 Hi, I'm Peter Burris, and welcome to another WikibonCube conversation. Today I'm joined by Eric Herzog, who's a CMO and vice president of channels in IBM's storage group. Welcome, Eric. Peter, thank you very much. Really appreciate spending time with theCUBE. Absolutely, it's always great to have you here, Eric. And you know, one of the, it's interesting, when you come in, it's kind of let's focus on storage because that's what you do, but it's kind of interesting overall, the degree to which storage and business is now becoming more than just a thing that you have to have, but part of your overall business strategy, increasingly because of the role that digital business is playing. Well, earlier today, IBM made some pretty consequential announcements about how you intend to help customers draw those two together closely. Why don't you take us through them? So, first thing I think with the digital business, it's all about data, and the digital business is driven by data. Data always ends up on storage, and it's always managed by storage software. So, while it may be underneath the hood, if you will, it is the critical engine underneath that entire car. If you don't have the right engine and transmission, which you could argue storage and storage software is, then you can't have a truly digital business. True, so tell us, what did IBM do? So, what we did is we announced a number of technologies today, some of which were enhancing, some of which were brand new. So, for example, a lot of it was around our spectrum storage software family. We introduced a new software defined storage for NAS, Spectrum NAS. We introduced enhancements to our IBM CloudObject storage offering, also to our Spectrum virtualize, several enhancements to our modern data protection suite, which is Spectrum Protect and Spectrum Tech Plus were enhanced, and lastly, from an infrastructure perspective, we announced a first real product around an NVMe storage solution over an InfiniBand fabric, and what we're going to do for the rest of year around NVMe and how that impacts storage systems, which are, of course, a critical component in your digital data business. And you also announced some new terms and conditions, there are new ways of conceiving how you can get access to the storage capacity, storage plants you want, why don't you give us a little bit of feed inside and then? One of the things we've done is, we've already created a couple of years ago, the Spectrum Storage Suite, which has a whole raft of different products, file software, block software, backup, archive software. So we added the Spectrum Protect Plus offering into that suite, we also had a backup-only suite, which focuses just on modern data protection, we put it in there, and in both cases it's at no additional fees. So if you buy the suite, get Spectrum Protect Plus, if you buy the backup-only suite, so you're more focused on backup-only, again, no extra charge to the end user. The other thing we've done is we announced in Q4 a storage utility model. So think that you can buy storage the way you buy your power bill, or your water bill, or your gas bill, so it can go up and it can go down, we bill you quarterly. We added our IBM Cloud Object Storage on-premises solution to that set of products. We had an earlier set of products built around Flash, we announced in Q4 of last year, now we've added object storage as a way to consume in basically a utility offering model. So we talk a lot at Wikibon about the need for what we call the true private cloud approach, which is basically the idea that you want the cloud experience wherever your data requires. And it sounds like IBM is actually starting to accelerate the process by which it introduces many of these features, especially in the storage arena. You've bought in more stuff underneath the spectrum, family, you're starting to introduce some of those new, highly innovative technologies like NVME Overfabric, and you've also introduced an honest-to-goodness utility model allows people to have or to treat their storage capacity more like that cloud experience. Have I got that right? Absolutely, and we've done other things too. For example, as you know, from a cloud perspective, everyone is moving to containers, right? Our Spectrum Connect product offers free support for dockers and Kubernetes. So if we're going to create a private cloud and you're going to do that on your own or even hybrid cloud where you're sloughing some of it into your public cloud provider, bottom line is that docker support, that container support is what you need to create the true private cloud experience that Wikibon has been talking about for the last year and a half now. Well, let's talk about the Kubernetes and docker and the notion of containers as it's associated with storage. And I'm going to take it in two directions. First off, tell us a little bit about how it works in kind of developer-oriented terms, and then let's talk about what that's going to mean to the ecosystem and how people are going to think about buying storage going forward. So why don't we start with, how does this capability work? Sure, so the key thing we've done with the Spectrum Connect product is provide persistent storage capability to a container environment. As you know, containers just like VMs in the past can come up and come down very frequently, especially if you're in a DevOps environment. The whole point is they can spin them up quickly and take them down quickly. The problem is they don't allow for persistent storage. So our Spectrum Container product allows for the capability of doing persistent storage connected to a containerized environment. So the way that this would work is you'd have, you'd still have a server, you'd still have a machine with some compute that would be responsible for spinning the containers up and down, but you'd have a storage feature that would make sure that that storage, or the storage associated with that container, would persist. So therefore you could continue to do the container up and down in the server, while at the same time persisting the storage over an extended