 Live from Las Vegas. It's theCUBE. Covering VMworld 2018. Brought to you by VMware and its ecosystem partners. Welcome back to theCUBE. Our continuing coverage of VMworld continues. I'm Lisa Martin. It's day one. Dave, we've had an awesome day so far. Got to co-host a lot together today. Yeah, keynote, some great, great, great technical talk. Some more technical talk coming up. Exactly, momentum. We're all excited to welcome back to theCUBE. Andrew Hillier, the CTO and co-founder of Densify. Andrew, it's great to chat with you. Great to be back. So I said you're an alumni. You've been on theCUBE a couple of times, but for our audience who might not be that familiar with Densify, tell us about Densify. What do you guys do? What do you do that is unique? Well, we're analytics. We're analytics of the service to optimize supply and demand in your hybrid cloud environment. So we do on-prem, but we do a lot of public cloud as well. So it's like learning the patterns of the workloads in the environment and then making sure it's running on the exact right infrastructure, which in the cloud can be very complicated. So it's really almost like supply and demand optimization for your cloud. I loved that I saw something in the prep notes for the interview that in terms of the challenges that Densify is seeing, cloud is like a giant Costco in the sky. It's got all these aisles full of stuff that you aren't even aware of. I loved that analogy. Unpack that a little bit for us and how are you helping customers to actually navigate those challenges? Sure, yeah. You can take the analogy quite far because it's like a giant Costco with huge aisles. You don't know what's in them. The stock is changing all the time. You've got a giant shopping list. It's kind of vague for people and you're in there trying to figure out what to get and people do it wrong. We see everybody, it's terrible what you see out there because I think organizations have really automated in the cloud the ability to turn on and off apps, but to figure out what they should be put on is still very manual. And so it's again, we see people trying to do this manually. It's not something that humans can do. It really needs to be done through analytics. What did you come up with the name? What's behind it? Densify, well, it really kind of captures we let you do more on less. So if you picture in a virtual environment, picture like a giant game of Tetris to make all the workload fit into less infrastructure, in the cloud it's the very same thing. We say, well, you're doing this much work, you're spending that much or you're deploying that many resources, you can do it on a lot less. And usually it's significantly less. So it creates kind of a cost challenge, but we almost see it more as an automation challenge where linking these two things together automatically rather than having people have to figure it out. But you guys are Toronto based. We're talking off camera. People may not know this, but there's an incredible developer community in Toronto. And I don't know, I really don't know why. Maybe it's a combination of the IBM labs is up there and the whole blockchain and Ethereum thing, but there really are some talented developers in that region, aren't there? Well, absolutely. I think there's some great universities, you know, University of Toronto, Waterloo is very close by and there is a concentration of certain industries. We were laughing earlier that the flight down here is just full of other tech companies coming down. We're all kind of on the same flight to come to VMworld or reinvent. And so there is certain concentrations that's quite a bit of talent. You know, we like it because we have a very strong team, development team to develop the analytics and it's a great market for that. So the cloud in a lot of ways is, I mean it should be super simple, but a lot of ways it's complex to figure out all right, what should I do when people talk about multi-clouds. I heard a stat today in the keynote that the average large organization has eight clouds. I was like, we're 50 people and we have eight clouds, easy. I can think of eight off the top of my head. So how does Densify help simplify that and really what problems are you solving? Yeah, so what we do is we start and I'll step back for a second. If you look at the cloud world, there's a whole lot of products out there. They all kind of start with the word cloud that kind of got to start in reading the bill and understanding what you're spending and giving visibility into that. And that's useful, but we think we've kind of moved past that now to the next generation. So what we do is we start with the workloads themselves. We learn the patterns of exactly what's happening in the workloads. Some patterns might be, you know, some workloads might be kind of steady, some might be very peaky and that will completely change how that translates into the cloud services you buy for them, right? So we kind of start by learning the patterns of activity and we understand well, that thing ideally you would be putting it on a T2 or now a T3 burstable instance, or an M3 or an M4 and just Amazon alone, I think there's 1.7 million different ways you can configure these things. So we kind of start with the machine learning of the patterns and say, okay, we do like a giant permutation to say, here's how all these workloads map to the optimal catalog and how you purchase them and do you prepay for them and all of those details. So it really kind of takes a guess where it's really not something humans should be doing. It's, you know, you can do this algorithmically with the right analysis and then just completely automate the whole thing. So prepay, do you reserve instances? How many, how much? Yeah, it's funny because we see, for example, with reserved instances in the cloud, you can prepay and you'll get a discount on that. So people will say, well, I'm using all of these certain types of instances, I'm going to prepay for them and save 30%. And we'll say, well, no, you shouldn't even be using those instances. Don't double down on that, fix it first and then you'll be running on a lot less and then