 Live from Las Vegas, it's theCUBE, covering AWS re-invent 2017, presented by AWS, Intel, and our ecosystem of partners. Welcome back everyone. This is theCUBE's special exclusive coverage of AWS re-invent 2017. CUBE's our flagship program, we go out to the events and extract the signal noise. I'm John Furrier, co-host with me today is Justin Warren, analyst. We have two sets here, here in Las Vegas. Our next guest is Sumit Singh, vice president of Cloud Analytics with Juniper Networks, formerly of App FormX, which was bought about a year ago. CUBE alumni back, new team Juniper, welcome back. Last time we chatted with you, you were entrepreneurial. Taking names, kicking ass, now you're... Bought Juniper Networks, yeah. You're part of Juniper Network, what's going on? So we've essentially been building, building more and more, and it's actually been a totally awesome experience. So last year, when we spoke, we were essentially looking at a whole lot of private cloud deployments, looking at OpenStack, looking at Kubernetes, looking at VMware, and since what we've now started really expanding into is of course the multi-cloud and hybrid cloud scenario, and looking at how to secure these clouds on-prem in the cloud, multi-cloud, as well as bring rich analytics into real-time operational insight as to what's going on in all of these environments and how to optimize them. Yeah, that whole multi-cloud, hybrid cloud thing has really exploded in the last sort of 12 months. I've been hearing from customers a lot more that they are pursuing a multi-cloud strategy, but it seems that there's just this proliferation of things that you've now got to try to monitor and secure. So how are you helping customers to do that? I mean, you've got to start with the basics, right? So the first thing that we've got to realize is, I mean, of course, there's companies that are born in the cloud, but then there's a whole bunch of others who have to normally run their own data centers, run their application stacks on-prem who are now looking to migrate to the public cloud and build out that entire multi-cloud scenario. In that situation, I mean, you need, I would say a little bit of hand-holding, right? You need to understand how your application is running on-prem, which ones can be moved to the cloud? How can they be moved to the cloud? You want to ensure that those policies that you were implementing on-prem, you'll be able to implement those same policies in the public cloud as well. So the monitoring, really, starts on-prem, right? All of those policies, the definition starts on-prem, and then you take them and you build them and you- So Shami, I've got to get your take too, Justin, on something that's going on that I see clear visibility on. Infrastructure, operations, data center cloud, get your house in order, networks, migration, hybrid cloud, multi-cloud, and then all that stuff. Then you've got this developer tsunami going on, renaissance of real new development, new kinds of development, multiple databases using an app, you've got IoT, so the software development methodologies are changing. Front developers, that's obvious. What's the impact to the infrastructure guys? Because you start to see Lambda and serverless as a way for saying, complete infrastructure as code. How does that change the notion of what the hell the data center is? Because you could argue that's just an edge now. So what's the software, what are some of the software practices you see that are notable? So what's truly amazing, like in all these things that you're saying is that you no longer need to use one approach to build anything. Any product that we put out, or any service that we put out now uses a combination of all of these things, it could be Lambda, it could be IoT, it could be a wholesale application that's orchestrated using Kubernetes that's spanning that multi-cloud environment. So the beauty of all of this is the power of choice. We have so much more choice available to us, right? And then when choice comes, that explosion and that complexity, the operations ask, complexity is key, but speed is also there. You're seeing, so the question is, at what point does the cream rise to the top and the people that have slow get run over? Well, we're just seeing another evolution in abstraction really. It's like we don't write in assembly code anymore where we're writing directly to the hardware, we add it in higher level programming languages. Now, in terms of the infrastructure, developers don't care about infrastructure. As much as people talk about DevOps and they think, oh, DevOps is the thing, developers don't want to deal with infrastructure. They want to deal with code, because that's where they live. And the infrastructure folks, well, a lot of them are actually becoming developers now. So they're learning how to use tools like Git, like using development tools to actually get their job done, which is where we see infrastructure as code. So there's a lot more of abstraction into pure software so that you don't have to worry about the underlying abstractions, at least not very much. Yeah, so many questions to you now on that is, that requires the network guys, Juniper, you're part of that, and all the analytics to think differently about what you're instrumenting. To do what he said, to make it free, you got to enable a lot of policy, a lot of data analytics. Take us through what's the current state of the art there. So the current state of the art is essential. If we talk about Juniper products, we have our family of SRX products where you can have on-prem firewalls as well as virtual firewalls in the cloud. And using these tools, you can have consistent security policies on-prem and in the cloud, you can create transit VPCs, connect up your applications in the multi-cloud world and do all kinds of fancy things. But where we are also going with our solutions is to make them much more simpler to consume. It's truly all about simplicity, right? Because now you have all this choice and you can have Lambda and you can have all these new ways to bring up your applications. What becomes key is that the policies that you want to implement become automatic. And the way to do that is, the way we are doing that is essentially doing this auto-learning of your environment. Automatically understand. Automation, right? But not automation in two parts. As in automatically detect what's going on but then automatically apply the policies as well. No matter where the workload is and where it's scaling, we automatically apply the policies to it. So it's a lot of investment in the smarts of underlight. Making something simple is actually quite complex to do. So you need to understand what are the right things to automate and what are the few things where you actually want to give humans that choice without it becoming overwhelming. So that, okay, I have to