 And now, SiliconANGLE TV and wikibond.org present a focus spotlight. Live from Las Vegas at VMworld 2011, host John Thurrier and Dave Valente illuminating virtual backup and recovery with support from EMC, where cloud meets big data. Back live at VMworld 2011, SiliconANGLE's continuous coverage. We're here with a spotlight on VMware backup and recovery. We're here with Steven Manley, who's the CTO of EMC's backup and recovery division. Now, Steven, welcome to theCUBE. I think you're a first-time CUBE person, aren't you? I have spent most of my life in spheres, these spheres and whatnot. This is my first CUBE time. One of the few, actually. I think most of the people in this event have been inside theCUBE, so it's good to have you join the other 90% that are here. But so, Steven, first of all, let me ask you, your role as CTO of the backup and recovery group at EMC, what does that entail? Are you looking at roadmaps, technologies? What do you do? So I spend, I probably spend about half my time really trying to understand what we consider the three C's. So what are the technology curves coming out, whether it's new processors from Intel or Flash or what have you on the technology side? It's also looking at customers, so I spend a lot of time trying to meet customers, understanding where they're trying to evolve their environments, who plays what role. And then, of course, some amount of understanding where the competition's heading, so we understand where we fit in vis-a-vis that. The rest of it really is then translating all of that into what do we do as an organization? What's our vision? How do we make sure that all the parts of EMC and especially backup recovery systems are pulling in the same direction so that we're building solutions that actually solve customer problems? So talk a little bit about your vision. What is EMC's vision as it relates to backup and recovery generally? And then maybe you can talk specifically about VMware. Sure, so one of the big things that we see that's really changed in the backup space. Yeah, I'd say in the last decade, obviously, is the rise of deduplication and what that's brought with disk. But a lot of the customers I meet, one of the interesting things they tell me is it doesn't feel like my experience has changed much. Sure, I've placed disk in my environment, but I'm still struggling with my backup windows. I still have the same kind of user experience I did before. And I think that's because a lot of what was happening was really under-the-surface changes. We were all sort of retooling, leveraging disk. I think the real trends coming forward and what we see on the vision, there's really two trends to watch for. And they're interrelated. The first trend really is more and more because of the size and the amount of data you have, and because of things like virtualization where servers are much more heavily loaded than they ever were before. We're seeing more and more that for backup to be able to hit its window, both from a backup perspective and a recovery perspective, we need the applications and the systems that own the data to take a more active role in the backup. So that could be VMware, that could be an application like Oracle, that could be a storage system like the VNX or the VMAX. And then the second thing we've seen is as our customers have understood across the board that their data is more and more integral to the function of their business, different customers in the environment, different administrators are taking a role in their protection. So it's not just, hey, let's call the backup team, hand it off to the backup team, the storage administrator, the virtualization administrator, the application administrator, wants to take a more active role in his or her backup. And so what we're seeing then as a vision is the backup team really needs to be able to provide an infrastructure where each of those different groups can centralize their backups, can control their backups, while providing that oversight into retention and to taking them off site and to meeting your compliance and regulatory guidelines. But again, leaving more control and integrating more deeply with the things and the people that actually own the data. And that's a lot of our focus is to really enable that transition as the next wave of really what disk enables for people. That's a very interesting answer and not one that I heard actually and I wonder if we could break it down a little bit. So the first one, he's talking about the application or other entities besides the backup appliance taking a greater role in determining the backup. Let's talk about what that means. Help me understand why that is, what's the benefit of that? Is it more efficient? Is it better recovery performance? And let's start there. So let's take, given that we're at VMworld, let's take VMware. So one of the challenges that you see people hit with VM backups is, you know, I used to have a system that was 20 to 30% utilized. Now that I'm running multiple guests on it, it's running 80 or 90%, oh my gosh, there's nothing left for the backup. So one of the things that we found is, VMware has added change block tracking and the virtue in change block tracking is instead of even something as optimized as, say, our Avamar technology with its source side deduplication where you're still having to look through all the data to find what's unique and interesting. VMware can tell us, here's what's interesting since the last time you did a backup and then you just apply your deduplication against that. That means that your backup window becomes much shorter. It means that the load on your system and on your storage system is much lower and then when it comes to recovery, you can apply things like a change block recovery to roll your VM back to a healthy point in time. So it really is all about being able to scale, being able to hit backup windows, recovery windows, and really just simplify the whole process. So you can use those processor cycles and the appliance for other things? Absolutely, that's exactly the thought that, and again, the bottleneck in a lot of these backups is actually on the host side. So it's not your backend server, whether it's a data domain or something else. The bottleneck is often, again, I'm running 12 guests on this one ESX server. I don't have the headroom to waste time on extra computation. And because VMware controls the data, because it watches the data, because it knows about the data, the fact that it gives us hints just makes sense. So the obvious question there, I guess, the follow-up is that if I've got a homogeneous environment with a lot of VMware, I'm in good shape, but what if I got a lot of different systems and let's say the application, maybe I got a Oracle maybe controlling the backups, I got some VMware over here, can you efficiently apply that approach, that vision across the application portfolio? How do you do that? So that's a great question, and it's one we get a lot. And a lot of times the conversation I have with customers is, well, you're telling me to almost unleash anarchy in my environment. You're going to let the VM guys control some of the backups and the Oracle guys control others and the storage team control yet others. And that worries me because who's going to make sure that, again, I'm meeting compliance regulations that I'm not backing up too often or too infrequently. And I think the answer there first is, look around, it's already happening. You