 Hi everybody, we're back, this is Dave Vellante with David Floyer, we're with Wikibon.org and this is theCUBE, Silicon Angles production of VMworld 2013, we're here in Moscone South, we're at the street level, come on in and see us, we're in the lobby, just look to the right, we have a great location here, thanks to the folks from VMware for setting us up and this is the spotlight on virtualization backup, it's a key topic, we don't hear enough about it, nobody talks about backup and data protection, sort of as an afterthought, but it's something that Dave and I have been tracking for a number of years. Aman Singh is here, he's a senior enterprise storage and backup architect with Jive Software, a very good, cool company out of Palo Alto, Aman, welcome to theCUBE, thanks for coming on. Thank you, thank you, pleasure. Let's start with Jive, Jive is doing some really interesting things, tell us a little bit about Jive and what your role is. So what Jive does is it's basically a social collaboration business software and we are spreading across the, across the lots of large enterprise environments where the, we're trying to see that how, and we don't use email for our collaboration, that's the main thing that, which we'll take care of many things from the collaboration perspective. What I do is that, so from the architecture perspective, from storage and backups, I manage end-to-end support and implementation. Yes, so Jive started in the early days of social networking and kind of had that vision of bringing that capability, that collaboration capability to enterprises. I think of my kids, my kids don't use email, you just mentioned, yes, it's something that is. Yeah, email would be like, after a few years, no one would know what email is. What is that? It's just a matter of time, is it? Exactly, yes, yes. Okay, so you look after the storage, maybe can you paint a picture of your environment generally and we can drill into the storage piece of it? Sure, I mean the way, so we have two aspects of the environment, one is you could have a hosting side of the environment where customers come in and we can host everything for them and then there's on-prem also, but most of the customers are coming into the hosted which is like a software as a service environment and from that perspective all the backend is how we run everything in the cloud. So you look at Amazon, how you can use it as an infrastructure perspective, you can use your own software, but from Jive perspective it's like you are running our software and on our infrastructure and so that calls for redundancy, resiliency in the environment and how things are done, if something goes, happens in the primary data center, what are the backend data centers and how things are run and also from the latency perspective what things could be done from overall perspective that the customers have a very good value feedback from the overall product perspective. So talk about the SLAs that you are sort of promising your customers from the standpoint of how frequently you protect data and what happens if something goes wrong. So it depends on customers, like customers can buy DR and they can also go to enhanced DR which where the RTOs and RPO times are much lower and so it all depends on what customer wants. So we can have something running in the data centers which we can just flip things up and they can have something running in the backup data centers or it could be we can restore it for you in a couple of hours and you can be up and running. So when we talk about, it depends on which level they choose and so forth. So when we talk about data protection as a service you're doing that. You're doing that today. So okay, so take us through how you're doing it. So I would probably go a little bit back how we were doing it. So basically we were using storage level technologies to replicate the data across. Now what happens is that when a customer comes in and he says that, okay, I want a policy where I want to retain my data for one year and so we go back and we're like, okay, we can do that. It's either for all the customers or it's for not all the customers. I mean, you have to keep a consistent policy across the board where you can eat up space so fast. It's like, okay, now what do we do? So you're keeping backups of one year for everyone or you are keeping backups for three months for everyone, right? You cannot pick and choose. There are no policy-based backups. Any service level you want as long as it's blue. Exactly, yeah. So with that, it's like there's no policy-based backup. You cannot have granular backups if you want to and you cannot tweak the environment with what we had as of today. So that called for something more sophisticated and also it also comes to money, that how much money we are spending on monthly basis on that given architecture. Is it scaling well or not? So it was not scaling. We were spending a lot of money in it and there was no granularity in the present backup infrastructure. So that's where we called for, okay, we need to re-architect this whole thing. Okay, and by re-architecting now, you're able to monetize it in new ways, yes? Not only to provide better service and better customer satisfaction, but you're able to charge for different service levels. Is that right? That is correct. And it's not just that. I mean, there was so many things that we could do with the present infrastructure is that, one is you can have granular backups. The other thing is everyone talks about DDo technologies out in the world. But so when you're running, so from a storage architecture perspective, that I don't want to run a DDo on a production system where I want to save space, right? It's like, okay, I want to get that DDo out of the production system where I can do all the DDo from the backend, my backups do all the DDo's. And now the DDo's, what Data Domain offers or AMR offers, basically going with CISL algorithms and variable block DDo's, they are like much more higher than what a given production system can do today. So you save a lot of space. And if you talk about DDo, they can give you run about 40X DDo on your data, which is like you can run a given system today for next five years and not spend a dime on it. It's just a management, even the management cost goes down because it's a centralized console where you can manage everything overall. That's been a huge cost avoidance for you. So huge cost avoidance. Yeah, I cannot give you a number, but it's something huge. Yes. So you can get, how do you get to specific problems? So somebody has corrupted a copy of something or other from a week ago. Can you come bring back a specific copy of a person's? That's the beauty of it. Yes, you can go into that. This system, the present infrastructure, collect on it and say, okay, this is what exactly you want. Even, it's not even if you talk about VMs, right? Even you can dig into the VMs and say, okay, this is the file that I need. And it has happened so many times in our