 Okay, we're back. This is Dave Vellante. I am a co-founder of Wikibon.org, and I'm here with David Floyer, who is one of my co-founders at Wikibon.org. He's the CTO of Wikibon. And this is theCUBE, where we bring you the smartest people in the industry. We invite guests in, CEOs, CIOs, analysts, practitioners, and extract knowledge and share it with you. These are people that are oftentimes independent as is our current guest. Jason Buffington, who is with the Enterprise Strategy Group based in Milford, Massachusetts, is an independent analyst firm. Jason follows the data protection marketplace, and is a real expert in that space. Jason, welcome back. Thank you so much. And second time on theCUBE, and we're talking data protection. It's an interesting marketplace. I was saying earlier, we had BJ Jenkins on, and I said the same thing with Steven Manley. When you look at things like purpose-built backup appliances, and you look at data duplication and the effects that it's had on backup, it makes sense that it's going to replace tape and cannibalize that business. What surprised me is the pace at which it did that and the incremental market that it's taken on. It's much bigger than the tape market. And it's going up and to the right, tape market's flattening and decline, and it's pretty amazing dynamic. It caught me by surprise. And why is that? Not that it caught me by surprise. Why is it working that way? I mean, I caught me by surprise, I'm a crappy analyst, but what's the market dynamic that I missed? You know, so ESG recently did some research and some of the things that we saw, first of all, for the first time in a couple years, improving backup is like the number one thing that folks are investing money in in 2012. It's always in the top 10, but it bounced back to the top of the list. And if you think about what else is going on in IT, I think it makes a lot more sense. It tied with number one, with increasing use of server virtualization. That's the good part, right? Server virtualization solves so many problems. The downside is VM sprawl makes backup so much harder, right? Then you take that. The VM in general makes backup. Absolutely, absolutely. Then you've got unstructured data growth that makes backup harder. You've got private cloud architectures where whole services and whole templates of machines come online that makes backup harder because it's more to move on. All those things are talking about, those are all growing the growth on production storage and whatever you're going in production storage, then you've got to have four to six X that in backup storage to make sure it's all protected, which is where I think why DDoop took over so fast because as that production storage kept growing, if it wasn't for DDoop, you wouldn't be able to backup everything that was growing in the production state. Jason, you were talking about VM sprawl, sort of pressuring backup. Can we dig into that a little bit? So why does virtualization make backup such a headache for people? Sure, for one thing, today with the commoditization of hypervisors, it wasn't that long ago that there was one, I guess what you call, enterprise credible hypervisor. But today, the hypervisors are more pervasive, it's not just enterprises, it's the mid-market that's doing it, and the reality is it's just too darn easy to stand up a VM, right? It's about a half a dozen clicks and all of a sudden this new VM comes out of nowhere and starts grabbing footprints of storage. Now the data... You've got to get me a server. Yeah, it's like, oh I want a server. The days of saying, well, oh, I want a server. That means I'm going to go ahead and, how much CPU and how much memory and how much hardware, and I'm going to go and figure out what the vendor is, and I wait four weeks. Thanks for budget. And I actually provisions it. No, instead, it's like, oh yeah, yeah, we got available resources, I'll click this much of this, change some dials and hit go, and it's there, right? It's too darn easy to stand up a VM. And that's when you're doing them one at a time, right? Now when you think about private cloud, when it's elastic and it's dynamic, now you've got this business owner who really, doesn't care what's running behind the scenes, he says, I want a new app. And when he clicks that new app button, all of a sudden there's three new VMs for the web front end, there's some middleware, there's two SQL backend servers on it, six new VMs stood up. How does the backup guy know how to go back that up? And he would just click in a self-service board. So virtualization makes it so much harder just because it's so easy to stand up. And the challenge is, ironically, is that when you're using templates, most of that data is redundant, right? Every one of them's got a Windows OS or a Linux OS, and it's all stuff you don't need 80 copies of. And on top of that, you've actually utilized your servers much more efficiently, so the spare bandwidth you have. Hey, talk about that a little bit. So you've got the simplicity factor. I thought simplicity was a good thing, which it is, but it's just, there's a ripple effect. No free lunch. And then David, you're saying that now, that I've virtualized. My observation on that marketplace thing is that whereas before you were using spare resources in the server to do that backup, that ability has gone away. There's just no longer there. So you're actually having to put in real resources to do the backup. So you have less physical resources now to do the backup. Because backup is one of those applications that does not underutilize servers. Okay, so now you're saying that because you've got less physical resources, when you go to do the backup, you don't need real resources. You don't need as much juice, not as much horsepower. So that pressures backup Windows. And then in addition to that, there's an increasing number of snapshots that are being used in the environments. And those snapshots need to be backed up, put somewhere else, replicated somewhere else. So that constant supply of snapshots, replication and real backup is putting a huge amount of extra capability in the hand of operations. And better RPOs and RTOs, but at the same time using resources. And using them because they're meeting the issues that