 Live from the Palace Hotel in San Francisco, it's The Cube at the HGST Press and Industry Analyst Briefing brought to you by headline sponsor HGST. Here are your hosts, Stu Miniman and Jeff Frick. Welcome back to the Sheridan Palace in downtown San Francisco. I'm Jeff Frick and you're watching The Cube. We're here at the HGST Industry Analyst and Press Day for a whole bunch of announcements and exciting news and I'm joining this segment with my co-host. Hey Jeff, I'm Stu Miniman with wikibond.org and joining us for this segment we have Dave Tang who's the VP of Corporate Development Strategy with HGST. Dave, thank you so much for joining us. Pleasure to be here, Stu. Alright, so a lot of announcements today. We've been digging through all of it with your whole management team, some of the industry watchers and the like and so, you know, you cover strategy. So, you know, HGST has gone through a bit of a maturation going from, you know, just doing this to doing many more things. Talk to us, you know, what's your involvement in the launch? Right. Well, overall the storage industry is changing. I think we've gone through a period where storage solutions for data centers were built exclusively by large system OEM companies. But with the growth, rapid growth in cloud data centers, that model has changed and evolved rapidly. So, strategically, we need to recognize that shift and the implications on the types of solutions, the scalability and flexibility and affordability of the solutions needed and look to how we leverage our expertise and knowledge on the fundamental building blocks of storage, whether it be SSD or hard drives, and create higher value capabilities for our customers. So as the leader of the strategy for the company, we're looking to understand what the next generation of solutions need to look like, what's important to cloud data centers as they grow and scale and add more services and how we fill the gaps, either organically through our own research and development organizations or how we look outside to fill those gaps. All right. So, a big piece of that, if we're to create building blocks, the announcement that you had today is what you call the active archiving platform. Yes. Which, there's a lot of pieces that made up that from my notes, that you've got the optimized hard disk drives, which are the helium, you've got tuned enclosures, you've got device affinity, and you've got to scale out software. I mean, wow, this isn't just take a component and do a little work on it. I mean, it sounds like this is something you guys have been working on for a while. So can you bring us inside and talk to how you guys created this new offer? Sure. So, the first step is always to get out of the box, and that's such a commonly used term, but for us, it really means if you're going to solve problems at scale, you really have to think about them differently. Solving problems at scale can't be done by just taking more of the same things that you have and gluing them all together, because inevitably you'll run into problems with manageability or other aspects of that type of scaling. So, what we did was we took a clean slate approach and really looked at what the objective was. So the objective was to create storage solutions that had the highest density, high scalability, capacity, and most importantly, affordability out there in the marketplace, because we fundamentally believe that data wasn't being stored because it wasn't affordable to store that data. So rather than take the typical approach where someone would take an enclosure that may store 48 drives or 60 drives and just start dropping drives in at the highest capacity possible, we asked ourselves, what would the system look like if we wanted to push the capacity of the drives to a higher level, a higher limit? And the reason why drives are potentially limited in the specifications they have is because to ensure the capacity and reliability at a certain point, we need to design the drives to work in any enclosure built in the industry, right? Different heat and power profiles, different vibrational characteristics. But if we design the enclosure, we know exactly where those drives are going, we can tune and optimize those drives to maximize the capacity. We get even more value out of them in terms of capacity and performance. So the type of approach of taking a clean slate approach and driving more value is really what got us to where we are. All right, so Dave, that's a great point, is general purpose devices have such a broad spectrum that they need to fit into. When we talk to Amazon, when they build an architecture, it needs to go in one data center, their own, and therefore they understand exactly what the characteristics is. So I guess the question I have is, when you build these enclosures, when this go in the cloud and in the enterprise, most enterprise data centers are limited on space and its power density is one of the big challenges that they have. So what's the appropriate places that this new solution can fit into? So these solutions are designed for any data center, whether it be cloud or enterprise, and the points that you made in terms of space and power are really what drove us to the design that we arrived at. So having the highest density solutions, being able to put over 10 petabytes in a single rack, is a breakthrough level of density. So we take up less floor space from the data center. And then because we're leveraging our unique helium technology, which consumes 40 to 50% lower power per terabyte within the storage device, we can lower the power consumption not only in terms of what is required to power the infrastructure, but the cooling cost of the data center, which we all know is significant. And then for new data centers that are being built, they don't have to pay the price for pulling high power infrastructure to each rack and pay the tax associated with that. So we're saving in the cost of the data center build out in a lot of ways. So can you share how much power per rack it does take just if I loaded this up to a full capacity? Well, we're not releasing