present infrastructure. Okay, we need this file. Okay, I cannot give you this file. I have to give you this whole blob where you go ahead and find from it. And the restore process is presently, or what we were doing previously is a nightmare. Takes so much time. Since our environment is, there are like so many millions of files. So taking a backup and restoring that in the present infrastructure is very hard. But the newer architecture is like, once you do a level zero backup, after that it's a breeze. Just could you describe that new architecture a little bit more? What are the things you're using within that? How are you organizing? So one thing is, I mean, it's a distributed backup. The thing is, going with data domain LMR, it fits into the present virtual infrastructure where you have proxies running, which are the brains for all these software. So it knows what has been backed up and the data domain takes care of all the deduces and everything. And LMR side takes care of, okay, what I'm backing up, should I be backing up this again? How much data I'm going to send across? So it takes care of everything. And it also, since we are like 99% virtualized, so it just fits in, you bring up the proxies and you create the policies that this is how I want to run it. You create the, when you want to run it and how long you want to retain it. And after that, it's just a breeze. So AVMR has the brains and data domain is providing the brains. That's the allocation of resources across the system. And the other thing is, and this thing is that you can replicate these data domains into another site and they can do a bi-directional replication back to the original site also. So you can actually run production in two different sites and not worry about backups. And the other thing is that, one is that you have a local copy in your local data center and the source are way faster than what we do today. Because we have to pull a data today from an off-site location because that's through the SOC2 compliance that you have to have an off-site location. So did you architect this? Yes. This is somewhat unique from other stories that we've heard, David, right? Yes, it is. I mean, you've got a tremendous amount of flexibility between where's the second site? How far away is that? Well, so we have a couple of data centers. So each site depends on what the backup site is. So we have a few sites in the U.S. We are in Amsterdam, we are in Singapore. So it all depends on what is sitting where and we can set up a bi-directional application however we need it. It's just doing that, okay, this is what would be my applicator site, this would what would be my primary site and you can pick and choose. So why EMC? Talk about that a little bit. Okay, so one thing is that actually this is the second time I am architecting the solution with Data Domain. And previously in my previous company, we actually were using a third party software net backup with Data Domain. Since LMR has come a long way from what it was providing in earlier years and there was not much, I wouldn't say confidence in it but it was not there from the enterprise perspective. We never went that route. Now from the EMC perspective, since LMR is part of the EMC now, one thing is if something comes bundled from a given vendor, it becomes much more easier for the support perspective. Let's say as of today, if I am having an issue even from the VMware perspective, if I'm having an issue, it's so easy to get the resources on a single call that here, this is an issue, they all work together. Otherwise, it's always, it's a finger pointing, right? It's not my issue, it's your issue, right? And then they come back, it becomes a nightmare. It's just like, so the biggest thing is single vendor solution works perfectly and it does what I need it to do. I mean, that's- I just did a survey and one of the questions we asked is, it was around virtualization, it was around multi-hypervisors. We asked, are you willing to trade the risk of lock-in for simplicity and about half the survey said yes and there wasn't any other response that was that close. Some wanted to open, some wanted lower cost, whatever it was but is that a fair characterization of your attitude toward working with vendors, vendor management, are you willing to trade the risk or are you willing to risk potential lock-in to get simplicity and service from a single vendor? I would definitely do that. I mean, the thing is, I mean, when you are in the enterprise environment, vendor management is a big thing. Everything comes and falls down to, it's kind of a backbone to help the whole thing and if you can get to a point where it's so simple to get support, so simple to support the whole thing, it's just like- So it's a conversation we have a lot in the keepers. Every vendor says we're not going to lock in, we're not going to lock in, but there's always some degree of lock-in risk. So as a buyer, you have to be cognizant of that and that in and of itself is not a reason not to do business with somebody, it's something that you have to evaluate in your spectrum of decision-making, isn't it? Correct, correct. Okay, good. How about advice for practitioners that you'd give that are facing some similar challenges that maybe you face? What would you say is the most important thing that you did or if you had to do it over again, you'd do differently? So I mean, when we started with this, we did a POC with all the big units in our production data centers. And so the main advice is that, I mean, the biggest thing is know your environment, like what you really want out of your environment and what needs to be done and if you present that to your management or whoever you want to show it that, okay, this is what it is today, this is what we will be getting to, right? And this is what we will be saving and this is, and do I need another person to do that? No, everyone is scared of backups. You talk about a backup today, the first thing you come into your mind is like, oh, sure, is it there? Then you're like, change it? Yeah, yeah, it's always like that. So you want to get away from that mindset, okay, like, someone, so for me, it's like, I can create policies in a given environment and say like, okay, here are the regionally policies. If someone needs a backup, go ahead here and this is the how to do it and do it yourself, right? I don't even need to look at it once I have set it up. Yeah, one job fills, I can go ahead and take a look at it, but I don't need to babysit the environment. That's, I mean, you want to get away from that. I mean, babysitting the environment is not, it's not that we have a lot of time in our hands, right? It's always something or else going on, so, yeah. Yeah, excellent. All right, Amon, thanks very much for coming on theCUBE and thank you, David, for sharing the spotlight here with me. Thank you. All right, keep it right there, everybody will be back with our next guest. This is theCUBE, we're live from VMworld 2013. Okay.