people have. So what? Which is the poor backup Windows and the poor recovery rates. No, I want to pile on there though. One thing you mentioned was, is that it's putting all those extra burdens on operations. But one of the things we're also seeing more and more of is that it's not just the IT operations guy who's focused on that anymore. Right? So the database administrator says, I want to backup some of this up. In fact, a great example of that during a BJ and Steven's keynote on Wednesday. You know, the DBA wants to be able to own it himself. The virtualization person wants to own it themselves. Absolutely. And then there's also the challenges where the senior level execs. One of the things we saw in Ishii's most recent research is that regardless of which group is choosing the backup recovery technology, it's often the business owner, the stakeholders in the backside, they're actually paying for it. Which means now they think they've earned visibility into how is the backup recovery process going. So you have pressures on a higher level visibility. You've got DBAs that want to backup themselves. You know, so you're losing control in many ways. Or you're raising liability and visibility and you're losing control of the environment. There's a lot of things that just make backup harder today. So what happens? Go ahead, Dave. I thought one of the most interesting things of Steve's keynote was the introduction of the catalog. The introduction of the catalog as being the key piece that they're going towards. So that you choose different ways of doing it. Different products, different snapshot technologies appropriate to the different applications. But you bring that together with a catalog which is keeping track of the copies of snapshots, the copies of replication, where they are, and being able to know where the data domain copies are, et cetera, et cetera. So it's a metadata repository around? So it's the metadata and the catalog that's going to become the center of importance of getting that right. And then using the resources to dial the specific requirements of the application. It comes down to application specific. And where's that metadata reside? Is it sitting flash? No, some of that metadata better reside at the flash level because when you need it, you're going to need that in space. So yes, there's going to be active data metadata. So it's archived data if you like but there's going to be an active component of that which you always will need to get color. So backup's driving the need for flash. That's a small amount. So let's talk about what that does for changing the role of the backup administrator. So, and some of the announcements we heard today were around, say, DD Boost and how it's actually enabling the Oracle DBA to be able to do its own backups. That's self-serviced. Putting it back towards the application. And that's the right thing to do. So from an evolutionary perspective, certainly backup to disk is faster and almost better in every way than backup to tape. That's great. But it drives all the storage and therefore you need DDoop. And so then the question is, how smart can you get DDoop? And the closer you get the discernment to the production workload, the better off you are. That's great, but. And then in the case of like R-Man, it doesn't even go through a networker or an MMR server anymore. Just go straight to the data domain appliance. That's the good part. The bad part is that backup administrator, his job is still a responsibility of assuring is the company's data protection. And so the idea of actually being able to make sure that okay, the Oracle DBA gets to be able to do what he needs to do, which is assure that he can recover and assure the resilience of the application data set. The backup guy still has to be able to say, am I meeting the SLAs of the company established? And so the idea of actually updating that catalog with that metadata, he actually becomes almost a data protection governance administrator, not a backup admin anymore. And the nice thing on top of that was if you have an architecture, which will tell you that level of protection and you can measure it, then you can put risk against it and measure that risk. And I thought that again was the first time I had heard a clear renunciation of what's been in tape for a long time, which is the invulnerability architecture there. Can you imagine what that conversation would have been like if they didn't have that feature? Can you imagine if it'd be like, the backup admin tells the Oracle guy, okay here, you can start doing your backups. It's not my problem anymore. It's right. And no one's going to buy that because the first time that something goes bad or they can't get their data back, they're going to go straight back to the backup admin and say, why weren't you making sure the DBA was doing his job too? Right. This solves that. Yeah. So what are the skill sets of a data protection governance administrator and how do they compare to a backup administrator? You know, the backup administrator is the guy who primarily thinks in terms of, I need to put an agent on the box and I need to know where my tapes are. The data protection management administrator, and I made that up so I'm not remembering it the right way, he's an enabler, right? He's not necessarily the doer. He's the enabler of how that works. He's the CTO setting the architecture. Absolutely. Providing services to the parts of the organization that want self-service. Exactly. And at the end of the day. And I think that's a scale, right? So in some cases, what you may find is that person's still doing the whole job. In other cases, they've completely given it away like to the Oracle DBA and he's just responsible for governance. And then there's that middle ground, right? Where he's the coach, okay? Let me give you those best practices to move you along the way. Another direction I think that same role is going to go is as folks start looking at, say, cloud-based backup as an additional tier, that same data protection management administrator, they're going to end up being the compliance officer for is that cloud provider providing the SLAs you require there? He ends up needing to be a negotiator as well. A lot of upside for data protection, a