 Live from New Orleans, it's theCUBE, covering VeeamON 2017, brought to you by Veeam. Welcome back to the Big Easy, everybody. This is theCUBE, the leader in live tech coverage, and my name is Dave Vellante. I'm here with Stu Miniman, my co-host for two days of wall-to-wall coverage of VeeamON. Gavin Cohen is here, he's the Vice President of Product Marketing at Nimble, a Hewlett Packard Enterprise company. Now, Gavin, good to see you. Congratulations on the exit. Thank you, Dave. So, new times, new, whole new lot of thought going on. I mean, first of all, what's it like to be part of HPE as it's early days, but how's that going? So we're a few weeks into it. It's extremely exciting so far. We're running at a thousand miles an hour. And what's been absolutely terrific is, I mean, the acquisition, it's an expansion acquisition. So that means the entire Nimble Storage product line continues to exist and stays alive, but we get access to a massive global sales force that we didn't have as an independent company. So, very exciting stuff for us. And a huge channel as well. I mean, I haven't talked to folks at Nimble on theCUBE anyway, since I thought it was back at one of the SNW's, Stu. So maybe you can give us the sort of Nimble 101 if you wouldn't have to have it. Sure, sure. So if you look, I mean, really the things, there's several things that set Nimble apart. We all, with a bunch of other flash startups, had first products to market around 2010. Nimble really accelerated that. So to the point of the acquisition, we had over 10,000 customers worldwide. And we really managed to very much change the game in storage from starting as a company focused on hybrid storage. We had a very successful launch last year of our all flash and managed to turn a very large portion of our business into all flash. But overriding that and probably the thing that sets us apart more than anything from not just the storage startups, but from all the large storage vendors is our use of predictive analytics and what we've been able to do with it. So talk a little bit more about that. Sure. So, I mean, our platform's called InfoSight. And the idea is in the infrastructure that we exist in, so the storage array spanning all the way through the networks, the compute all the way up to the hypervisor. Every day we collect millions and millions of sensor data points. Actually as a collective base, we're processing every second millions of these status sensor data points. And what we're doing with it is we're passing it through all these techniques of predictive analytics and machine learning. And we use it really to predict and prevent problems. So our goal is not just delivering fast flash performance from the array, but really this end-to-end delivery of data up to the application in a better way than otherwise possible. So you kind of had in the early days, it was the original EMC phone home, right? We all remember that. And then around sort of the virtualization guys, the three-pars, the compalence, they had what we used to call the hero reports. And it was good. And it was kind of a phone home on steroids. What you're talking about is a whole new advancement in analytics that drives, you know, anticipatory actions potentially. Is that right? Is that not your question? Very much. I mean, there are probably three numbers that speak to it. So 86% of problems that would normally involve a call to vendor support. So in this case, nimble support. We completely end-to-end automate all the way from recognizing the problem before the customer even sees the problem through to resolution. And it's pretty remarkable because it's not just the stuff you'd expect from a phone home where we recognize a power supply is going wrong or a SSD is not working correctly. We can recognize, you know, misconfigurations on the host or a bad HBA or a multiparthing, you know, setting that's not correct that's impacting performance and then proactively tell the customer about it. They may not be aware and actually tell them how to resolve it. So it's kind of a remarkable one. Yeah, Gavin, I say, if this announced today, you'd probably say that you're an artificial intelligence company that's going to help. We just hadn't quite coined that word when we came out and, you know, it really is because you look at, you know, I said there were sort of a few interesting metrics. The other one that's sort of been astounding, particularly for, you know, new technology in a world where storage has been around for many, many years, we've hit well over six nines of measured availability across our install base, but not just across one configuration across every nimble array out there running every version of OS in every kind of environment we're well over the six nines of availability. And then probably the most astounding of all is 54% of the issues that InfoSight resolves are not actually tied to the storage. So they're all these problems that are outside of storage and that's the stuff that customers just love because these are these needle in a haystack problems with VMware settings or, you know, problems on the network that get blamed on the storage and end up having a root cause outside of. How do you get visibility on, you know, beyond your own little world of storage? Yeah, so that's part of our secret sauce where we have these, you know, collectors. So it all stems from the array, but we also collect up through the stack. So we have our vCenter agent as an example and they all feed in analytics. So a lot of what was built into Nimble from day zero was just this infrastructure of sending out sensor data and then collecting it and processing it. And then over the years we've just expanded. So we started where we just collected from the array. Now we push out of the array and sort of covered most of the infrastructure. And that's really where the differentiation is because when you correlate all those different data points you get some really interesting insights. So you ingest that data and essentially in real time. And then process it and spit it back out. And help the customers. I love this new metrics. You know, when you can, I think it gives substance to disruption when you have new metrics that you're creating, like in particular 86% of the failures are automated or the problems are automated, no humans. So I guess that's the fourth metric which it's hard to get, but how much time you save people. It is and it's almost impossible to measure because no one publishes their amount of time wasted on storage. But we know just anecdotally when you talk to any