 from Las Vegas, it's theCUBE. Covering IBM Think 2018, brought to you by IBM. Welcome back to IBM Think 2018, everybody. My name is Dave Vellante, and I'm with my co-host, Peter Verst. You're watching theCUBE, the leader in live tech coverage. This is day three of our wall-to-wall coverage of IBM Think, the inaugural Think conference. Good friend, Eric Herzog is here. He runs marketing for IBM Storage. They're kicking butt. You've been in three years, making a difference, looking great, new Hawaiian shirt. Welcome back, my friend. Good to see you. Thank you, thank you, always loved being on theCUBE. So, this is crazy. I mean, I miss Edge, I loved that show, but, you know, one stopped shopping. Well, a couple things. One, when you look at other shows in the tech industry, they tend to be for the whole company. So, we had a lot of small shows, and that was great, and it allowed focus, but the one thing it didn't do is every division, including storage, we have all kinds of IBM customers who are not IBM storage customers. So, this allows us to do some cross-pollination and go and talk to those IBM customers who are not IBM storage customers, which we can always do at a third-party show like a VM world or Oracle world, but, you know, those guys tend to have a show that's focused on every division they have, so it could be of real advantage for IBM to do it that way, give us more mass, and it also, you know, helps us spend money on third-party shows to go after a whole bunch of new prospects and new clients in other venues. You've attracted some good storage DNA. I think it's, you know, yourself and some others. Ed Walsh was on yesterday. He said, Joe Tucci made a comment years ago. He said, somebody asked him what's your biggest fear if IBM wakes up and figures it out in storage. Looks like you guys are figuring out four quarters of consistent growth, you know, redefining your portfolio towards software-defined. One of the things we've talked about a lot, and I know you've brought this, was the discipline around, you know, communicating, getting products to market faster cycles because people buy products and solutions, right? So you guys have really done a good job there, but what's your perspective on how you guys have been winning in the last year or so? Well, I think there's a couple things. One is pure accident, okay, which is not just us, is, you know, one of the leaders in the industry where I used to work and Ed used to work has clearly stubbed its toe and has lost its way, and that has benefited not only IBM, but actually even some of our other competitors have grown at the expense of, you know, EMC. And they're not doing as well as they used to do, and they've been cutting headcount, and you know, there's a big difference in the engineering spend of what EMC does versus what Michael Dell likes to spend on engineering. We have been continuing to invest. Sales resources, marketing resources, tech support resources in the field, technical resources from a development perspective. The other thing we did as Ed came back was rationalize the portfolio. Make sure that you don't have 27 products that overlap. You have one, and maybe you have a slight overlap with the product next to it, but you don't have to have three things that do the same thing, and quite honestly, IBM before Ed showed up, we did have that, so that's benefited us. And then I think the third thing is we've gone to a solution-oriented focus. So can we talk about as nerdy as tracks per sector and TPI and BPI, I mean, all the way down to the hard drive or to the flash layer? Sure we can. You know what, have you ever, you guys have been doing this forever. Ever met a CIO as a storage guy? No, no, CIOs don't care about storage. Exactly, so you got to talk- I've met a couple of ex-CIOs in the store. Yeah. So you really got to talk about applications, workloads, and use cases, how you solve their business problems. We've created a whole set of sales tools that we call the Conversations Bill to the IBM sales team and our business partners, which how to talk to a CIO, how to talk to a line of business owner, how to talk to the VP of software development in a global enterprise who doesn't care at all, and also to get people to understand that it's not, storage is a critical foundation for cloud, for AI, for other workloads, but if you talk latency right off the top, especially the CIO or the senior executives, it's like, what are you talking about? What you have to say is, we can make your cloud sing, we can make your cloud never go down. We can make sure that the response time on the web browser is in a second. Right, as you know, Google did that test about if you click and it takes more than two, two and a half seconds, they go away. Well, even if that's your own private cloud, guess what, they do the same thing. So you got to be able to show them how storage enables cloud and AI and other workloads. Let's talk about that for a second, because I was having a thought here, it's maybe my only interesting thought here, I think, being pretty much overwhelmed, but the thought that I had was, if you think about