 The spotlight on VMware data protection, VMware backup and recovery, I'm Dave Vellante of Wikibon.org and we're here at VMworld 2012 with two very strong domain experts from EMC's BRS division, the group that does all the backup, the AVMR, the data domain, the networker piece, the software piece and it's really come together as one of the big forces in the industry is as we've reported here in theCUBE a number of times, EMC actually has two thirds of the purpose built backup appliance market. It's a dominating market share, something you don't see often in the IT business, that kind of share for such a large market. And I'm here with Guy Churchwood who's the new Senior Vice President of Engineering at BRS. Guy, welcome. Thank you very much. Good friend Steven Manley, multi-time CUBE guest, Steven. Welcome back. Good to see you again, Dave. Last time we saw you was, I believe, at EMC World and before that great vision. We're going to talk about that some more but I'm actually going to start with the new guy and Guy ask you, you're new to EMC, you've just joined the company recently, how have you been spending your time? You've been cruising around meeting people and what have you been doing? Kind of, yeah, I've been onboarding myself, let's just say, making sure that my business cards aren't leaking ink and traveling around to each of the locations and just making sure we had the priorities on all the projects, so that makes sense. So you have a number of engineering facilities around the world, have you had a chance to visit any of those? I've visited every wonderful location. Every one? Yeah, so about five weeks. Racken up the miles. Yeah, yeah, I've literally been in the head office now for about three weeks only for the time I've been here. So what's your initial impression? What are your previous thoughts? It's good, I'm really excited. I mean, obviously I took an XE app from the industry for the last few years as CEO of a small startup in the Bay Area, prior to that obviously I was over at NetApp and strangely, this is my third tour of duty now with Mr. Manley. So I'm not sure if that speaks well for my intelligence or not. Well, it's a punishment, it's a history. After stupidity, I guess. So let's talk about where we are with data protection in the VMware environment and how we got here. I mean, I talked earlier in the previous segment about VMware's initial attempt with VCB, which was kind of a flop, not kind of a flop, it was a flop and then VADP really was a major effort. Of course, it allowed the ecosystem to solve the problem and then we saw this week, VDR replaced by VDP. So that was big news and we had Chad on earlier, he was talking about the effort that went into that, Steve, and it was not trivial. This was a significant effort on our side. Just again, for the viewers, the basics in VDP is more or less that similar to what we've talked about in the past, more and more backup is really moving into the application layer or the hypervisor layer or the storage layer. And in this case, it really is about really embedding the protection inside the hypervisor, inside of vSphere, leveraging some of the Avamar technology to be able to build that product. And certainly on our side, yeah, this was a significant investment, well over 50 engineers were working on this. So this is not some small project, this was a significant investment. Why is backup such a challenge in virtualized environments? You know, for me, it's actually two things. So the first, which I think most people will hit on very quickly is the backup isn't fast enough and more importantly, the recovery is not fast enough. And so they're just constantly searching for ways to be able to meet the performance window for the backup and the recovery in a way that can scale with the degree that they're trying to scale out the virtual environment. So that's the problem most people talk about and that's absolutely there. I think the second issue that people don't focus on as much is the fact that much like the app administrator in say an Oracle type of environment, the VM administrator wants more visibility and control over his protection than what they've been able to experience in the past with more legacy backup applications. And so one of the things we find in these environments is more and more they're rolling out one-off solutions. The VM team is rolling out their own solution for backup because they're either frustrated or just in the dark with what the backup team is doing and they want that control. So it's really two parts. It's the performance, but it's also the fact that since data is king, the VM teams want more visibility into what's happening in protection. So Guy, I want to ask you as, again, the sort of relatively new outsider coming in. You heard me earlier say, talk about EMC's dominance in this business. Now, Steven, in May at EMC World, put forth this new vision for EMC, which utilized snapshots and off-sighting and things like that. Why does there need to be a change? You guys are dominant. Why change it? Well, I think it's definitely progression. In other words, you have to move yourself forward. I think one of the key points that Steven's been talking about is whether you would look at it, as you say, as the Burger King and make sure I go around, which is you have it your way. So in other words, what's important about things like VDP is that the administrator of the backups can actually work inside the virtualized environment. In other words, the VMware UI. So we've really got to be much more cognizant to the way in which people want to interact with this technology as opposed to slamming them with technologies. So I think that what we've been working very hard at is almost like reversing the way in which we look at the technology. So it's not so much the underpinning infrastructure in the center, but it's more on the edge, which is the way in which somebody interacts with the technology. So Steve, can you summarize that vision for our audience briefly and then talk about how you get customers from where they are today to 8.8 to point B where they want to go? Absolutely. So if you think about it, again, it really comes down to those main two factors. How do I get backups moving faster so that I can recover more quickly? And that really means your protection environment. You need it to be application or hypervisor aware because that's the object that you're looking to protect, but you need to be able to leverage all the technologies underneath, all the technologies embedded to be able to hit those performance windows. So for the hypervisor, it's things like change block tracking and change block recovery for VMware. For Oracle, it's things like incremental merge and block change tracking. At a storage level, it's things like clones and