 Live from Barcelona, Spain, it's theCUBE! Covering Cisco Live Europe, brought to you by Cisco and its ecosystem partners. Hello everyone, welcome back to theCUBE's exclusive coverage here in Barcelona, Spain of Cisco Live Europe, 2019, I'm John Furrier with Stu Miniman and Dave Vellante here this week covering all the action in cloud, data center, multi-cloud, our next guest Jim Frey who's the vice president of strategic alliances at KenTek Technologies, groundbreaking report that came out of the Amazon re-invent conference with a lot of customers. Part of the multi-cloud discussion. Jim, great to see you, welcome to theCUBE. Thanks, it's Frey by the way. Frey, I'm sorry. No worries, no worries. Multi-cloud, your report has some interesting data. Talk about the survey, the results, what is it telling us? Yeah, we've been working hard at KenTek on extending our solution to start covering the cloud, multi-cloud server, hybrid environments. And so we were at the AWS re-invent show and we decided to take the opportunity to talk to some of the attendees and just sort of get their view of what some of the challenges are. So we talked to a little over 300 of them and we asked them a few questions, not a rigorous thing, you're doing it on the show floor, right? But we found some really interesting things out of that. So the first thing is, is that it really is a multi-cloud world already. And it more so even than hybrid. So we had nearly 60%, 58% of the people who we talked to had more than just one public cloud in play. They almost all had AWS, of course, because it was an AWS event, but not all of them, which is really interesting. But they either had AWS plus Google or plus Azure or plus some other cloud, more so than even hybrid. So we also asked, are you using AWS in conjunction with your own private data center or a third-party hosted Colo Center? Only 33% were doing that. So we were surprised. And the reason that's really significant is monitoring and management of these environments is much more complex in a, well, it's complex in a hybrid environment. It's even more complex in a multi-cloud environment. So it sounds like there's some real need for some help there. What are the challenges and, what are some of those complexities, what are the challenges in the monitoring? Well, so that was the next question. What's the key challenge? And usually, whenever you ask someone about the challenges, the number one answer is always, oh, security. That's my biggest concern. That did not turn out to be the case here. The biggest overriding concern across all the different sort of levels of people we talked to was actually cost management. And cost management is, it was a bit surprising, but usually you hear security, security, security and then something else. This was cost management, either number one or number two and number one for most of the constituencies. And in some of the subgroups, like VP level, SVP level, architect level, it was overwhelmingly the first choice. 40 and 50% of them are saying, yeah, cost control is our biggest issue. Even ahead of other things like performance, like visibility, like actual control of the environment. So cost was really the biggest concern. And that's a big issue. Jim, something we've been tracking, especially at shows like this, at the Cisco show is the challenges, I used to understand kind of the stuff that I had in my data center. I could get my arms around it. I might not love the management tools that I have. I might complain about some of the cost, but it's all very well-understand. It's bought mostly as a capex. Right, when you get to the public cloud, I totally understand what you're saying, multi-cloud, is now I've got all these different pieces and how well do I have them defined. There's different skill sets between them. And when it comes to cost, right, the big unknown is, oh, hey, am I getting surprised by what happens in that environment and across all of them? I mean, I've talked to plenty of companies that will dedicate an engineering resource just to manage cloud. Or I have many friends in the industry that are helping, you know, cost optimization is something that, you know, software consulting, there's huge business in that because we're still early in this, so we're getting to those steady states. Help us connect the dots. Where does Kentek play into this then? So you've talked to all these customers. Thank you. Our viewpoint is network, and we're trying to give a viewpoint of what's happening in this environment by watching the network. And that's always super valuable because it helps you localize where things are, what activity's happening, it helps you see which workloads are talking to which workloads. And that reveals sometimes things you don't expect. And this is where the cost control comes in because, you know, in the cloud environment, you have to pay for certain network traffic, especially between availability zones or when you're shipping out of the cloud back to your other, you know, your home environment. And we have talked to a lot of customers who've said, hey, end of the month comes around, I get my bill and there's this big number there for data transfer. I don't know what drove that. And why am I being surprised time and time again by this? Well, the network view point's really awesome for seeing that. And if you can do it with a monitoring system that's watching for that all the time, the good news is, is you can catch it, figure it out if it's real or not, need it or not, and fix it before 20 days later you get a big, fat bill. What is fixing it means? It means like keeping it contained in the cloud or on-prem or managing what's moving around. Could be combination of things. One of the things that we've seen in some of our early deployments are someone moved a workload into a different availability zone. Well, there was an application dependency that didn't recognize. And, you know, that workload was talking to, you know, a home data center or another availability zone and creating traffic across there and just running up the meter on the network costs. So if you can see that, and it becomes very obvious if you watch the traffic patterns, you can at least have someone go say, okay, that's a surprise. They had a big rise in my zone-to-zone traffic or my, you know, cloud-to-home traffic. Let's just take a look at who's driving that and whether or not it's something that should be or shouldn't