data protection management administrator. I'm going to have to do that into a blog for a while. With a strong dotted line to the chief security officer. Yeah, yeah. Yeah, data protection in a security context. In a security context, obviously linking to other parts of the organization other than just ours. By the way, that's a much better career path. We just had a couple of minutes. I wanted to bring it back to this. This is always astounding. This is IDC data. I don't think ESG does market sizing. Not yet. Not like this. IDC did this market data. They do it every year, every half a year, whatever your corner, whatever it is. The pie is amazing to me. The EMC has so much of this market. Now, I guess it's not that surprising because you acquire data domain, it's the leader. But the fact that they've been to sustain it now for a couple of years, even grow their share with companies like IBM and HP and Symantec and Quantum and Oracle, you know, attacking them. Do you guys think this is sustainable? Or is this going to just, people are going to slowly chip away at this? I think the model is sustainable. People want to have outsource the complexity of backup and have an appliance. That's very, very clear. The amount of deduplication that you will do in the back end through data domain, I think will come down a lot because you'll have the snapshot technologies, you'll have the applications themselves, the avamars doing a lot more of them. So there'll be a shrinkage of the amount of... Opportunity to do that. But the marketplace itself, and especially if EMC goes through on their current strategy is theirs to take and theirs to own. You agree with that? I do. I do think the secret sauce that makes that pie wedge so big though is DDBoost. The fact that it's accelerating an ecosystem so that other backup players can grow on that and leverage the data domain back in, I think that's the thing which continues to make that compelling. So just the attract, that vision you're saying, the attractiveness of that or the actual sales of that? No, just the fact that whether you... So what they got right now? So they've got the backup exec, they've got, they just announced Quest, the Netback. So they've got all these different backup providers. So regardless of what you want to talk about for the how, DDBoost makes that back end fill into that blue pie. So as long as they're the leader from an accelerator to deduplication, I think that's what's going to fuel that back end. When other folks start creating those accelerators, I think it becomes more interesting to watch. Well, and I thought the way to compete with this would have been the sort of vision of a time machine for the enterprise using Snapshots, CDP, sort of a whole new vision event and data protection. Which for instance, Jim McNeil of Falcon Store has been on theCUBE and he put that forth two years ago at VMware, but now you got EMC talking to 13,000 customers about the same vision. NetApp have the similar vision. But once EMC's on that curve, the customer's going to go, okay, that sounds pretty good. I'll wait, you know, maybe it'll take them six months, nine months, maybe it'll even take them a year, a year and a half to get there. But if there's a roadmap there, customers trust EMC because they got the great service, they're doing a lot of other good stuff with it. And so EMC's becoming, you know, that sort of trusted infrastructure partner. So that vision was exciting. You guys got excited about the keynote. I think some of the smaller players are challenged now. They got to come up with a way to attack that pie, not by copying what EMC's doing, but by innovating and providing transformable services. I think they're going to find ways of putting the technology into specific environments. So in the same way, Avamar has done very well with VMware. There are other products being others that do a very good job and they will, they can put in part of the solution. There'll be Oracle themselves, may well come out with something for Oracle. I think it would be, and Microsoft is going to be a big component there. Yeah, so there's a lot of players that can take from that pie and make money. Interesting dynamic, but I just, you know, it's been, but I agree with you that EMC have got the driver's seat. Now that they've put that vision forth, I think it sets a lot harder for people to steal it. That's a game changer and was an opportunity for others that I, quite frankly, I think the market missed and now you got EMC talking about it. And so, and they're so good at execution. And even when they're not good at execution, they're so good at marketing that they can close the gap until they can figure it out. So it's a real conundrum for the competition in my view. And if they own that catalog and if they are smart enough to make that catalog heterogeneous so that they can bring in to that catalog other components of it, then they're going to be really very well positioned in every aspect of selling storage to the, this shared infrastructure, this shared storage which is going to be a combination of backup and archive and active data, all three of them together. Jason Buffington, I'm going to give you the last word, any stuff you're doing at ESG, any research you want to highlight, activities you got going on, predictions you want to make? Sure, so follow JBuff, J-B-U-F-F. That's where you find all my tweets. You'll find out about it first. As far as research goes, data protection modernization 2012 just came out. We took a long look at what are people doing for cloud, for D-Doupe, for how they're protecting their virtual environment differently than physical, and where snapshots and replication play into a larger data protection strategy. So it's been some great research to put out and the public reports almost out. Great, at JBuff, Jason Buffington from ESG. Thanks very much for coming back on theCUBE. He's at D-Floyer, I'm at D-Valante. Thanks very much, great to talk to you. Great job. Thanks everybody for watching. Keep it right here. We're going to cover end to end. We'll be back. We've got another couple of segments today. My co-host John Furrier, we'll be back. Keep it right here. EMC World Live 2012. This is theCUBE, SiliconANGLE.tv. Keep it right there.