specifications yet. Those will be available in early 2015. But our goal is to hit the absolute lowest power per terabyte being shipped in the industry. So a lot of your customers, are you finding they need this now because they've just hit the peak in terms of the power capacity that they can get into that particular facility? Or are they taking advantage of this new capacity you're freeing up in the form of footprint and power and using that now to capture that extra data that they couldn't really hold economically before? What are you kind of seeing as the actual business impact in the marketplace? Well, there's both. The great thing about the place that we're in in the industry is that data is growing at such a fast rate and an exponential rate that there are a lot of new data centers being designed and built right now. So in addition to the refreshes that are going on, we're looking at a lot of new data center build out projects with our partners and customers. And what they're really interested in is they know that the data is growing at an exponential level, but they don't want to increase their expenses at an exponential level to keep up with that. So they have to find something that breaks the model of the traditional storage solutions. And they're really excited about the innovations that we've developed. I'm curious, on a lot of the new data centers, a great quote I heard at VMworld the other day from a friend. He goes, you know, even turkeys can fly in the storm. You know, the market is so hot. There's just so much new capacity coming on. Are they using industry standard servers and stuff? Or are they looking to more specialized solutions that you guys can deliver on the storage on these kind of green field new data centers that are going in? Well, it depends whether you're talking about servers that are running the applications or servers that are part of the storage infrastructure because part of the software defined storage solutions is that they have software that needs to run on a process or on a server somewhere. So what we're finding is the vast majority of users are looking for complete solutions. So even though there are software capabilities in object storage and there are high capacity storage arrays that could be purchased and white box servers that can be purchased, the integration of those elements into a complete solution are really hard. So we've heard time and time again that customers and users are looking to deploy complete solutions as opposed to investing in that bring up because at the end of the day, cloud service providers want to develop new offerings for their customers to attract and retain their customers. They don't want to spend time integrating their infrastructure and keeping the infrastructure running. That should just be part of a utility that they rely on as opposed to necessarily a core competency that they have to invest a lot of money in. So Dave, the 10 terabyte drive breakthrough in the industry. Absolutely. Can you give us a little bit of insight? What had to happen? You pulled together SMR and the helium to be able to make this. Is the ecosystem ready for all of this? And what's it going to take to drive that forward? Sure. So SMR does take some adaptation from the ecosystem in that because of the shingled nature of the tracks, you need to be sensitive about how you're writing to the data. The drives like to see long sequential rights to it as opposed to a lot of random writing. And that can be adapted in mass, but the risk is that you get performance that is less predictable. So some adaptation has to take place either within the operating system or a file system or within the storage solution altogether. Part of the strategy that HGST has in delivering these complete high scale out solutions is that we can, in parallel with the drive development, we can develop the system level capabilities that take advantage of things like the 10 terabyte breakthrough capacity and deliver those to market and let users enjoy the benefits of that lower cost per terabyte in a more accelerated fashion. They don't have to wait for the rest of the ecosystem to come together, which oftentimes can take years. All right, so in your presentation earlier, you said that 75% of the new apps in the cloud are big data or analytics related. I'm curious, what are you doing to get ready for what sometimes called the cloud mobile or third platform applications? Right, well, precisely the elements that we talked about today. So number one is storing more data so that we can accommodate those big data intensive applications. But once that data is available, it needs to be processed or transformed quickly. And that's where our application acceleration solutions that Gus talked about, flash based solutions with clustering software and virtualization software come into play. So we're looking at providing a full portfolio of capabilities that address the diverse needs of the cloud that not only need big data capabilities, but big data analytics and processing capabilities. So back to the act of archiving, there's been a ton of new solutions in that marketplace. Everything from what Amazon does with Glacier through traditional storage players, releasing archiving as many different technologies out there. Can you explain to us how is active archiving different from, say, cold storage? Sure, so I think cold storage covers a pretty broad set of use cases. I guess the simplest way to think about it is that cold storage is storage that's used for data that's not hot. So it could be data that is past its create and modify phase, but it could go all the way down to storage that's used for pure archival purposes. So when we talk about active archive, we're talking about a subset of what's typically thought of as cold storage. And the active part is the key word in that in order to make use of vast amounts of data that are being stored, you need to get to it quickly, much faster than what you could do with a tape library or an optical library, you would need access times, certainly in the sub-second type of time frame. So that's a distinct