be. One of the interesting trends we've been watching, obviously with cloud and hybrid cloud, is kind of the consumption and deployments of cloud. And hybrid's interesting because hybrid's about cloud operations on-premise, which has been slowest to deploy. Wikibon's done a lot of research on private cloud and why that's happening. But it seems that cloud sprawl on the public side has been there. So yeah, I got some Amazon, easy to stand up. Got some Azure and now Google. So it's probably easier to get stuff in the clouds. And then now they got a repurpose on-premise to kind of have this seamless cloud native environment, hence Cisco's announcements, et cetera, et cetera. So as that's happened, what have you guys learned and seen in terms of the customer behavior? They wake up, obviously the bills are higher. So it makes sense that multi-cloud is higher than on hybrid and the cost containment is a concern. How did they get there? What are you seeing? What's the psychology of the customer? Just share some insight into the customer behavior. Is it oh shoot, I got to unwind this? Do I double down? What's going on? I think it really depends a lot on what the projects are and what the objectives are and what the skill set is. But one of the things that we found in this survey is that that network viewpoint that helps you understand what's really happening in the production environment is often being underutilized or underappreciated in the cloud environments. In the cloud deployments and cloud infrastructure. So one of the things we asked about was how many of you folks at this event are actually taking advantage of, for instance, VPC flow logs, which can tell you exactly what's happening within AWS between availability zones. And it was surprising. They've been around, VPC flow logs have been around for years as a technology and as an additional service available. But only about a third of the respondents were actually using them. So they weren't taking advantage of this important insight and viewpoint telemetry set. About a third kind of knew about them but wasn't using them yet. And then another third didn't even know what they were. So I think there's still some maturity happening, some maturation happening in terms of understanding what can I do about this? How can I get ahead of this? What's at my disposal? So part of the challenge, of course, then, is I've got that piece covered. But as you said, now, how do I cover my home front? And where do I find some sort of tools that can be put these things together so I can see it all as one? And that's where you guys fit in. That's where we fit in. So let me get some anecdotes from you. One, it's clear that there's a pain point. Take the aspirin, understand what's going on. Contain the bills. Give a scenario of what they're doing to contain that you mentioned a few of them. But also give an example of where they're using the data to be proactive. So there's the vitamin inside of it. So vitamin, aspirin, whatever metaphor. So I got to contain my cause, I get that. How are people using the data to be more proactive in either architecting or deploying? So I think the, I don't know that anyone's being proactive yet. That is certainly the promise and the opportunity. Most organizations are simply want to be more aware of what happened or more effectively reactive and you start there. And once you start to realize, hey, I can't do this, then you can start turning towards being more proactive. So for instance, our solution was built to allow you to trigger corrective actions back to the environment. We don't take the actions where we can trigger the systems that would change configurations or change policy and inform those systems of what's happening and what sort of parameters can we recognize that indicate an issue. So we believe that, especially watching the change in patterns of activity, noticing the anomalies. Anomaly detection is oftentimes used around security use cases and we do that. But also it should be applied to operational use cases. When does a new workload pop up or a new volume of traffic show up that I didn't expect? And if it's something I recognize happens on a regular basis and I know the answer, let's automate the corrective response. So that's kind of our theory of provide you the understanding of what's happening then with the tools to trigger an automated corrective action. All right, so Jim, we're talking a lot about multi-cloud this week with Cisco. Of course, Cisco dominant in the networking space, really feeling out where they live in multi-cloud and how networking plays across all of them. What's the relationship between Kentik and Cisco? How does that work? Thanks, so we're a member of the CSPP program. We're a partner. We joined because we manage a lot of Cisco gear. So a lot of our customers have Cisco. A lot of our use cases historically have been at the edge of the network, in particular the service providers. So those that are delivering internet services are using the internet to reach their customers in some way. So what's really different about us is we do a really deep and detailed approach of integrating BGP path data and BGP route data and correlating that with the traffic, as well with other enhancements and augmentations to the data that give business and service context to the network traffic. Makes it more actionable. Yeah, and what are you doing in the container space? You mentioned edge computing. Got some interesting use cases. Maybe explain a little bit where you play there. So when I say edge, I'm talking about internet edge, not edge computing, although we're fascinated by what edge computing represents and the new challenges that's going to bring. Now when it comes to containers, actually we're very fascinated in working in that area too because, John, as you mentioned, moving and implementation of new cloud workloads is cloud native, using Kubernetes, using things like Istio. That changes the environment. So we've actually built a connector into Kubernetes so that we can use that to pull service information. In terms of what workloads, what containers are out there, what are they doing, what's their purpose. So when we show you an activity map of site to site communications, we can say, here are the actual services that are being, that are participating in this activity. Istio is another place we're really interested in to look at the service mesh that's being set up to run and