customer, any customer with any vendors products, when they run well, they run well. When they don't, hunting down those problems and dealing with multiple vendors and everything, it's an absolute nightmare. I think that's what we've managed to sort of crack into and really deliver something better for our customers. And I mean, the other while we're on numbers, net promoter scores get thrown around a lot, but as an independent company, Nimble has the highest storage net promoter score in the industry, so we crossed over 85 as a net promoter score. And it's mostly when you talk to customers, it's just that support experience. They've never seen anything like it from a vendor. That's great. I couldn't help but notice when the keynote was going on, they put up the key sponsors that at different levels, both HPE was there and Nimble was there. So your team was already planning to be here prior to the acquisition. Tell us a little bit about the partnership, any specific products you have at Nimble that fit in this space. Yes, so probably two pieces that are interesting. We have very deep integration with Veeam and we were actually the first of the smaller storage vendors to be integrated with Veeam. If you looked initially, they integrated themselves with the big players that you'd expect. We were the first of the others. The integration lends itself exactly to what we do well. We do a really good job of snapshots and replications and supporting the number of snapshots and replication points. So it's just a really slick integration where you can drive the entire backup process through Veeam but actually behind the scenes Nimble does all the data movement and the snapshot creation under Veeam's management. The second thing, and this is actually a product that we showcased for the first time at Veeam on. So at our stand, now the HPE stand is what we call a secondary flash array. And it's kind of a very unique device because when you think about backup, most backup repositories, they're a one-way repository. You put stuff in, you access it when you need to but when you access it, it needs to come back. You need to copy it back and it's slow. And what we've done is we've built a secondary storage device that's great at accepting Veeam backups. It's got inline D-Doop and compression and everything that's very efficient but you can actually run real workloads on this device. So we've come up with this idea of put your backup data to work. Instead of having it sit there idle, you can spin up DevTest and QA and do things with that data or verify your backup because now you have performance. Yeah, it's always been the problem with storage, right? If you make replicas or if you have backup so you've got a certain amount of resources that aren't being used or the other piece is backup's great but recovers everything. So you need to be able to be fast. You need to be able to be nimble, I guess, would be the case, right? We love that. So it's really, I would infer that what that is is a productivity tool that you can also use for backups. Is that a fair way to do it? I think it's actually how it'll end up getting used. I think the use case always starts with backup. You need to put your backup data somewhere and most people will choose to put it somewhere that's highly cost optimized, knowing full well the trade off is when you need to restore it, it's not going to behave like your primary device. This is opening up a whole new, as you said, a new use case where you get the data there but then really the interesting thing comes that you use that every day. So you can run all these other secondary processes on it or you could fail over to it and actually run production on it if you needed to. And you can be cost competitive because of your data reduction techniques. Is that right? Is that right? Can we, okay, so for those of you out there that don't believe that, let's push on that a little bit. The spinning disguise will tell us it's true for so-called high spin speed devices but when you get to the cheap and deep stuff, we're still much, much cheaper. Your counter would be that you can't reduce, data reduce that stuff effectively or is that right? I think really you got to look at the usual, the cost trade-offs, right? If you want, and now the portfolio that we're part of is a perfect example. If you want the most cost effective place to put your backup data, it's the HPE Store Ones product. It's totally designed around being an efficient destination for backups. It's got DDoop like nothing else that will crunch that data down and you can store it for months or years, very cost effectively. Then you're done. Then you're done, right? Now you can get the data back and it's absolutely rock solid but it doesn't behave like a primary storage device. Our secondary flash array is somewhere in between the cost of primary disk or primary flash and hybrid disk and that sort of cheap and deep in that it's got a lot of the low cost attributes because of compression and DDoop but it's now got IOPS so you can do things with it and that's really where no secondary devices has gone before. Data sharing and it's got a cherry on top and some sprinkles. That's it. Gavin, last question I have for you is, the acquisition's done. You talked a little bit about the channel. Many people look at InfoSight as kind of the gem of your portfolio. Can you give us any guidance as to where we can expect to see that driven throughout the HPE portfolio? Yeah, sure. So the best thing is I'm not yet, I think, subject to all knowing about all the rules of what I can and can't talk about. I'll let it all out. No, I mean, as a very clear stated direction, HPE acquired Nimble, a large reason was InfoSight and just looking at what we've done as an independent company, I mean imagine if you could start to transform the support processes that HPE could offer and bring some of these capabilities to their own product line. So we're already embarking on looking at doing that first with the three power product line and while I won't give you dates, I can say that there are a lot of people aggressively working to get something out. I think you'll see that spread pretty quickly because the IP that we have and the data scientists and the sort of infrastructure that we've built to perform these analytics is extensible and we're pretty excited about that. So yeah. Excellent. Gavin, thanks very much for coming on theCUBE. Thank you very much. Appreciate it. Appreciate you coming on. Thank you. All right, keep it right there everybody. We'll be back with our next guest shortly.