 Live from Vancouver, Canada, it's theCUBE. Covering OpenStack Summit North America 2018, brought to you by Red Hat, the OpenStack Foundation, and its ecosystem partners. Welcome back, I'm Stu Miniman, and you're watching theCUBE's live coverage of OpenStack Summit 2018 in Vancouver. My co-host, John Troyers here, and happy to welcome back to the program Mark Baker, who's a product manager with Canonical. Mark, how's the show treating you so far? The show's been going very well. So we've seen people coming to us on the show floor, coming to the sessions. We're seeing real interest in building scalable production clouds. And so, and coupling that with other container technologies and a lot of other complementary technologies like machine learning. So a lot of the discussion is, can we build clouds? But also much more about the workloads and the kind of integration with parallel, if you like, or adjoining technologies. Great, I want to talk about the customers really, Mark. So as you said, you've been to a few of these shows. Oh yeah, yeah. You know, we've been to a few of these also. And the makeup of the attendees has changed a bit. One of the things I heard, it is 2X the number of cloud architects with their title compared just to last year. Little bit of a broadening into the scope. What are you hearing from customers? What brings them here? What's exciting them in this environment? So, I mean, yes, certainly cloud architects and economical, we regularly talk to cloud architects. And because architecture or the cloud is something that evolves, it's not something that's pinned, right? As workloads evolve and new technologies come along, you need to be able to evolve that architecture. And therefore people that understand that are important. I think it's also noticeable, you know, I'm sat here wearing my blazer. It's noticeable seeing quite a few people around the show wearing blazers. And so you go back in a couple of years ago or even a year or so ago, it was very much a sort of developer-centric type of event. We're seeing more business conversations now and even discussing things such as money and economics, which weren't necessarily conversations that we were getting too heavily in a couple of years ago. Yeah, there's still a bunch of the hoodie set. Oh, for sure. You hear lots of cool T-shirts and, yeah, ironic facial hair and the like. So, maybe from your standpoint at Canonical, talk a little bit about those constituencies of Huda-Selter. We've got the operators, you've got the developers, you've got the, you know, C-Suite. I'm sure the answer is, you know, yes, but, you know, who do you find yourself? Maybe help walk us through some of those roles that you're talking to, some of the biggest concerns that they're having and how you're helping them. So in most enterprises that we're going to talk to, we're typically talking to initially operations, right? Because they know that they need to be able to provide these services to cloud services and container services to their customers internally within the business. And they're looking at, okay, how can we operate this? How can we secure it? How can we scale it in smart ways? And they're looking for our help and assistance doing that. Very soon after that, we'll need to go and talk to developers or engage the line of business developers. Primarily because we need to, this represents change for them. Moving into a cloud or moving their applications to containers represents change. And we want to get them onboarded into this environment and to start to begin that change as quickly as possible. The cloud to succeed, it needs to have many running workloads on it. And so engaging with the developers to take advantage of the capabilities the platform can provide is really important. We'd love to be able to go and talk to at that sea level. And we are starting to have more of those conversations. But I think the type of infrastructure that OpenStack and Container Technologies provides, the initial interest is very much coming from those operators, from the architects, and from the developers. Well, let's talk about operators for a minute. I mean, once upon a time, there was a tribe of people called Sysadmins. They were kind of surly and they took care of things like Linux, right? And now out of that Linux framework, there's a huge set of technologies that have grown all based on Linux, all that Canonical works with. And there's a new set of skills required. Can you talk a little bit about what the new operator needs to know and how you can help train people? And Canonical can help train people? And if you're an assistant man working with Linux, what different things do I need to care about now in the cloud management world, cloud operator world? Yeah, sure. You're right, it used to be relatively simple. And that you would run a VM or you'd run an application on top of their metal and there'd be certain things you'd need to be able to tweak, to scale it, and to up the performance. But when you're running in, as we say, more agile infrastructure, so whether it's cloud and containers or combinations of both, there are very many different variables and how an application is able to take advantage of the storage or the capabilities that a platform provides means there's many different knobs and dials that you can turn. We tend to be advising right now, people on bringing in services such as CICD, Continuous Integration, Continuous Deployment so that they can start to adopt some of these newer