 From the SiliconANGLE Media office in Boston, Massachusetts, it's theCUBE. Now, here's your host, Dave Vellante. Hello everyone and welcome to this exclusive CUBE conversation. Here's the setup. The storage industry has been drowning in complexity for years. Companies like Pure Storage and Nutanix, you know, they reached escape velocity last decade primarily because they really understood how to deliver great products that were simpler to use. But as we enter the 2020s, virtually every player in the storage business is trying to simplify its portfolio. And the mandate is coming from customers that are under huge pressure to operationalize and bring to market their major digital initiatives. They simply can't spend time managing infrastructure that the way they used to. They have to reallocate resources up the stack, so to speak, toward more strategic efforts. Now, as you know, post the acquisition of EMC by Dell, we have followed closely and been reporting on their efforts to manage the simplification of the storage portfolio under the leadership of Jeff Clark. IBM is one of those leading companies along with Dell EMC, NetApp and HPE that are under tremendous pressure to continue to simplify their respective portfolios. IBM as a company has declared the dawn of a new era. They call it chapter two of digital and AI, which the company claims is all about scaling and moving from experimentation to transformation. Chapter two, I will tell you one questionably is not about humans managing complex storage infrastructure. Under the leadership of General Manager Ed Walsh, the company's storage division has aligned with this chapter two vision, and theCUBE has been able to secure an exclusive interview with Ed, who joins me today. Great to see you, my friend. Thank you very much for having me. So, you're very welcome. I mean, you heard my narrative. How did we get here? How did the industry get so complex? I liked the way you kicked it off because I think you nailed it, but it's just how the storage industry always been. And there's a reason for it 20 years ago, but it's almost, it's run its course. And I can tell you what we're announcing, but everyone, there's always a difference between high-end solution sets and low-end solution sets. In fact, they're different. There's custom silicon on the high end. So think about EMC Symmetrics in the day. It was the ultimate custom hardware and software combination. And then the low-end storage, well, it didn't have any of that. And then there's a mid-tier. But we actually, everything's based upon, dude. So the thing about, you know, you need the right availability, the right price ports, feature function, availability features that made sense that you had to have that unique thing. So what's happened is we're all doing, you know, sustaining innovation. So we're all coming out the next high-end array for you. EMC's, the next one is the hashtag next generation storage, right, or mid-range. So they're going to redo their mid-range and then low-end, but they never come together. But this is where the complexity, you're nailing it. So no one's a high-end or low-end shop. They basically use it all. But what they're having to do is they have to manage and understand each one of those platforms, how to maintain it, it's kind of specialized, how to report on it, how to automate it. Each, the automation requirements are different, but then different APIs to actually automate it. Now the minute you want to say, okay, now help me modernize that and bring me to a hybrid multicloud, now you're doing a kind of complex thing over multiple ways, and it gets different platforms, which are all completely different. So, and now the key thing is in the past, it made sense to have high-end silicon with high-end software and it made sense and different low-end. And basically, because of some of the innovation we've driven, no longer do you have to do that. There's one platform that allows you to have one platform meet those different requirements and dramatically simplify what you're doing for enterprises. So, we're going to talk a little bit more about what you guys are announcing, but how do you know when you get there to this land of simple? Yeah, one is hard to get there. We can talk about that too, but it's when clients, so we just had a call this morning with our board of advisor for storage, our division, and they're kind of the bigs of the bigs, more on the high-end side, but just so you know the sample size, but literally in the discussion of what we're talking about, the platform, the simplification, how do you get to hybrid cloud, what we're going to do on the cyber instant response type of capability, cyber resiliency, and literally in the call, they're already emailing their team saying we need to do something more strategic, we need to do that, we need to look at this holistically. They love the simplicity. Everything we just went through there, they can't do anymore, especially in chapter two, it's about modernizing your existing mission critical enterprises and then putting them in a context of hybrid multi-cloud. That's hard. You can't do it with all these different platforms. So they're looking for, let me spend less, you said, forget my team to do upstack things, they definitely don't want to be managing different disparate storage organizations. They want to move forward and use that, free up a resource to do other things. So when I see big companies just literally jumping at it and giving example, we've talked about the cyber resiliency thing. I've had four of