period of time. Right, so what that means is any of our customers who have our Spectrum Accelerate software defined storage for block, our Spectrum virtualized software defined storage for block, and the associated family of arrays that ship with that software embedded. Remember, for us, our software defined storage to be sold standalone is just a piece of software or embedded in our arrays, which for example with Spectrum virtualized means there's hundreds and hundreds of thousands of our software defined storage between the software only version and the array version. So for people who have those arrays, the container support is absolutely free. So if you've already bought the product and you're on our maintenance support, you just download the Spectrum Connect, boom, you're off to the races, you deploy your containers for your private cloud environment and you've got it right there. If you're a brand new customer you're going to buy, let's say for example, next week, you buy next week, you get the Spectrum virtualized, let's say for example on our stores, V7000 FL flash rate, because that software comes with it, and you could go download Spectrum Connect at no fee because you type in your customer, put in your serial number, boom, they can just download it and we don't charge anything for that. And now your storage guys and your developer guys are working a little bit more closely together as opposed to being at each other's throats. I'm saying what happened to the storage. Oh wait, I thought that was going to be, well no it's not persistent, this case it's persistent, they can take it up, they can take it down, they can do whatever they want and that container product is free so the IT guy doesn't go, oh now I got to pay more money because he doesn't. And then the guys on the DevOps side and on the deployment application side are, okay now I know the storage is, I don't have to worry about that as an issue anymore. The IT guys took care of that for me. So you get everybody working together, you get the persistence storage that's not, you know comes when you get a container environment you get the exact opposite, it's not persistent and now we've offered that. And again it's a no charge for the user so it's easy to deploy, easy to use and there's no fee. And so Eric the reason I ask the question is because it's the compounding of these little annoyances that make it difficult for companies to accelerate their entree into digital business and how they engage their customers differently. And so this is one of those examples where as we said data is the asset that distinguishes a digital business from a regular business competitor. What types of changes is this going to mean to the way the business thinks, the way the business buys and the way the business perceives storage? So I think the first thing is they need to realize that in the digital business data is the oil. It is the gold, it is the silver, it's the diamonds. It is the number one. It's the value. It is the value of your digital business. So you have to realize that the underlying infrastructure if it goes down, guess what? Your digital business is no longer up and running. So from that perspective, you need to have your underlying foundation from a storage perspective. In this case, think a storage system need highly, highly available, highly, highly reliable. And it needs to be incredibly fast because now you're doing everything from a digital business and so everything is pounding on your server and storage infrastructure. Not that it was in the traditional data center but if certain things need to be slow, it's okay. But now that you've gone true private cloud with a full digital business, it can't be slow. It has to be resilient and it has to be always available. And those are things that we've built into both our storage software layer, the Spectrum family and to all of our storage arrays, the StoreWise family, our DS family, our Flash system family, all are highly redundant, highly available and they're all Flash. And let me add two more things to that because I think it's pertinent to the direction that IBM is taking here. Because digital, or because data is not exactly like oil or not exactly like diamonds in the sense that the oil and diamonds and whatnot still follow the laws of scarcity. The value of data increases and I know you've made this point, as you use it more. So on the one hand, the storage has to provide the flexibility that developers can go after the same data at different times and in different ways but still have that data be persistent. And then related to that obviously is that you want to ensure that you're able to drive that throughput through the system as aggressively as possible without creating a whole bunch of additional administrative headaches. So if we pivot for a second to NVMe, what does that mean to introduce things like NVMe to those five things that we just talked about? Especially on the performance and the flexibility of having multiple applications and groups being able to go at the same data, perhaps through snapshots and copies. So a couple of things, from a software perspective that sits on top of all of our products, we've taken an approach of modern data protection. It's not just let's do an incremental backup like in the old days. So what we do today is we have basically incessant snapshotting, which is a full boat copy. What you can do is you can check those out with our Spectrum Content Data Manager, which we didn't announce anything new on that but we announced it last year. And with that, you can have an ending snapshots. The DevOps guys can grab a real piece of software, real piece of data. So when they're doing their development, they're not using a foe set and that foe set often can introduce more bugs. It doesn't get up as quickly. So now you got a real data, right? So they take the snapshot. By the way, it's self-service. They could check it out themselves. Now when you look at it from the IT guy perspective, guess what? There's a log of who's got what? So if there was a security issue, they could say, oh, Eric