prepay for that. So we see a lot of people actually locking themselves into the wrong configuration because they just went and prepaid without actually optimizing first. It's a big problem. And the corpus of data that you need to be effective, I mean, how much time do I need or how much data or what do you, what do you find is the optimal situation? It's nice in the cloud because if you're talking about a VMware environment, we're at a VMware world, you know, a lot of times you'll talk to vCenter and have to start gathering data and you want to get a business cycle to understand if there's a monthly peak or something. A lot of the cloud APIs give you history right away. So you can get up and running really quickly. So we find that the combination of a SaaS delivery model and these APIs and the providers means that you can kind of, in 15 minutes, you can sign up and set this thing up and be getting historical data back and you get answers very quickly. So you don't, it's good if you're not very patient because you don't have to wait around for a month to see answers, you get them right away. So I'm going to kind of fumble through this question but Andrew, maybe you can help me. So a lot of the cloud providers will have what I would call hardcore primitives in their APIs. I think Amazon is an example. We were just talking about the data pipeline and how complex it's becoming with, there must be 15 different proprietary APIs for whether it's Kinesis or EC2 or S3, DynamoDB, et cetera, et cetera, et cetera. Other cloud providers, maybe try to simplify that. What are you seeing there? Is that a problem for clients? How do you sift through all that? I mean you guys have to be like super cloud geeks to figure that stuff out. It's very complicated and you have to treat each cloud differently, so one of the things that we do which is quite important is we don't water them all down in the one model. We will roll up all your clouds into one view but each, the analytics of each provider is its own nomenclature, because things are differently. The database services, the scale groups or scale sets, they all behave differently. So you treat them each in their own native way and you got to stay on top of what they're doing and they're constantly releasing new stuff. So Amazon just released a T3. It's actually a fairly complex thing to optimize because of all these credits. So we have a team that gets on that and adds it to the analytics very quickly and so all our customers just start benefiting from that right away. So if you're a densified customer, we will start informing you, yeah, that new thing that came out, you could be using it on 40 of your instances right now. So you kind of automate that process. The way I think of it is that customers should be focused on their apps and their unique business differentiators not following the news of what's happening in the cloud. Because you can just keep, you can forever be reading new things about what's happening in the cloud and we do that through the analytics. So you can just focus on your business. So you customize that model for each client? Is that right? For each cloud provider, so not for each client, but for each provider, the analytics would look are different and the way you reserve, Google has committed use discounts. They're very different than our eyes in AWS. So they each have to be treated natively and one of the things it does is, I think a lot of providers do like people to kind of not move around very much. So, you know, they like to kind of up level you in, once you're in that provider, if you start using the other services API, then it's really hard to move somewhere else. So I think if people go into it thinking they're going to be able to be portable between clouds, usually when you land in a cloud, you start leveraging that provider's features, you're going to stay there. In the words of the great Paul Moritz, the cloud is the mother of all lock-ins. He was right about that. It's true. But it's so alluring that's that value of trade-off. I don't know, but I've found that people will trade off the risk of lock-in for business value. Yeah, I mean, I think if there's a benefit and if you settle on a vendor and you choose them, and I think people like to not have to retrain everybody on freedom and cloud providers as well. So there's natural things that lock you in anyway. Once you've made that decision, I think a lot of people choose two vendors to keep them honest, or maybe for two different types of workloads. We see analytics or ML going to Google, but other workloads going to Amazon. But you've got to be foolish not to leverage all the unique and very powerful services they offer. So last question, Andrew. I saw on your website a big, big teaser. It says, the next cloud revolution is coming, October 15th, 2018. What can you give us? Well, I don't want to talk too much about it, but we've been doing a lot on automation. And I think just characterize that when we come up with these answers, one thing that we do that's very important is we are very precise that we give an answer that you can actually automate. So you can't come up with an answer and say, yeah, it might be better on that, or you might want to do this. It has to be very precise. And we really specialize in that. And so we've been doing a lot of work on actually making it happen automatically. And it's kind of like a next generation of automation. It's pretty cool. So that's coming out on October 15th. You can go to our website and kind of pre-register to keep on top of it. But it's a really new thing because again, we see that as the kind of the next generation of automation. We've gotten somewhere in the cloud right now, and this is really the next thing in our view. And it's really cool. And it's October 15th. We'll be waiting with bated breath. Andrew, thanks so much for stopping back by theCUBE and we look forward to hearing what's coming out. Thank you. And thank you. In about six weeks. Yeah, in six weeks. We want to thank you for watching theCUBE. I'm Lisa Martin for Dave Vellante. Thanks for watching. From VMworld, we'll be right back with our next guest.