 Live from Gillette Stadium in Foxboro, Massachusetts, extracting the signal from the noise, it's theCUBE, covering VTUG's New England Winter Warmer 2016. Now your host, Stu Miniman. Welcome back, I'm Stu Miniman with Wikibon and this is theCUBE live from Gillette Stadium for the VTUG 10th Anniversary Virtualization Technology User Group. Been talking a lot about virtualization and cloud and happy to have on the program for the first time Sarah Zellahusky who gave the AWS VPC keynote this morning. She is the lead site reliability engineer from Reactive Ops. Sarah, thanks for joining me. Thank you for having me. All right, so Sarah, before we get into it, can you tell us just kind of a little bit about your background? Tell us for those in the audience that don't know Reactive Ops, you know what they do. Sure, absolutely. So my background is actually in science. I started my career as a physicist and astrophysicist but a lot of the problems that you solve there in science are computer-based, Linux-based. So from there I got into SysAdmin and then became a developer for a little while and I've kind of hit all the bases. And so I ended up in the ops world managing infrastructures and things like that and from there I kind of made a jump into DevOps. And that's what I'm doing now at Reactive Ops. Reactive Ops is a small consulting firm with a lot of smart people, a lot of people who know what they're doing and what we do is we help companies transition from either their co-location data center into AWS, build their infrastructures, architect them a solution, help them get continuous deployment, continuous integration type deals. And then we also carried them on to managed services so like help with pager and monitoring and alerting and things like that. Well, that's great. I like the keynote this morning. I think about two years ago was the first time they really brought in the cloud pieces and there was kind of the AWS 101. But you gave a little bit more practical, a little bit more how-to on VPC. So can you explain how VPC fits into the picture and why that's a good kind of starting block that people that looking at it ends up? Yeah, sure, absolutely. So VPC is the building block these days in AWS. EC2 Classic was the original incarnation of AWS and it was very different now. Everybody was shoved into the same public IP space and there were not as many resources and not as many walls and VPC was the offering they brought to allow you to have a private cloud inside this public cloud. So it's kind of classic networking, but not. It's a little bit different, right? So what you want to do is transition people's knowledge from what they know about their physical servers, their physical networks and their co-locations into the cloud and so you do that with VPC and really people need to know the building blocks. What does my network look like? What are the routing tables? What are the security options I have available to myself? And so we went in depth with VPC today and I'm hoping that a lot of people took out from that, hey, I have knowledge about these types of things and here's how I can move into the cloud and carry that knowledge through. Yeah, when we look at AWS, VPC is one of the leading points when they talk about their hybrid solutions because it kind of bridges from my on-prem environment to the cloud. What do you see? How are customers, do those terms mean anything to them? How do they think about kind of their on-prem and the public cloud and how those things tie together? Sure, we have a huge range of customers. Some customers are in an on-prem situation and they want to make the full jump to AWS and that's a migration and some customers have sensitive databases or storage, things like that that they want to keep on-prem. They want to keep it secret and safe and what they feel is controlled. And so, you know, VPC gives you the ability to have both of those situations. You can connect your on-prem, your co-location to the VPC securely and you can run services in both locations or you can migrate yourself and we see customers that run the whole gamut. Yeah, so what would you say are some of the kind of the largest misconceptions that people have when it comes to using AWS resources? A lot of people are frustrated. They think that it's not secure because it's public cloud, right? And that was, I think a lot of the misconceptions came from EC2 Classic. It is a much different beast now than it was when it started. And as you can see with AWS, they come out with new features every day, every week. There's a huge list, right? And so people don't know what the now is. And so I say most of them are worried about security, obviously. And Amazon has come out with a lot of new features lately to help with that. Another of the misconceptions is that you are sharing hardware with others. And for a lot of security compliances, that's not allowed. But not only can you have single-tenant hardware in the public cloud, but Amazon actually will work with you. They will sign BAAs and things like that for particular security compliances to help you mitigate your security. So. You bring up a couple of really good points there. Number one is, of course, that cloud is changing so much that the cloud that, gosh, if I looked at it, not even last year, but last month, it might have changed. And the second thing, we found in our survey and our talk to customers is if I haven't done it, security seems like this big wall. And then if I dig in and I look at it, and most people, you go look at your own environments and you think about the security that I have there. I mean, one of the number one things I talk to most customers and they're like, you know, wear security on your priority list and then wear security on your projects. And it was like, oh, if I only had more time, I would handle more of my security issues. But if I could go to the cloud, boy, that they're updating it, they're changing it, they're much more proactive than anybody is inside their own data source. Absolutely. I mean, it's in Amazon's best interest to give you security that you can control. And actually, you're right. A lot of people are worried about digging in. It seems like a big wall. But it's actually pretty simple in AWS. They give you the ability to have public and private subnets. They give you the ability to have network ACLs and they give you the ability to have security groups. And those are three very basic security tools that will allow you to get 99% there. And they're adding new things every day with AWS config, which will monitor your instances configurations and other things that will give you insight into the security that you have in your BPC. So you mentioned continuous integration and continuous deployment. How are customers wrapping their heads around this? Because we think it's kind of the typical IT environment. And my refresh cycle for hardware, five to seven