difference of creating disk-based active archive solutions as opposed to some of the other either tape-based or hybrid solutions that are out there. The other piece that was talked about quite a bit this morning was the device affinity. And I wonder if you could drill down into that. It was described as really being able to use the hard knocks that you guys have already learned in the bumps and the bruises and save those from your customer by having a group of things that have been put together and you've kind of learned the hard lessons. Talk about how that's a business advantage and really what the benefit is to the customers. Sure, so for us it means understanding what the underlying technologies are in the drives and where we have opportunities to push the capabilities of the devices even further. And specific to what we were talking about with the active archive platform, it's ability to know how to design an enclosure that contains the drives so that we can squeeze more capacity out of the drives. So that's an example of device affinity. And it can come in the form of hardware, mechanical design, as well as firmware capabilities that get more value out of the sum of the parts than just the bare sum of the parts. So the value that we can pass on to our customers and their users is higher value. So looking to provide significantly lower costs per terabyte to them enables them to store more data, which enables them in turn to extract more value out of that data and that continues to fuel the cycle. Yeah, and you had also talked quite a bit about the percentage of the data in the digital universe that's going to be usable, correct me if I'm wrong, is growing. Yes, absolutely. We said from 17% to 22% last year to 37% in 2020. How much of that is the actual value of the data or the changing economics that now make the value of that data go above the break even threshold? So based on the research that's been done to date, that increase would appear to be based solely on the value of that data, the value that you can extract from that data, because as I mentioned, only 40% of that data out in 2020 is expected to be stored based on projections of capacity shipments. So I think the missing piece of the puzzle that the market needs to embrace or understand better is what we're working on, on fundamentally different scales of economies that allow you to store that remaining 60%. It's like the Google, right? They store everything, nothing ever gets deleted. So that's kind of the theory that you're. Right, but you have to be able to do it on an economical basis, right? So Dave, I want to talk about the storage administrator because you talked about how we're going to get eight times more data by the time we get to 2020, but we know headcounts pretty much flat. So is the manageability of this something that your customers, the cloud service providers or your OEMs are going to handle, or do you handle that piece of it all? Yeah, that's absolutely key. And I think in terms of providing a solution to the marketplace, we have to take a strong level of contribution there and ownership in that experience with the customer. Because as I said, our cloud customers want to focus on developing new services for their customers. They don't want to spend time managing the infrastructure. So the easier we can make it for them to deploy and manage the infrastructure, the more data they can afford to give, the more services they can create for their customers. And that serves the entire ecosystem. So we know that cloud guys do things a little bit differently. What do you see from the OEMs that, you know, buy your stuff and the enterprise end users? Are they understanding the cloud operational model and how to, you know, not need to touch all the gear, not need to hyper-optimize everything based on the application, but to build kind of self-managing infrastructure? Absolutely. I think we're seeing a big change in the industry right now where our OEM system customers as well as enterprise customers are realizing that the long qualification periods and rigorous testing and validation of systems is no longer really required in a cloud environment. It's really about gaining agility through the elasticity and the scalability of the hardware and software that they're using. So I think there's a mass movement towards these cloud environments. And in fact, a lot of enterprises have split IT organizations, right? They have those that manage the traditional enterprise data centers, and then they have others that work on projects that leverage the value of the cloud. And that gets better utilization and lower cost out of the infrastructure. But more importantly, for the people that are using that infrastructure within their company, the developers, that self-serve capability to get a hold of and provision computing resources makes their entire corporation a lot more nimble. All right, so for the active archiving platform, obviously there's a good fit for that in the cloud service providers. Should we expect to see OEMs of this from some of the traditional server and storage players? So it really depends on the vertical markets that the enterprise is playing. There are certain verticals that are very data intensive, right? Biotech and pharma and other healthcare players have to retain data for compliance reasons. So they're very focused on economical solutions for the storage of data. And there are other verticals as well. Oil and gas tends to be a typical one in research as well. So we do expect a lot of enterprise users in key verticals to be adopting this technology. Great, well Dave, thanks for stopping by. Exciting times. Yeah, very exciting times. Pleasure to be here. You guys are playing both sides. You're playing the Flash. You're playing the active archive, the high capacity drives. You guys are all over the place. Well, there's a tremendous amount of opportunity. It's great. So I'm Jeff Frick. You're watching theCUBE. We're at the HGST Press and Industry Analyst Event at the Sheridan Palace in downtown San Francisco. We'll be right back with our next segment after the short break.