can find that in VMware environments, one of the, certainly earlier this morning there was a session on VMware VDR packed to the walls. VM admins want to be able to control their own backup. You got in environments, Oracle administrators are already executing their own backup policies through RMAN, so it's already happening. The question is how can you gain control of it as a backup team and really manage that transition because I think if you don't embrace it, you're going to get displaced. And so that's really where you want that centralized backend infrastructure where all the data flows into that spot with the backup application then really providing that centralized oversight. So give those local administrators control to a point but make sure you're watching over them so that, again, you're meeting your best practice guidelines for your company. So it really is all about finding that balance and that's where the backup software still plays a role. Something has to be maintaining it. Where the backup hardware comes in is a centralized place where you can store all the data so that, frankly, it's not distributed across your enterprise, which, again, is already starting to happen today for people who aren't intentionally taking steps in this direction. And Stephen, you started to touch on it, that second part of your vision, really the organizational issues. My reaction is, okay, who's in control? Who's the single point of control? Can you talk about that a little bit? So today, if the backup team isn't watching, then, in fact, when, say, the CIO comes in or the CISO comes in and says, can you tell me if I'm hitting all my backup policies? Can you give me a report on what's backed up where and for how long? A lot of times the answer is no, you got to go ask the Oracle team and the VM team. We can give you the data for what we protect but there's a lot of other protection mechanisms. That's not the answer you want to be given as a backup team. The answer you want to be giving is, no matter how it was backed up, no matter how it was protected, yes, here's the report, here's how it's being protected. And that's really, again, where we see people using our tools like the backup applications, be it Avamar or NetWorker, with something like a data protection advisor on top to really give them those centralized reports and dashboards to help them understand what's happening in the environment. So I think the backup team really has to step up and take the reins in this new role, which is you still need to have that control, you still need to be that single point of focus, it just might be that you're not touching the data with your clients anymore, but that doesn't absolve you from the need to be in control. I've talked a lot in the last couple of years about backup traditionally has been sort of a one-size-fits-all, maybe it's a two-size-fits-all. If you got a high value application and you got a zillion dollars to spend, go get SRDF, all right, great. And then everybody else gets daily incremental weekly full. Right. There it is. Sort of a bolt-on to the applications. Is that changing, or first of all, does it have to change and is it changing? So on the first one, absolutely I think it has to change because one of the things you find from a lot of people who have applications is they say, just because I'm not Morgan Stanley or JP Morgan Chase doesn't mean my data's any less critical to my business, I might be a commercial company, I might even be a small and medium business, data's still what I thrive on. And so regardless of what the application is, they want to be able to protect it. And I think the good thing that you've seen over the past, I'd probably say three to five years, is more and more different parts of the stack are stepping up to optimize backup. So again, it's VMware with change block tracking. It's Oracle with their own change block tracking and applied incrementals. It's obviously the NAS appliances using things like an NDP accelerator. So more and more what we're finding is the owners of the data are stepping up to take that role because yeah, you have poor man's backup or rich man's backup, you can't afford a bimodal world like that. There really has to be the ability to have this, I guess sliding window where you can have more choices for the value you need. And tailor it as a service, as you were talking about at the beginning of the spotlight. As an industry, we've been sort of bowing down to the backup window for a long, long time. Is it a pipe dream? Maybe it's my Dave Vellante's pipe dream, but can we eliminate the backup window in our lifetimes? I think so. I think, again, I think a lot of what you're seeing out of the different data owners, again, the applications and I mentioned Oracle, but there's a lot of others right behind them. You see it with VMware and given how much is moving to VMs and you see it with the storage systems more and more, they're really enabling you to have to backup only changes and to get those backups done in minutes as opposed to hours or days. And so, yeah, I think obviously the legacy infrastructure that you installed in 1996 will be the last to fall. But again, the more that you virtualize those infrastructures, the more your backup window starts to become a thing of a past and the more you start to think about how many versions do I need to keep? How long do I need to keep them for? And what's the best way that I'm going to be able to recover them? And that's really the conversation we want to be having, not how can I brute force two terabytes of data over a wire in eight hours? So I asked you off camera a little bit. I said, I might throw out a question about CDP. And my question is, I have a Mac. I use a time machine. I envisioned, put forth this vision, a time machine for the enterprise and that involves CDP. Is that a viable vision? Is it part of EMC's roadmap? What are your thoughts on that? So I think certainly the most important part of CDP that I think you see across the board, again, whether it's something like EMC's recovery point, or again, sort of the change block tracking mechanisms that we've talked about, it's the ability to very easily take the set of changes that have happened and then apply them to a backup. And of course, for that to work, you really need your backup to be on disk. And for that to be scalable, it's got to be deduplicated. So I think CDP, that first part of, you've got to be able to efficiently get the data off without starting each night with, I have no idea what's happened during the day. Let me go figure it out. That's a lot better. The second part of being able to roll back to any particular point in time. I think that's going to be very app dependent. So I mean, already today, you meet a lot of Oracle administrators who will roll back to a fixed point in time, an app consistent copy, then apply logs forward. I think that model will start to gain more and more traction. But I think the first step is really getting rid of that backup window like we talked about. And the CDP mechanisms are really, I think the way that you get there, whether you consider, regardless of how you term it, change block tracking, snapshots, that's the basic idea. All right, Steven, I could go on for a long time. I love this topic, a great meeting you. Really loved the vision. Thanks very much for coming on theCUBE and good luck. No, I appreciate it. Thanks for having me and enjoy the rest of the show. All right, thank you.