 All right, we're back. This is Dave Vellante and this is SiliconANGLE.TV's continuous coverage of EMC World. We are live in Las Vegas and we're bringing you the hall-to-wall end-to-end coverage, Monday, Tuesday, Wednesday, all day. You've been a great audience. We really appreciate your feedback. You can tweet me at atDVellante, at Furrier is John Furrier. And we're here with Steven Manley, who's the CTO of the EMC BRS division. This is the future of data protection, spotlight here at EMC World. And we're also joined by David Floyer, who is Wikibon's CTO. Gentlemen, welcome. Thanks for having me. Thanks very much. Yeah, this is Steven. You're a three-time now, a CUBE alum. I'm going to tell you, I think that's going to get the hang of us. Yeah, we first met last year at VMworld, talking about a lot of the changes that virtualization had brought, particularly around backup. And we talked about an Oracle Open World a few weeks later, some R-Man specificity as a little different world in Oracle. And then some time has lapsed. At that time, we talked about concepts, new concepts that you started to talk today about in your keynote. I heard you had a great keynote. Unfortunately, we were broadcasting, so I could see you. I couldn't hear you. So why don't you bring us up to date on what's new, what's changed since we last talked, and maybe give us some highlights of your keynote. So I think if you're really to hit on what's changed since we last talked, a lot of it's remained the same, right? The backup pressures remain the same. I've got increasingly painful backup windows, recovery windows. I need to find techniques that are going to help me reach those windows. A lot of that's remained the same. Disk continues to expand, I think, faster than most people expected. And we're seeing more and more users who want visibility and control under their backups, like we talked about at Oracle. DBA is wanting to run their own backup. So a lot of that's remained constant. I think two of the big changes or accelerations that we've seen, one is a lot more of people, I'd say, not just displacing tape with disk, but really replacing tape with disk and going for a fully tapeless infrastructure. And I think the second one is more and more people really trying to get control over their backup infrastructure. So I'd say trend-wise, those are two of the biggest shifts is that we've seen a real acceleration in those two areas. So tapeless. So I know David's going to challenge you on this, so I'll set it up. Tapeless for backup, correct? Tapeless for backup. So is the tape still around for archive and maybe disaster recovery? Yeah, though even then, we've seen, I'd say, an increasing shift, especially in the disaster recovery space, I think we've seen an increasing shift away from tape. And on the archive space, I think there was a wave where archival platforms like the Centera came out and took a chunk out of tape. I think we're seeing a second wave in terms of archive with things like the data domain extended retention. So I think more aggressive is removing tape for backup, but you're starting to see it even in the disaster recovery in the archive space as well. You know, I showed this slide and the setup, and David, Floy, I don't know if you could see it, but so this is IDC data on the purpose built backup appliance market. This is terabytes in this axis and revenue in that axis. Everything's up to the right. This little line here is the approximate size of the tape market. Now, I was astounded because I assumed that purpose built backup appliances could never be bigger than the TAM for tape. That set the ceiling. Didn't work out that way. Why is that? I think a lot of it has to do with the pain and backup. What drives the backup industry, first and foremost, is I can't get my backups done. I can't recover my data. And that gets in the way of other business processes. It gets in the way of virtualization. It gets in the way of application consolidation and refreshment. It gets in the way of what people are trying to do with their storage infrastructure. And, you know, if that's your main backup pain, and the only way you can solve it is to start to move to disk, and the disk market's going to surpass the tape market. Because in the end, the tape market was just a vehicle, a medium for how you did backup and protection. The protection is the core of what people are really trying to do with their data. Well, the irony is people, you know, started doing the purpose built backup appliance because they thought, you know, deduplication started to change the economics. Hey, we can save a bunch of money. It never got cheaper than tape, but it got close enough and then just overwhelmed the market for backup, didn't it? Yeah, I think if you look at a lot of the environments, if you look at it end to end, we've got customers who are showing, and I think you're talking to one soon, after me, Loomis, that showed an ROI on their disk investment that showed that they were saving money versus their tape infrastructure in a fairly rapid period of time. So, you know, when you take into account all the factors that go into to backup, it's not just the hard cost, but also the soft cost as well. You know, disk appliances can actually, for backup, show a serious ROI, in some cases, six months or less. So handling the media, the processes around it, the labor intensity, all of that, when you take a full TCO, you can make that business case very, very well. Now, David Floyd, you listened to Stephen Manley's keynote discussion and you said, I think, you know, roughly quoting, I'll paraphrase, that was the best and most cogent presentation I'd heard on backup in a long time. What was it that you liked so much about that talk and then we can get into it with Stephen? In particular from BRS, yes. What I thought you tackled for the first time was the addressing fully of the emerging technologies from Oracle, from VMware, with CBT, from Snapshots, also also Snapshots, so software and hardware, as effectively the only way of being able to solve the backup window issue and to be able to give specificity, application consistency to back to the applications and drive it as a service. And the analogy you started to give or you half gave was a time machine, which we all know and maybe love sometimes. As much as you can have any backup product, right? So is that a shift in the way that you're going to market? You emphasise Avamar a lot. You obviously have got a very rich investment with things like Network, all the other products that you have. Is that a shift towards that model and are you going to adapt those products to reflect that model? I think there's a definite shift in the model. I think that old world backup of, I install a client and I set a schedule and I read all the data and I put it in a proprietary format and I send it across multiple networks through multiple devices just to get a copy somewhere. I think that model, it's an antiquated model. Whether it's disk or tape doesn't matter. It doesn't matter. It's just, it's a model that's not going to scale. So what you're seeing is you're seeing us trying to integrate the things that own the data, the hypervisor, the storage stack, the application, they really have the tools and techniques to actually allow me to get backups done faster. And so you're going to see us really adopt to a model that says, look, we want to make sure you get backups done the fastest way possible and we want to give you centralised management of that. And where appropriate, we obviously want to sell