, and I think IBM really needs to be confident. You guys have done a tremendous job around open source in terms of Java, in terms of Bluemix or what's going on with Cloud Foundry. I mean it is very much a belief that, like you said, choice, and then contributing back to communities as well. So it's not just package software, it's also open source software, open source communities driving a lot of things in that space as well that's looking at next generation flexibility, next generation opportunities. Well, and we're doing that across the board, as you know, at VMworld, we introduced the mainframe, the Z system, supporting VMware. At the same time, on our spectrum scale, we not only have scale out network attached storage, but we support OpenStack Swift, an open object store. And what you're going to see next week as well, I'm spilling the beans in the lirly, is support with Hadoop Connectors. And Hadoop is an open source database, if you will. So a lot of areas of IBM, server division, storage division, clearly the software groups, are all about supporting the Linux community, things like OpenStack. Java, Docker. Yeah, Docker, we do stuff with Docker. So IBM goes out of its way to offer ultimate flexibility, and as you know, we have a reputation, let's say in the Linux community, for example, as having more Linux programmers than almost any other company in the world, and contributing more code across all the various divisions into that, the various elements of the Linux community. So we believe it's all about choice and flexibility, and that's critical for us at IBM. All right, so Eric, you know, want to give you the last word. People walking away from Oracle OpenWorld 2015. What's your key takeaways, and the message IBM wants their customers to hear? Well, I think the key thing is, the world is changing dramatically. People can't think about hardware, they have to think about the data and how to get the most out of that data, and the storage underneath that needs to be secure and reliable, like it always did, but performance is becoming more and more important, which is why Flash is so critical. Things like our spectrum scale with scale out NAS, it offers parallel access across thousands of servers, giving you incredible analytics performance, and what's really important is how does storage allow you to derive value from the data, because people are swimming in these oceans of data, how the heck can they make sense of it? And that's what you got to look for from your storage is to help you make sense of these oceans of data, not just, oh yeah, I need some storage, that's reliable, or boy, I need a low cost per gig, because you know what, we have really low cost per gigabyte with our Flash, that's not the real value. The value is, how does the Flash take a report from 30 hours to two hours? That's the value of things like Flash systems. So customers need to think about their data first, not about what's my cost per gig, because the metrics are about data, and if something's a little bit more expensive, but report runs in two instead of 30 hours, that is money for the company, even if they spend a little more to get it. All right, well, Eric Herzog, Vice President of Product Marketing with IBM Storage, always a pleasure to catch up with you. We'll be back with lots more coverage here from Oracle Open World 2015. This is theCUBE, thanks for watching.