those this week, that's exactly what we need to have done. So it's just, I haven't had a conversation yet that clients aren't actually really excited about this and it's actually pretty straightforward. So I'll give you a benefit of the doubt. Again, we'll get there, but assuming you're there, why do you think it took so long? I mean, you kind of mentioned it's hard, but yeah. So transformations aren't never easy and typically whoever is a transformation agent gets shot in the back of the head, right? So it's really hard to get teams to do something different. So imagine every platform, now EMC has nine now, right? So it is through acquisition or others. You have VPs, you have VP of development, offering and maybe sales, but then you have whole teams or you have founders you've acquired, right? So you have real people that are, they love their platform and there's no way they're going to give it up. They always come out with a next generation and how it's going to solve all ills, but just it's a people transformation. How do you get, you know, we're going to take three and say, hey, it's one platform. Now, to do that as an operational transformation challenge, it's actually driving the strategy. You don't do it in a matter of a week. There's development to make sure that you can actually meet all the different use cases that will take you literally years to do and have a new platform, but I think it's just hard to do. Now, anyone that's going to do that, let's say, you know, EMC or HP wants to do it, they're going to have to do the same thing we did, which is going to take them years of development, but also it's meaning, I'm managing the transition and the people involved or the founders you've acquired, or it just, it's amazing. In fact, that's the most wonderful part of my job. It's dealing with people, but it can frustrate you. So we've seen this over the years where I mean, look at NetApp with the waffle. It was one size fits all for years, but they just couldn't cover all markets and then they're a face with TAM expansion. Of course, now the portfolio expands. Do you think the same? And they have three and, you know, and David Scott at HPE used to or HP at the time used to talk about how complex EMC's portfolio was and you see HP has to expand its portfolio. Do you think, do you think Pure will have to face the same sort of? We are seeing Pure with three, right? And then that's without the file. So I'm just talking about what we do for physical, virtual and container workloads in cloud, right? So if you start going into what we're going to do at scale up object, we all have our own there too. So, and I'm not even counting the three to get to that. So you see Pure doing the exact same thing because they're trying to expand their TAM and you have to do some basic innovation to have a platform actually meet the requirements of the high-end requirements, the mid-range and the entry-level requirements. It's not just saying I'm going to have one. You actually have to do a lot of development to do it. All right, let's get to the news. What are you guys announcing? So basically, we're announcing a brand new, we're a dramatic simplification of our distributed storage. So everything for non-Z. So if you're doing physical boxes, bare metal, Linux, you're doing virtual environments, VM or Hyper-V, or if you're doing container workloads or into the cloud, our platforms are now one. It's going to be one software, one API to manage, but we're going to actually, we're going to do simplification without compromise. We're going to give you what you need. You're going to need an entry-level packaging, mid-range and high-end, but it's going to be one software, allows you to meet every single price requirement and functionality. And we're going to give you some, what I'll say is surprises on the upside for what we're bringing out to you because we believe in value innovation. I think we can up the value we're bringing to our clients, but also dramatically take out the cost of complexity. But one thing we're getting rid of is saying the need, the requirement to have a different hardware software platform for high-end, mid-range and low-end. It's one hardware and software platform that gets you across all of those, and that's where you get a dramatic simplification. So same OS. Same OS? So normally what you would do, you would optimize the code for the high-end, mid-range and low-end. Why are you able to address all three with one OS? How are you able to do that? It took us three and a half years, so I'll talk about a couple of innovation pieces. So on the high-end you have customized silicon. We do, everyone does. So we had a Texas Memory Systems acquisition. It was a flash drawer to you, about 375 terabytes, uncompressed to you do. Pretty big chunky, you had to buy in big chunks. So it's on the high-end. That was the unit of granularity, right? But it gave you great value, but also you had great performance, latency better than you get in NVMe today before NVMe, but you get inline compression, encryption. So it was wonderful, but it was really ultra high-end. What we did is we took that great custom silicon and we actually made it onto what looks like a custom, or it should be a standard NVMe SSD. So you take a Samsung NVMe or WD and you compare it to what we would call our flash core module, they look the same and they go in interchangeably into the same NVMe standard slot. But what's in there is the same silicon that was on this ultra high-end box. So we can give the high-end exactly what we gave them before, ultra