ways of working. Operators now, they need to be much more aware of what the workload characteristics are and how that might behave on a hypervisor or how it might behave within a containerized environment. I just came out of a conversation with a customer, for example, that was asking detailed questions about storage performance. They have applications that require certain levels of storage performance and different types of storage that we can bring to bear in conjunction with an OpenStack which is going to be the appropriate one and how do they segment them? And so it's definitely become more complex but I think through collaboration events like this, we're actually getting much better at being able to provide them with the information and the choices they need to make. Yeah, Mark, speak to us a little bit about the community. OpenStack started heavy users in the community, contributing to the community. How do you see that dynamic playing out today? Well, there's still lots of contribution coming into OpenStack so that's good to see. We are starting to see, as OpenStack has matured, as the marketplace has matured, some of the focus no longer being purely on contributing code but now sharing experiences around operations. And that's starting to move into this area where people use this phrase infrastructure as code to be able to access infrastructure programmatically. I think we're seeing collaboration now in the OpenStack community and the adjacent communities around collaborating on the operations, especially when those operations themselves are encapsulated in code. So very simple things, sounds simple, not necessarily easy to do but being able to upgrade and update in place, how you would sort of suspend a system whilst you perform some maintenance and evacuating the workloads and bringing them back in and those kinds of very common tasks for cloud operators. We saw even just a few years ago how operators would each have their own way of doing it, their own preferred methods and this was generally not so efficient so collaborating on those and sharing best practices is one of the really interesting things to see within this community today. Sure, sure, I think the evolution goes, everybody then starts to write scripts but you all write scripts in your own ways and eventually you have to come up with a framework. Yes. And you all have developed a couple of different frameworks in terms of installation and upgrades and things like that as well. Absolutely and it's one of the things that once the customers start to understand that we've developed a framework around operations, those operations are encapsulated within code and it means that if we have a customer, Deutsche Telecom for example, one of our customers that is understandably very security conscious because they run the telco network, has best practices around the security of their cloud and we're able, when they start to make recommendations or updates to that, we're able to take those and share them with a broad audience and get that sort of collaborative spirit around what's the best way to be able to do this, right? Yeah, so you mentioned security there, any other kind of key pain points what are you hearing out in the marketplace is GDPR, something that a lot of your customers are beating on you and what's the canonical position there? Yeah, absolutely, so GDPR has been a real catalyst for people to look at areas for security that they probably meant to get around to at some point but never had, so. Some people said it's the Y2K of this generation. Yes, exactly, it's definitely a forcing function and so one of the areas that we've seen a lot of activity around and certainly we've committed resources to it within the last couple of months has been around encryption of data at rest. So obviously in the cloud, you're going to have a lot of data that's there with the relevant workloads and some of that regulation, GDPR regulation is about what happens if somebody removes a disk from a server, does that mean that they have access to the data as we start looking at things such as edge cloud? So we've very many clouds close to the customer or close to the edge which don't necessarily have the same data center infrastructure around them. How do we secure the data there, right? So encryption of data but doing it in a way that doesn't require you to manually type passwords in to be able to access them all of the time is not a simple problem and we've spent quite a few resources working out how do we address that, how can we do it in a way that's going to allow it to be dealt with economically and scalably. We've been a lot of talk about open infrastructure in general here at the show and open stack obviously is designed to manage infrastructure but we've already talked about containers here with you in this segment. There's a lot of container news, Kubernetes news, open dev summit going on at the same time. So how do you as a product manager you can't just be worried about one part of the stack how do you and your teams worry about that integration and that unified platform and bringing together these interactions with all these different open source projects. Oh yes, for sure and that's, I mean certainly it's one of the things that Canonical has been cognizant of or working on for quite a long time is a Linux distribution at its heart is really the integration of very many different components from a kernel and libraries and compilers and all the various other pieces that go