you the backend storage for that. But it really is all about, it's again that model of on the backend team, I'm trying to deliver services and I can't give you a one-size-fits-all anymore. I've got to use the techniques that work for my end users and for my conservation. And work for the workload. Absolutely appropriate. So VMware, Avamar is obviously appropriate. Got Oracle, you're going to have to use Armand. Armand, and as you move into a world of snapshots, sometimes that's going to be simply managing snapshot replication policies. So are you going to put in the catalogs for those snapshots and replications into your database products? Is that... So that is, in fact, I think if you look at the evolution of backup software, I think that's probably, again, you sort of look at the traditional backup system today and you say, Armand, as a backup software vendor, I've got some value in sort of the storage node, media server, whatever you want to call it. I've got some value in my proprietary clients and then I've got some value in the management stack. I think as we see the market evolve, more and more the value is really going to be in the brains. It's going to be in that management stack and then that bit that does the application intelligence. And that's where you're going to see us focus. And that's going to be about the reporting, the policy management, and the catalog. The catalog, I think, is the anchor. The anchor as you go forward, because it's not just going to be about backup either. It's going to be about archive. It's going to be... It's going to be a signal source of data. It's that source of truth. And as you look at where backup software is going to be in five years, I think we're going to be drawing our differentiations based on the catalog, because the data movement's going to be embedded in other parts of the stack. That's why I found it exciting today. And that's the first time I've really heard from EMC an articulation of that model. And we're seeing it in the lower end with things from NetApp and things sort of, you know, little fringe products here and there. Not, I mean, good products, but in the lower end. There's a side load thesis. Yes. And so they're proving the model in the business case for it. So that's exciting that you're bringing that to the market as a whole. That's very exciting. It's... And, you know, obviously it's my background. I was a long time NetApp guy who worked on the Snap Vault Snap Air products. So, yeah, I think, you know, really critical part of the market. And I think what we want to do at EMC and why I joined EMC is the ability to take that particular part of the model and expand it to all layers of the stack. Like I said, the VMWare layer, the Oracle layer, not just that one off storage layer. And the other thing I found very exciting was your concept of, let me get my notes here. Is it data invulnerability architecture? Right. Is that something, I mean, you applied it to disk and obviously you don't have many tape products so you wouldn't. But what, could you talk more about it? Why couldn't you apply that just as much if necessary to the tape for the areas where it's appropriate? So I think, you know, in some sense, you know, the data invulnerability architecture or at least the methodologies behind it, you really should have those across your environment, whether it's disk or tape or some other media in the future. And if you look in the traditional tape model, backup formats have done things to try to be more robust in the face of tape failures, whether it's inserting checksums and headers or it's putting headers in two locations on the tape so you can be resilient to media failure. I think with this, it just gives me the opportunity to do a little bit more because the data's online and spinning. But, you know, it really takes that model that a lot of us had in the tape market and it says, now we can do more of what we wanted to do that you couldn't do because you couldn't just constantly be spinning your tape scrubbing data. And that for, I think the data domain team especially, they grew up with that as a model. I think as they started to work more closely with the networker team on data formats and long-term data retention, you know, the networker team actually brings a lot of value to bear in that discussion as well. Because of the catalogs and because of their knowledge of everything going on. And because of their experience with their version of data and vulnerability. So yeah, I think, you know, we highlighted one particular aspect of data and vulnerability, but yeah, we could go on for hours about it because, again, ultimately as a backup vendor, I can do everything fast, but if it doesn't come back at the end, that doesn't really matter. Backup is one thing. Restore is what matters. Cover is everything, yeah. And so now we had BJ Jenkins on this morning, Steven, he was talking about this notion and I think he's indicating that you were the first one to put it forth was that, and you might have just alluded to it just a minute ago, is that it was software that was very sticky historically. And that's starting to change. And I presume if you start to think about a new architecture of data protection where you're using, you know, CDP and you've got a more holistic view and there's still a lot of software in there, but it changes the backup application as the sort of tail wagging the data protection dog. Can you talk about that a little bit and then we got a break? Yeah, absolutely. So I think what happens and is already happening in the market is you can look at it and say, all right, you know, because I have different data movement techniques and because it's more snapshot or CDP style oriented or however you want to look at it, one of the things that's shifting is we're doing more and more of storing data in their native formats, which gets rid of that vendor lock-in from software. I have seven years of historical tapes in a proprietary format. You can't get rid of my backup application because you bought me seven years ago but I own your data, right? And more and more shifting towards keeping data in native format so you don't get that lock-in. And I think that's good because now it lets you start to go best of breed. Second piece is, again, the data movement starts to be, again, more open and heterogeneous because you're seeing it at different layers. And again, the stickiness in the backup software stack is really going to come down to who has the best backup catalog and management interface. And I think that's less sticky than the old model because, again, I don't have seven to 10 years of historical tapes that's keeping you locked into me but there's still the stickiness of if I have the best product and I can help you find all your data, you're still going to want me. So software is still relevant but I think, again, the value chain and what is delivering is shifting a little bit much more to, again, being that brains of the outfit. Steven Manley, putting forth a vision. When we first met you, you're relatively new to EMC. It's obviously got your fingerprint now on the vision, the division, congratulations on a great keynote talk and thanks very much for coming inside theCUBE. Always good to see you there. Appreciate it. Appreciate it. And so keep it right there. We've got Bill Holmes of VP of IT Infrastructure at Loomis coming up next. He's going to give us the practitioner's perspective. This is the Future of Data Protection Spotlight. We're here live at EMC World. Keep it right there. We'll be right back.