low latency, better than NVMe, but also you can get inline compression, Ddupe and all the ware leveling and all the stuff you expect in the custom silicon level. But we can take the same NVMe drive and we can put it into our lowest-end model, an average sale price, $15,000. Allows you to literally not no compromise on the high-end, but have unbelievable surprises on the mid-range and low-end, where now we can get the latencies and the performance and all those benefits on, to be honest, a much lower box. Same functionality, same functionality, so you lose nothing. Now that took a lot of work, that wasn't easy. You talk about people, there was roadmaps that had to be changed and we had to know that we're going to do that and stick to our guns, but that would be one. Other things is you're going to get some things on the upside that you're not expecting, right? So because of this custom silicon, I'm going to have unique price performance, but also cost advantages. So I'm going to have best price performance or density across the whole prog line, but also I'm going to do things like on the high-end, you're used to unbelievable operational resiliency, right? So two-site, three-site, hyper-swap, you know? Now with two boxes that act like one, have a whole outage or a site outage and you don't literally miss a transaction or multi-sites. So, but we're going to be able to do that on the low-end and the mid-range as well. Cyber resiliency is a big deal. So I talked about operational resiliency. It's very different coming back when it's cyber, but cyber instant response becomes key so we're going to give you special capabilities there which are not available for anyone in the industry, but is cyber instant response only a high-end thing or is it a low-end thing? No, it's across everywhere, right? So I think we're going to shock on the upside, but a lot of it was the development to make sure the code stack, but also the hardware, we can really say no compromise, you want an entry level, I'm going to meet anyone at that moment. In fact, because of features of it, I'm able to compete in an unfair level against everyone on the low-end. Same thing, mid-range and high-end, but you're not losing anything because you're losing the Cusm of Silicon. So let's come back to the cyber piece. What exactly is that? All right, so listen, this is not for data breaches, where you have to put in, so if a data breach happens, they steal your database or they steal your customer names, you have to report to the press, you have to let people know, but it's typically they don't call the storage guy and say, hey, solve it. It was stolen at a different level. Now the one thing is that doesn't hit the media, but it happens all the time, actually more frequently, and it definitely gets called down to the operation team or the storage team is for cyber or malicious code. They've locked up your system. Now they didn't steal data, so it's not something you have to report. So what happens is, call comes down and you don't know when they got you. So it's an iterative process. You have to literally find the box, bring up, maybe it's Wednesday. Bring it up, give it to the application group, no, it's there, bring up Tuesday. It's an iterative process. It's like drilling for oil a hundred years ago. Nope, drill another hole. So what happens is, if it's cyber, without the right tools, you use your backup, but one of my board advisors literally made your bank, I had four of those, I'll give you one. It took me 33 hours to bring back a box. It was a large database, 30 terabytes, 33 hours. Now why is you backup? Why didn't you use his primary storage against DR copies and everything? Well, they didn't have the right tool sets. So what we're able to do is, tape is great because it's air gap, but it takes time to restore and come back when you're running. The modern data protection like we have, or Veeam or Cohesity allows recovery being faster because you're mounting backup copies faster. But the fastest is your primary snapshots and your replicated DR snapshots. And if you can leverage those, but the reason people don't leverage it, and we came upon this almost accidentally, we're seeing what our services brother and IBM doing from IBM SOs or G outsourcing or GTS, when they did have a hit. And what they want to do is bring up your snapshots, but if you bring up a snapshot and you're not really careful, you start crashing production workloads because it looks like the VM that just came up. So you need to have, and we're providing the software, allows you to visualize what your recover points are, allow you to orchestrate bringing up environments, but more importantly, orchestrate it into a fenced network environment. So it's not going to step on production workloads and addresses, but allows you to do that and provide a URL to the different business users that they can come and say, yep, it's there or it's not. So even if you don't use a software before this incident, it gives you visibility, orchestration, and the more importantly, a safe fence network, a sandbox to bring these up quickly and check it out in an easy to promote production. So that's your safe zone. Safe zone, and it's just not there. So you start bringing up snapshots, and now it's not like a DR case where you're just bringing things up. It's like, you have to be really smart because you're bringing it up and checking it out. So without that, they don't want to trust to use a snapshot. So they just don't use primary storage. With it, it becomes the first thing you do because