years, my update on major software, 12 to 18 months, even patching, might only be once a month if you're likey versus if I go to the cloud, I mean, they take care of it, but kind of the DevOps culture and CICD, is we should be flipping things on its head as to... How fast we do this. So where are our customers and how does your company help them? So actually we find a lot of customers that we take on are in a position where they're treating their cloud account as if it is classic. They don't update instances all the time. They have long running instances that aren't patched, that aren't updated and they're afraid to do so. But the more we talk to customers and say, you can have an immutable infrastructure. You can stand up an instance configured and with your application deployed on it. And then you can throw it away and stand up a new one five minutes later, you don't have to worry about sysadmin tasks. And I would say most customers are very excited about the fact. They are interested in not having to deal with all of these baby servers, right? And so I say a lot of them are very excited about it. They just don't know where their developers fit in. So they're developers who need to tweak things or make things special for particular situations. That's where customers have a hard time like wrapping their head around immutability. Okay, so let's maybe talk a little bit about DevOps then. It's kind of a major focus of reactive ops. What does DevOps mean to your customers and how does it help them, your typical customer? Yeah, DevOps is an interesting, nebulous concept and a lot of people are unsure. To me, DevOps is, it's not all about the tools. It is a culture. It is a culture in which your developers need to be open to working with your infrastructure and you need to work your infrastructure so that it is usable from your devs. And a lot of the conversations we have with our customers is what do your developers need? How do they be more productive? And why does your infrastructure have to be so strict? Why can't we make it flexible? Why can't we make it work for them? And a lot of the conversations are around productivity and making your business work for you. And so at ReactiveOps, we're really about making all of the tools, which a lot of people think tools defines DevOps. Ansible, Chef, Poppe, configuration management, automation, those types of things. They really just steps along the way to get your developers to be more productive. Yeah, it's interesting. In some ways, it sounds like an extension of what we've talked from virtualization, which was getting out of your silos, being more flexible. And IT needs to be an enabler of business to be able to respond to the business and move fast. So does that make sense? Absolutely, yeah. I mean, really, at the end of the day, you just want to be able to show your product to your customers, right? And any roadblocks, any bottlenecks that you have need to be addressed. We are not in the inflexible, traditional IT world anymore. Things change from day to day, and there's no reason to not be flexible. Okay, do you have some examples of specific types of applications that kind of VPC and DevOps really make sense for customers? Oh, absolutely. So just think of financial companies. So say you have an investment platform. You need to make changes with the market. You need to make changes as your competitors have other offerings. And these changes may be to the minute. And so your code needs to be flexible to the minute. And so if you can deploy multiple times a day to adjust to the market to customers, current events, things like that, that's the only way you can do business. So if you were to deploy once a quarter, once a month, you would never be able to be competitive in that field. So that's one of the examples that I would say. Okay, so I believe this is your first time at the VTUG. Yes. I guess when we look at events, there's been a lot of changes. When I talked to most people that are in the DevOps, very software development standpoint, meetups seem more popular than even, the user groups are relatively small but happen on a regular cadence. So what are you seeing out there? Where do people go to learn from their peers, get involved? So if we're talking about conferences, meetups, I would say yes, in the DevOps-y world, the meetup, the user group is the core of it. You're going to the city that you live near, so Boston or New York or whatever big city, and you meet once a month, and what you're doing is the community, they don't even have to be big name people, will come in and they'll show you their experience with a tool, an open source tool, or here's how we did this or here's how we did that. And people are making it easier for others to pick up those projects. Another thing that helps a lot is social media. There are people who just live on Twitter and they say, here's my new project, I've posted it on GitHub, please take a look, and so those types of things are big in DevOps, and the bigger conferences by DevOps tools, like DockerCon or AnsibleFest, things like that are very popular in learning to bridge the gap between DevOps. Any final feedback you'd want to give to the community? Sounds like meetups and online, there's a lot of places to get involved. Everybody just make sure they've got a GitHub account and start plugging away on code. Yeah, so I would say if you're interested in pursuing what DevOps can do for you, start using the open source tools. All of the tools that the big companies, the Netflix's and the Disney's and all of the world use are mostly open source tools, and the community in GitHub and Twitter will show you how to use those to be effective, so there's no end to the possibilities. So you made me think of one last question. I think of open source is hugely impossible, but we seem to have this dichotomy of, you talk to most users and if it's build versus buy, I don't necessarily have the understanding, so I'd rather somebody put something together to buy it. Hyperconvergence is a big discussion at this event that can put together pieces. Open source tends to be more of a build it. How do you reconcile those two worlds and where do you see things going? Absolutely, there are plenty of companies like my own ReactiveOps that will do this for you. So we have the knowledge and expertise of all of these open source tools and what we've done is put them into a framework that can get you from A to B because we've had that experience and there's plenty of products out there, ours, RancherOS is one I can think of that will use the same tools, give you the accessibility if you want to then manage it your own, but will get you from A to B faster without you having to have all that knowledge up front. Okay, well, Sarah, really appreciate you taking time. Thanks for sharing with the community and we'll be back with lots more coverage here from the VTUG, thanks for watching.