with that. So understanding how these components plug together that's open stack with containers and open V-switch for the networking and set for storage for example, that's very much part of what we've been doing. We're learning with customers as we go very much that how they want to plug these things together with Kubernetes could be Kubernetes is running alongside open stack, Kubernetes is running on top of open stack. Open stack even running on Kubernetes some of them are looking at. So understanding how they, people want to be able to plug technologies together and we standardize very much on sort of reference architectures of combination of open stack plus Kubernetes as a really simple example. But then as part of our QA process, testing process all the reference architectures that we build with hardware partners and other partners too is ensuring that we're able to deliver that as a standalone products as required but also as effectively solutions together that are fully integrated, fully supportable and they're going to deliver the capability of the customer needs. Okay, first of all the open stack on top of Kubernetes really is that something you'd recommend to customers or is there a specific use case for that? It's not something that we recommend today. So there's been certainly a lot of discussion in the open stack community around the control plane and what's the best way to deliver the control plane. Canonical made a very strategic or specific choice several years ago that actually containerizing the services is the right way to do this. So we containerize basically all of the control plane services apart from Neutron Gateway which can be a little tricky to do that but we containerize all of those services and it gives us flexibility when we want to perform updates and migrate services between different systems for example. How do you manage those containerized services though? There's lots of diversity of opinion. Some people want to be able to do that with Kubernetes and that's great and we certainly track those efforts and work with those people if they're using Ubuntu or some of our technologies but I think it's still yet to be decided what's the best way to be able to do this. Yeah so you have an interesting job as a product manager you always want to productize and in general standardize as much as possible and in these communities you have the diversity of opinions oh I'll take this piece, I'll get rid of the core, I'll do something over here, I'll flip it upside down. How do you balance that giving customers choice that making sure you can deliver solid offerings that you can support? And so I mean that's very much it, it's a choice we can say look we can deliver a robust high performing cloud with these reference architectures we've learned that through experience with customers and working with our partners we understand customers all believe they're special and they all have their own special requirements often with good valid reason and so but we'll always try and start from a base and then say let's start to iterate through that adding in additional capabilities or maybe tweaking something for your particular use case if you do that and see how it impacts the cloud because for us to be successful of us the OpenStack community to be successful we need to ensure that those clouds can live and breathe and evolve over time and if they're making too many or too heavy a customization of that cloud then it can start to impact their ability to do that. So we're all for that choice. Speaking a little bit on the line of standardized services I'm really intrigued by managed OpenStack from Canonical can you talk a little bit about what customers it's right for and when it comes into the conversation and then we're in the life cycle because I guess then it can also eventually go away as the control gets handed back over to the customers when they're finished with managed. Absolutely so we started providing what we call BootStack as a fully managed OpenStack service primarily to address the skills gap within the OpenStack community. So we saw a lot of companies interested in deploying OpenStack a lot of enterprises looking for OpenStack but they couldn't find the talent or the people with the experience of deploying a managing OpenStack. There weren't the people around right hiring was hard so and that was becoming a blocker for us to be able to deliver clouds to those customers and so we started to offer a managed service. We had a lot of those reference architectures and best practices pretty well nailed down but it was a facilitator for them to get up and running with the cloud and it's a point where they became comfortable in operating it, managing themselves, hand control back. We've seen that as a very popular model and that period where they're having us manage it can be six months or 12 months or 18 months but the customers know they have the reassurance that they can take it back control in house that they can operate it themselves and they can manage their own environment. They become self-sufficient but they're not doing that from day one. We're holding their hand and taking them along that path. So that's been a very popular offer. All right, Mark Baker really appreciate you giving us an update on really the broad spectrum of customer use cases and all the updates from Canonical. For John Troyer, I'm Stu Miniman back with more coverage here from the OpenStack Summit 2018 in Vancouver. Thanks for watching theCUBE.