snapshots and replication and being able to leverage de-deplication and tiers of storage. But no matter what level you're doing it at, you need to be able to leverage the intelligence inside of the application, the hypervisor, or the storage array to be able to meet the performance window and be able to recover quickly enough. The second part in the vision which Guy talked about is it's not just about the backup team driving anymore and somebody always having to pick up the phone to get the recoveries done. So I think we use the analogy of sort of an enterprise time machine. When it comes down to it, if it's my file, if it's my application, if it's my virtual machine, I should be able to recover that on my own as quickly as I need to to be able to keep the business up and running. I shouldn't have to be picking up the phone, trying to communicate with other teams and slowing the process down. So again, at the heart of it, to be able to scale, I need to leverage the intelligence inside of the things that own the data. And to be able to meet business needs, I need to be able to expose those technologies to the users, whether it's an end user, an app owner, a hypervisor owner, a storage owner, that's going to make the business go forward. So Guy, in thinking about the portfolio that EMC has, it's pretty vast. It's involved acquisitions of Avamar, acquisitions of Data Domain, and even acquisitions of Legato, actually, for Networker. So now you've got this very rich set of assets. How do you make them work together? Is that a primary objective of your engineering organization and how do you do that? So yeah, it's definitely a primary objective of what we have to execute against. I mean, I think it's important to note that, slamming technologies together for the sake of it isn't worthwhile. You have to be cognizant of what the customers want. So what we have been doing internally is looking at common services and then exposing the common services across multiple platforms and iterations. And so I think that the vision that's been set out is very good. We have to be a little bit more aggressive onto it. Can you give an example of those common services? Well, actually, so deduplication, what we've done around DDBoost, I mean, that's a centralized function that we have. We're doing some work on the UI now, which is going to be all-encompassing across multiple platforms. And even the work we've done around VDP, you know, the engineering, the 50-plus engineers that we have kind of shout out to the Irvine team and Boulder. I just have to kind of do that, otherwise they get upset with me. But what they've been doing is basically driving that and then we actually have strategies to run off the back of the technology center and allow the same use, in other words, almost like a bridge to the virtualization environment, both in the Avamar world and the network of world, you know, to drive that thing forward. So these are common services. Are customers typically installing, you know, to both of those technologies in Avamar and a data domain together? Are you seeing that increasingly or is it more different use cases? What are they asking you for? So I think one of the things we've seen is since we first integrated the data domain in Avamar, actually two AMC worlds ago, we've seen a constant increase in the uptake of people leveraging Avamar with data domain together. The other thing we've seen is since Networker 762, and now especially with the new Networker 8 launch, more and more networker customers are also integrating, deploying the integrated solution between Networker and data domain. So while it's absolutely true that we still sell a lot of data domain into other backup applications or directly behind, say, an Oracle or a SQL type of backup, we are seeing increasingly aggressive adoption of Avamar plus data domain or Networker plus data domain. So it really is coming together. Yeah, so it's very important for us to make sure that each of our client bases, I mean, we have a very large established base of Networker customers and data domain customers and, you know, Avamar and Source One, you know, each of the products, we have to make sure each of these ones move forward. So I don't see that we would end any of the technology streams that are basically leveraging common services across them. So you're seeing this notion as well, change block tracking in the VMware space and block change tracking in the Oracle space, for example. I mean, it just kind of makes sense to only back up the changes. How does that change the way in which you architect products and go to market? So on the architecture side, I think probably the biggest shift is understanding that, you know, on the data domain system, you're not always getting redundant blocks the way you used to. So five years ago, you know, the vast majority of data that piled into a data domain, you pulled out immediately because it was already duplicate. Now, given that we're only getting the unique change blocks, the real value now is in being able to store them efficiently as possible because I still want to store all my backups as fulls. I want to back up changes because it's fast but when it comes time to recover and we all know recovery is the reason I do backup, when it comes time to recover, I want fulls. And that's where the data domain technology of being able to share blocks is critical because now I store all these backups, even if they come in as changes as fulls, so the recoveries are fast and efficient. So from an architecture perspective, it's just forced us to take a step back and say the real value is in the block sharing and the recovery performance, not as much in the, I just deduplicate streams of bytes coming in. Yeah, so that's really talked to from my market question. It used to be a huge emphasis on dedupe rates, which really doesn't mean that much, especially in a virtualized environment. You're emphasizing things like recovery rates and that really is what matters to customers is what I hear from customers and I'm glad to hear that your strategy is aligned with that, so that's good. All right, well, we love the vision, we love the disruption, we love the fact that you guys are willing to disrupt your own base and make some changes there. I mean, even VDP, you know, is going to eat into your Avamar base a little bit, so that didn't get lost on us and so I think that's what tech companies increasingly have to do to grow and protect their franchises, so it's a good move in our opinion. Congratulations, Guy, on joining EMC, we'll be watching, good luck and Stephen, I appreciate you coming on and sharing your insights with us. Thank you, appreciate it. Always a pleasure. All right, keep it right there, we'll be back with customer segment, VMworld Live, 2012. Keep it right there.