choose between one of 800 different ways of doing this. That's just not something that humans cope well with. Whereas machines are actually really good at that. Yeah, and that's the value here. We want to hide all the complexity under the hood. You know, use those advanced ML algorithms. Use, you know, whether they be on-prem or in the cloud, running all of the analysis, implementing all the right policies for you. Right, new workload comes up, it should automatically get the policy. Right, and we are now able to do that both in the private data center as well as in the public cloud and bridge those policies together for you automatically. The common theme we're seeing in cloud, we just had a guest on from Thorne where they automated essentially police officers writing down notes on a notebook to fully spotting with machine learning, all this great stuff to find missing and exploited children. Manual, that sucks basically. Right, you got to do manuals, low. The workload's too dynamic now for you to think about. I want real time. So what's going on there? How do you guys help there? What's the progress? Oh, so this is actually, great question by the way. So, and this is part of the reason why we, like as a company, as a startup, maybe we were like doing all this cool stuff and in a way thing and, you know, not really thinking about all the, hey, this is slowing me down. The reason why we went to Juniper, if you look at the history of Juniper and the product portfolio and the stack at Juniper, when it comes to automation, when it comes to things like APIs, when it comes to things like policy, they've always kind of like led the pack in that networking space. And now this is the opportunity to take that, that wealth of knowledge and scale it out and take it to the broader, multi-cloud hybrid cloud space. But that's truly where it is. And even if you kind of like go down low level to the devices, all Juniper devices are able to stream real time telemetry. We're able to do ML in real time, even on the physical devices, right? Similar for our virtual devices. And now with our Formics, we even bring in the performance and operations inside from the running infrastructure, whether it's on-prem and in the cloud, not just networking, but the compute, the databases, your applications, your Kubernetes clusters, all of that to build for you this end-to-end view, right? Not just the networks. Your servers, VMs, workloads, the underlying network, the connectivity, all of it. Right. Because the developers, they live in application land. And again, they don't really care about that infrastructure. But as it turns out, sometimes it's quite useful to know which particular network devices or what the infrastructure is that underpins things. Like where you sometimes need to be able to drop into assembly code to really optimize things. So are you making that information about the infrastructure visible to developers in a way that they like to know and consume? Absolutely. So one key thing about our product portfolio and how we are releasing our services is essentially we've wrapped everything around these role-based access interfaces, where both the operators are able to get their views, they're able to construct views that the developers are able to see, and then both can implement their own policies. Right? If let's say there's some infrastructure that's down or is unhealthy, then having that global topology view helps you in real-time correlate and in real-time informs you what the impact of that outage is. Like who are the developers who will be impacted? What are their applications? And we can bring that insight and then consume it to run the automation. So if let's just say some infrastructure is unhealthy, can you read out the traffic? Excuse me, talk about what you guys are doing here. Obviously Amazon, big learning conference, but it's a massive show, 45,000 people here across multiple hotels, a lot of sessions. What are you guys talking about? What's the big cloud piece for you guys? So for us, really, first it's just visibility, right? We won't, we have a product portfolio that gives you visibility, right? Like both for your physical infrastructure and your virtual infrastructure, right? Then the next thing is of course, you know, yeah, you have the visibility, but then at our scale, no human can consume all of that information, right? It's too slow, it's too slow. So you've got to have the machine learning built in. So it's converting that visibility into insights in real-time, right? And then it's about how do you secure your workloads, right? Consuming all of that insight to implement all of the policies, implement all of the automation to ensure that everything is running as you want it to. All right, so what's your Juniper message to the developers here? Is there a new face to Juniper? Is there a new vibe? You mentioned Juniper's always had great products. They can move packets around at lightning speeds, you know, wire speeds, all that great stuff. How do you, what's new? What does it mean for me as a developer? What does Juniper, how does it make my life easier? What's new is that now it's easier for developers to consume our products. Our products are now available in the Amazon Marketplace, right? Our visibility products, our machine learning products, our security products, right? You can just click, install, and start using them. That's new for Juniper, right? I mean, traditionally you would think of- You don't get the Juniper goodness just by treating it like a library. That's it. You can just download, not even download, right? You're running an Amazon, right? Serverless, it's routerless, it's device-less. There you go, you can just start consuming them. And then if you do have that knowledge of how to use those devices on-prem, you can apply that knowledge in the cloud and use them, or- There must be a grid computing back in, like, what, 20 years ago. I mean, isn't that just a grid now? Almost, pretty much, yeah. It's a fabric. It's the same, it's like, if you already know how to use it in one place, you know how to use it everywhere. I mean, it's, but really the value in the cloud is making it even simpler, right? Running all of that automation. Like, we talked about Lambda. Like, even within our product family, we can use Lambda to constantly see what's changing and that's how we process lots of our internal transactions as well. So, to me, congratulations on your acquisition and your entrepreneurial journey. And now you're at Juniper, looking forward to keeping in touch. Thank you. So, to me, seeing Vice President of Cloud Analytics, and now at Juniper Networks, formerly of App4Mex, Cube Alumni. Thanks for having us coming on and sharing your commentary. I'm John Furrier, Justin Warren, here on theCUBE, Bain Stage in Las Vegas at 8 of Us ReInvent. We'll be back with more after this short break.