you hope you got it within a week or we can have your snapshots. If it's in the environment for nine days, now you're going to tape. Now, if you do this, if you put the software in place before an incident, now you get more values. You could do orchestrated DR testing. Because we're doing this orchestrated bring up application sets, it's not AVM, it's a sets of VMs, fence network, bring it up, does it work? You can use it for test dev data. You can use it for automatic DR. But even if you don't set it up, we're going to make it available so you can actually come back from these cyber incidents much faster. And this is the capability I get on primary storage because everybody's targeting the backup corpus for ransomware and things like that nature. This is primary storage. And we do it in our backups. So our backup allows you to do the exact same thing and do the bootable copies, right? And so if you have our backup product, you could already do this on primary. But what we're saying is, regardless of who you're using, because you're still sent, you need to do backup. You need to air get your backup because like WannaCry was in environments for 90 days. You know, your snapshots only for a week or two. So the fact matters, you need it. But in this case, if you're using the other guys, you can also, we're going to give it just for this tool set. How does immutability, does it factor? I know like, for instance, AWS and Reinvent, I think announced an immutability capability. I think IBM may have that because of the acquisition that you guys made years ago is I think clever safe and fundamental to their architecture. Is that a way to combat ransomware? So that's helpful. Immutability allows you not to change it. So ransomware and malware typically is either encrypting or deleting things. Encrypting is what they do, but they have the key. So the fact of matters are deleting things. So if it's immutable, you can't change it. Now, if you don't have the right controls, you can delete it, but they can't change it. They can't encrypt it on you. So that becomes critical. So what you're looking for is we do, like for instance, all of our, this flash system allows you to do these snapshots, local remote, that allow you to have and go to immutable copies, either in Amazon, we support that, or locally in our object storage or in IBM's cloud does allow us to do that. So the different platforms have this immutability that our software allows you to integrate with. So I think immutability is kind of critical. How about consumption models? The way in which you're packaging and pricing. People want to, the cloud has sort of changed the way we think about this. How have you responded to that? So you hit upon our chapter two, we, IBM actually resonates with the clients. It's chapter one, we're doing some lift and shift and we're doing some new use cases in the cloud and they had some challenges, but it worked in general. But we're seeing the next phase two is really now looking at the 80% of your key workloads, your mission critical workloads, and basically how you transfer those in. So as you're, basically, if you look at your chapter two, but you're going to do the modernization and you might move those into the cloud. So if you're going to move into the cloud, they would say, well, I like to modernize my storage free my team up because it's simple. I don't have to do a lot of things, but you need to simplify so you can now, modernize so you can transform. But I'm going to be in the cloud in 18 months. So I don't want to modernize my storage. So what we have is, of course we have, you can buy things, you can lease things, we have a utility model that's great to three to five years, but we have now a subscription model, which think of just cloud pricing. No launcher commitments, use what you use up and down. And if it goes to zero, Carlos will pick it up and there's no expense to you. So no long-term commitments in return. So in 14 months, I've done my modernization, you help me free up my team, let me go, and then we'll come up and pick it up and your bill stop that day. Cancel anytime. Cancel anytime. Do you expect people to take advantage of that? I mean, is a ton of demand at this point in time? I think everyone's on their own cloud journey. So we talk a lot about meeting the client where they are at, right? So how do I meet them where they're at? And everyone's on their own journey, right? So a lot of people are saying, hey, why would I do anything here? I need to get there. But if they can modernize and simplify what they're doing, and again, these are your mission critical. We're not talking, these are how you're running your business. If you can make it work better in the meantime and then modernize it, get it to containers, get it to a new platform, that makes all the sense in the world. And because if we can give them a flexible way to say, it's cheaper than using cloud storage, like an Amazon or IBM cloud, but you can use it on-prem, free you up. And then at any time, just return it. That's a big value that people say, you know what, you're right. I'm going to go do that. You're able to give me cloud-based pricing down to zero when I'm done with it. And now I can use that to free up my team. That's the value equation. I don't think it's for everyone, but I think for a segment of the market, I think it's critical. And I think IBM's kind of perfectly positioned to do it with a balance sheet to help clients out. So how do you feel about this? I mean, obviously you put a lot of work into it. You seem pretty