Herzog, you're the one that had that. So it looks like it leaked out from you. Even if it was inadvertent, the point is the DevOps guys could go in and grab from this new modern data protection paradigm that we have. At the same time, the IT guys can at least track what's going on. So it's the same thing. Then from an NVMe perspective, the key thing that NVMe adds is A, all of the existing infrastructures in Finneban fabric, fiber channel fabric and ethernet fabric will be supported. Over time, we're announcing today a Finneban fabric solution, about all of the arrays that you buy today. If you, for example, bought a flash system V9000 and you want to do NVMe over ethernet later in the year, software upgrade only, you buy the hardware now, you're done. Our A9000 flash systems, fiber channel connect, you buy the fiber channel now, you just upgrade the software a little bit later. So the key things within NVMe configurations, A, the box is already highly resistant, highly available. They resist failures. They're easy to fix if there is a hardware failure, for example, failed power supply. You know it's going to happen. The smart business has an extra power supply sitting on the shelf. He pulls it out, he swaps in and sends it back to IBM. And when it's under warranty, boom, we take care of it. So that's the resiliency and the availability aspect from a physical perspective. But with NVMe, you get a better performance, which means that the arrays can handle more workloads. So as you go to a truly digital business built around the private cloud that Wikibon has been talking about now for 18 months, as you go to that model, you want to get more and more apps pounding on the same storage, if you will. And with an NVMe fabric solution, NVMe, over time in the subsystem itself, all that gives you more apps can work on the same set of storage. Now you have, do I have enough capacity, which is a separate topic. But as far as can the array handle the workload with NVMe from a fabric perspective, NVMe in a storage subsystem, you can handle additional workloads on the same physical infrastructure, which saves you time, saves you money, and gives you the performance for all the workloads, not just for a few Nietzsche workloads and all the other ones have to be slow. So Eric, you're out spending a lot of time with customers. Tell us a little bit about how they see their environments changing as a consequence of these and other related announcements. Are developers going to be looking at storage more as a potential source of value? How are administrators dealing with this? And give us some examples, if you would. Sure, sure. So I think the key thing is with things like our content data manager, is we've got customers right now and they're able to check it out to all the test dev guys, which they couldn't do before. They're getting work done faster with real data. So the amount of bugs that come up with internal developers, just like commercial developers like IBM or any other software company, the Microsoft, the Oracle's, everybody has bugs. Well, guess what? In-house developers got the same bugs. But we help reduce that bug count. We make it easier for them to fix because they're working on a real data set, not a fake data set, right? The IT guys love it because the DevOps guys don't say, can you spin this up, spin this down? They do it on their own, right? Which accelerates them in doing their work and the IT guys aren't bothered for it. That one concern on security, guess what? You got that log saying who's got what. Right, right. Burris has this, Herzog has that. That's a big deal because the IT guys ultimately, if something leaks out or there's a security issue, they get the call from the chief legal officer, not the DevOps guy. So this way, everybody's happy. The DevOps guys are happy. The IT guys are happy. The IT guys can focus on not spinning up and spinning down for the dev guys. You can build it all yourself. Our copy data management and all of our storage software is API driven, REST APIs, integration with all of the object storage interfaces, including S3. So it's easier and easier for the IT guy to make the DevOps guys happy and give the DevOps guys self-service, which is, you know, self-service, one of the key attributes of the private cloud that Wikibon is talking about is self-service. So we can give that self-service through the software side. So I got one more question, Eric, as we think about kind of where this announcement is, most important to businesses that are trying to affect that type of transformation we're talking about. Is there one specific feature that is your conversation with customers, your conversations with the channel, since you also very, very close to the channel, that keeps popping to the top of the list of things to focus on as companies, as I said, try to figure out how to use data as an asset differently. Well, I think what the key thing from a storage guy perspective is one, interfacing with all the APIs, which we've done across our whole family. Okay? Second thing is automation, automation, automation. The DevOps guys like it in a smaller shop, there may be only one IT guy who has to take care of their entire infrastructure. So the fact that our Spectrum Protect Plus, for example, can do VMware Hyper-V backup, except it can be done by the VMware Hyper-V guy or a general IT guy, not a storage guy or a backup admin. In the enterprise, sure there's a backup admin in the big enterprises, but if you're a Herzog's barn grill, there is no backup admin. So that ease of use, that simplicity, that integration with common APIs and automating as much of the process as possible is critical as people go to the digital business based on private clouds. Excellent. Eric Herzog, CMO, Vice President of Channels, IBM Storage Group talking about a number of things were announced today as businesses try to marry their storage capability and the digital business strategy more closely together. Thanks for being here. Great, thank you very much. Once again, I'm Peter Burris. This has been a wiki bond or a CUBE conversation with Eric Herzog of IBM.