operate communications between containers because that's a new sort of virtual cloud network. It's a way that these containers are communicating. And again, the more you understand about the communication patterns, the better you can recognize problems, the better you can balance and plan, the better you really get a handle on what's really happening. Jim, I want to get your thoughts since you brought up edge of the internet. Multi-cloud and hybrid cloud as data moves around certainly brings up the question of which routes the packets are moving around. There's always been debates about SLA's around direct connections versus go through the internet. It's China looking at it. So there's a security kind of concern. What's the trend that you're seeing with respect to say either direct connects. If I'm a company and I have multiple clouds, I have the connections in there, I'm concerned about latency certainly cost. Whether it's cat videos or whatever or applications, it still costs money. So latency's important. So each cloud has its own kind of latency issues. What have you seen? Well, getting to the cloud and then within the cloud. Yeah, exactly. So it's complicated. So this is a new dynamic, but it's similar concept. Is there standard latencies? Is it getting better? What's the trend look like? That's a great question. And I honestly don't have a good answer for you, but I recognize and agree that those are common concerns that we hear. And the best thing that at least for what Kentic is doing is to provide the means to measure and understand that. So you can compare what's working. You can document a baseline, your different options and your different paths and recognize when there's a real problem occurring. When you start to see licensees spike to any particular cloud service or location or zone, so that you can try to get on top of it and figure out what you can do. So it's a classic case of evolution. Get it instrumented. If the providers get better at what their service is, that's really out of your hands. It's not really. Okay, so getting back to the survey to kind of wrap things up. Interesting, it's at Amazon, the biggest cloud show. Azure pops up on the list as pretty high. Makes sense, Microsoft's got great performance. I mean, Azure's kind of like they move a lot of stuff into Azure, pre-existing Microsoft stuff, plus they're investing. What's the bottom line summary? As you kind of, you know, the aroma of the rapport. What is it, what's coming out of the rapport? What's the key insights that you glean out of this? So I think it indicates a normal pattern of adoption and sort of we're growing into this marketplace. It's evolving as we go. You know, we saw big early adoption happening in like lift and shift approaches to just move stuff into the cloud. Throw it in the cloud, it's going to be cheaper. It doesn't turn out to be cheaper. It can be. Then you've got another set of organizations that are born in the cloud, right? And they've started out from the beginning. But so those two early approaches are now merging into how do we really use this as a true strategic approach to IT? What are the real world complexities we're going to deal with? And how are we going to deal with those? It's really no different from the way that technology has evolved within the traditional data centers and why the way virtualization came in and changed the way we built and architected data centers. It's awesome, it's great. It saves you money in one area but then it created huge blind spots because you couldn't tell what was going on in those virtualization layers. So we had to adapt our operational monitoring and operational practices to accommodate the new technology. I think we're going through the same thing now with the cloud. People recognize that they don't necessarily want to be beholden to a single cloud provider. They want alternatives. They want cost, you know, competitiveness. They want redundancy. And so multi-cloud, I think, is becoming more and more real in part because people don't want to put all of their eggs in that one basket. And cost certainly looks good on paper at the beginning. But then, as you said, the side effects, the system, so these consequences, the system when you start growing or whatever. And that's where people just have to work it better, right? That's pretty much the operational. I mean, let's apply the same rigor that we used to apply to traditional data center environments and let's start embracing the cloud, right? So, Jim, you talked about the multi-cloud bit. Why don't you put a fine point on it? There's a reason why you jumped from being an analyst into the vendor world. Some people on the outside will be like, well, you know, cloud's been going on for 10 years. Seems we understand where this is going. But tell us why, you know, now is so important for this multi-cloud environment and the opportunity that you see at Kentic in the ecosystem. For Kentic in particular, what we're starting to hear very loud and clear amongst what our traditional and initial base of customers was facilities-based service providers and digital enterprises that managed big-routed networks and needed to understand and can better control their relationship with the internet and delivery across the internet. They're coming to us and saying, hey, look, we're splitting. We're adding cloud workloads. So we're moving our content that we're serving up into the cloud. You know, more and more of our systems are moving to the cloud. And we rely on you for this visibility in our production environment. We need you to add this. So we saw a demand from our customers to accommodate this. And in parallel, we're just really inspired by this next generation of cloud-native application development. It seems to be starting to reach that point where it's becoming reality and it's becoming mature and it's becoming a reliable approach to IT that now is the time to really get serious about bringing these other best practices from the traditional world and applying them there. And the survey data is great. It proves multi-cloud and hybrid all here. Costs can run out of control. You got to work, you got to operationalize cloud. And the same rigor, I love that. Great insight, Jim. Thanks for coming on theCUBE. Appreciate it. LiveCube coverage here in Barcelona for Cisco Live Year of 2019, it's theCUBE. Day three of three days of coverage. We'll be back with more after this short break.