all the things that IBM is talking about, blockchain, analytics, cloud, go on down the list. None of them would have been possible if we were still working at 10, 20, 30 milliseconds of wait time on a disk head. The fundamental change that made all of this possible is the move from disk to flash. Right. Storage is the fundamental change in this industry that's made all of this possible. What do you think about that? So I would agree with that. There is no doubt, and that's part of the reason I said storage is a critical foundation for cloud or AI workloads. Whether you're talking not just pre-performance, but availability, reliability. So we have a public reference, MediCat, they deliver healthcare services as a service. So it's a software as a service model. Well, guess what? They provide patient records into hospitals and clinics that tend to be focused at the university level, like the University of California Health Center, right, for the students. Well, guess what? If not only does it need to be fast, if it's not available, then you can't get the healthcare records, can you? And while it's a cloud model, you have to be able to have that availability, characteristic, reliability. So storage is, again, that critical foundation. If you build a building in a major city and the foundation isn't very good, the building falls over. And storage is that critical foundation for any cloud, any AI, and even for the older workloads, like an SAP HANA or an Oracle workload, right? If again, if the storage is not resilient, oh, well you can't access the shipping database or the payroll database or the accounts receivable database because the storage is down. And then obviously, if it's not fast, it takes forever to get Dave Valente's bill, right? And that's a waste of time. So it's plumbing, but the plumbing's getting more intelligent, isn't it? Well, that's the other thing we've done is we are automating everything. We are viewing our software and we announced this that our arrays are going to be having an intelligent infrastructure software plane, if you will, that is going to help do diagnostics. For example, in one of the coming releases, if a customer allows access, if a power supply is going bad, we will tell them it's going bad and it'll automatically send a PO to IBM with a serial number of the address and say, please send me a new power supply before the power supply actually fails. But it also means they don't have to stock a power supply on their shelf, which means they have a higher cost of CapEx. And for a big shop, there's a bunch of power supplies, a bunch of flash modules, maybe hard drives that they're still dinosauric and how they behave, and they have those things and they buy them from us and our competitors. So imbuing it with intelligence, automating everything we can, automate. So automatically tiering data, moving data around from tier to tier, moving out to the cloud, what we do with the reuse of backup sets instead of doing it the old way of backup. I know you've got Sam Warner coming on later today and he'll talk about modern day protection, how that is revolutionizing what DevOps and other guys can do with their, essentially what we were called in the old days, backup data. Let's talk about your spectrum launch, spectrum NAS, give some plugs for that. What's the update there? So we announced on the 20th of February a whole set of changes regarding the spectrum family. We have things around spectrum protect with GDPR, spectrum tech plus as a service as well as some additional granularity features. And I know Sam Warner is going to come on later today. Spectrum NAS, software defined network attached storage. Okay, we're not going to sell any infrastructure with it. We have for big data analytics our spectrum scale, but thing of spectrum NAS is traditional network attached storage workloads, home directories, things like that, small file servers, where spectrum scale as one of our public references and they were here actually at Edge a couple years ago. One of the largest banks of the world, their entire fraud detection system is based on spectrum scale. That's not what you would screw spectrum NAS for. So, and it's often common as you know in the file world to have sort of a traditional file system and then a big one that does big data analytics and AI and is very focused on that. And so that's what we've done. Spectrum NAS is software only, software defined, rounds out our block, now it gives us traditional file, we had scale out file already and IBM Cloud Object Storage is also software defined. Well, how about the get put world? What's happening there? I mean, we've been waiting for it to explode. So, the get put world is all about NVMe. NVMe, new storage protocol, as you know it's been SCSI forever, SCSI and or SATA and it's been that way for years and years and years and years, but now you got flash. As Peter pointed out, spinning disk is really slow. Flash is really fast and the protocol of SCSI was not keeping up with the performance. So, NVMe is coming out. We announced an NVMe over Infiniband fabric solution. We announced that we will be adding