excited. You feel as though this is going to sort of help re-energize your business, your customer base. And how do you think competitors are going to respond? Good question. So I think simplification, especially if you get, when you talk about value equation, I think I can add more value to you, Mr. Customer. I can bring things you're not expecting, right? We'll get to the cyber in a second. That'll be one of the things they're not expecting and reduce the cost of complexity. So we've already done this a couple of times. So we did it with our mainframe storage launch in the fall. If bar none, the best box for that workload, lowest latency, most integration, encryption, pervasive encryption, and Christmas flights. But also we took it from nine variants to two, because we could. And people, why did you need all those? Well, there was reason for it in the past, but no longer. We also got rid of all the hard disk drives. We also add a little non-volatile cache and allowed you to get rid of all these battery backups, all these custom things that you used to have in this high-end box. And now it's dramatically simple, better. By the way, no one asked, hey, where did my other seven variants went? It was simpler, it was better, faster, but then it was the best launch we've had in the history of the Procline. So I think we can add better value and simplify for our clients. So that's what we do. You asked about how people respond. Listen, they're going to have to go through the same thing we did, right? A Procline has people behind it, and it's really hard, right? Or a founder behind it, right? You mentioned a couple of their acquiring companies. I think they're going to have to go through this. It's a transformational journey that they're going to have to go through. It's not as simple as doing a PowerPoint. I couldn't come to you and say, I can simplify without compromise. I can help you on the low-end, mid-range, high-end with the same platform, unless I did a lot of fundamental design work to make sure I could do that. Flash-core modules being one of them, right? So I think it's going to be hard. It'll be interesting. Well, they're going to have to go through the same thing I did. How about that? Usually, when you make a major release like this, you're able to claim Top Gun, at least for a while, and things like latency and bandwidth and IOPS and performance. Are you able to make that claim? So in fact, you'll see it, basically you saw it in the launch day, but basically you saw the latency, which is one, because we're bringing a custom silicon down, our latency, I'll give you a pure bragging on the websites, their lowest latency is 70 microseconds, which by the way is pretty, it's going to be 150 microseconds, pretty good, right? It's good bragging rates. We're at 70 microseconds, but that's on their X90 using storage class memory. So literally we're 2X faster than on the latency, how fast you can respond to something. But we can do it not only on our high-end box, but we can also do it on our average sale price, $15,000 box, because I'm bringing that silicon up and down. So we can do the latency. Now EMC, the highest and power max box rate though, two big chassis put together, that can do 100 microseconds. Again, still we're 70 microseconds, so we're 30% faster, and that's the epitomize of the high-end custom silicon and software. So latency, we got it, IOPS, right? So look at the biggest, baddest two boxes of EMC, they'll do 15 million IOPS on their website. We'll do 18 million, but instead of 2X, it's 8U, I mean it is 12X better IOPS for rack space if you want to look at that way. Throughput, which if you're going to do, it's all about building smart businesses, it's journey to cloud and building smart businesses. Everyone's trying to do this. Throughput and analytics becomes everything. And you're going to do analytics and everything. Your DBAs are going to run analytics. So throughput matters. Ours is for every one of our boxes that you can kind of add up and cluster out. It's 45 gigabytes a second. Pure for instance, their bragging rights is 18 and they can't cluster to get any more. So what we will do is on any of the, and most of them are high-end, but I'll say I can do the same thing up and down my line because we're on bringing the custom silicon. So on bragging rights, and that's just a kind of website, big bragging rights, I think we got it cold. And then if you look at price performance and just overall price per capacity, we're in line to be the most cost-effective across everything. The up and down the line is very interesting. It's kind of, it's unique. And then you mentioned resiliency. I'll tell you, that's the hottest thing, right? So you mentioned the cyber incident response. So that is something that's just, we did on the mainframe. So we did the last mainframe cycle. We allow you to do the same thing. And it literally drove all the demand for the product sets. It's already the number one thing people want to talk about because it becomes a, you're right. I needed that this week. I needed it last week. So I think that's going to really drive demand. What worries you? Well, on this launch, not much. I think it's how fast and far we can get this message out. Right, so. Okay, so execution, obviously. So you feel pretty confident about that. And then, yeah, getting the word out, letting people know. Well, congratulations Ed. No, thank you very much. Appreciate it. Appreciate it. Coming in. And thank you for watching everybody. This is Dave Vellante for theCUBE. We'll see you next time.