a fiber channel, NVMe fabric based and also an Ethernet. Those will come and one of the key things we're doing is our hardware, our infrastructure is already to go. So, all you have to do is a non-disruptive software upgrade and for anyone who's bought today it'll be free. So, you can start off with fiber channel or Ethernet fabrics today or Infiniband fabric now and we can ship but on the Ethernet and fiber channel side they buy the array today and then later this year in the second half, software upgrade and then I have NVMe over fiber channel or NVMe over Ethernet. Explain why NVMe and NVMe over fabric is so important generally but in particular for this sort of new class of applications that's emerging. Well, the key thing with the new class of applications is their incredibly performance and latency sensitive. So, we're trying to do real artificial intelligence and you're trying to, for example, I just did a presentation and one of our partners, Mark III, has created a manufacturing system using AI and Watson. So, you use cameras all over which has been common but it actually will learn. So, it'll tell you whether cans are bad. Another one of our customers is in the healthcare space and they're working on a genomic process for breast cancer along with radiology and they've collected over 20 million radiological samples of breast cancer analysis. So, guess what? How are you going to sort through that? Are you going to sort through 20 million images? Well, guess what? AI can do that, narrow it down and say, well, it's this type of breast cancer, that type of breast cancer and then the doctor can decide what to do about it and that's all empowered by AI and that requires incredible performance which is what NVMe delivers. Again, that underlying foundation of AI in this case going from flash on with SCSI, flash to NVE, increasing the power that AI can deliver because of its storage foundation. But even those are human time transactions. What about when we start taking the output of that AI and put it directly into operational transactions that have to run like a bat out of hell? Which is where NVMe will come in as well. You cannot have the performance that we've had these last, you know, almost 30 years with SCSI and even slower when you talk about SATA, that just not going to cut it with flash. And by the way, you know there's going to be things beyond flash that will be faster than flash. So flash two, flash three, it's just the way it was with the hard drive world, right? It was 2400 RPM, then 36, then 54, then 72, then 10K, then 15.5. More size, more speed, lower energy. Which is what NVMe will help you do and you could do it as a fabric infrastructure or you can do it in the array itself. If you roll inbox and out-of-box connectivity with NVMe, increasing the performance within your array and increasing the performance outside of the array as you go out to your hosts and out into your switching infrastructure. So I'd love and thank kids, too many people to count I've been joking all week. It's 30,000, 40,000, we're still tallying up. I'm going to miss edge for sure. I'm going to miss the updates in the late spring. But so let's get them now. What can we expect? You know, what are you trying to accomplish in the next six to nine months? What should we be looking for without giving it any confidential information? Well, we've already publicly announced that we'll be flushing out NVMe across the board. So we already publicly announced that, that will be a big to do. The other thing we're looking at is continuing to imbue what we do with additional solution sets. So that's something we have a wide set of software. For example, we publicly announced this week that the VersaStack All Flash Array will be available with IBM Cloud Private to the Cisco Validated Design in May. So again, in this case taking a very powerful system, the VersaStack All Flash, which already delivers ROI and TCO, but still is, if you will, a box. Now that box is a converged box with compute, with switching, with All Flash Array, and with a virtual environment. But now we're putting, again, as a bundle, IBM Cloud Private on there. So you'll see more and more of those types of solutions, both with the rest of IBM, but also from third parties. So if that offers the right solution set to cut CAPEX, OPEX, automate processes, and again, for the cloud workloads, AI workloads, and any workloads, storage is that foundation, the critical foundation. So we will make sure that we'll have solutions wrapped around that throughout the rest of this year. So it's great to see the performance in the storage division. You're great people. We're not even, we're undercounting it. We're not even counting all the cloud storage. That goes, you know, accounts somewhere else. You guys are doing a great job. You know, best of luck. Really keep it up, Eric. Thanks very much for coming back on theCUBE. Great, thank you very much. Thanks again, Peter. We appreciate it. All right, keep it right there, buddy. We'll be back with our next segment. Right after this short break, you're watching theCUBE live from, think, 2018.