 Everybody, welcome. Thanks for coming to another edition of the OpenShift Commons briefings. We host these every Wednesday, and we hope you bookmark it on your calendar. Come back and see us every Wednesday at noon. Today, we're lucky enough to have Michael Ferranti, the vice president of product marketing here talking from Portworx, and he's going to be talking about enterprise data protection for OpenShift applications. Welcome, Michael, to the show. Hi, Michael. Great to be here. Great. So tell us about what you do there. Yeah, it's kind of a loaded question. I don't know if some of your audience saw recently that we were just acquired by Pure Storage, so we announced that last week. I was really about that on the internet just the other day. Congratulations. That's a pretty good thing for you, folks, I take. Yeah, we're all really excited, and I'm sure we'll get into a lot of those details as we go along, because I think it really relates to kind of our approach to the topic of today's conversation, enterprise data protection. All of that's to say, kind of, you know, there's a lot of change. I've been with the company for three years now, and when I joined, I was, you know, maybe the second or third kind of non-engineer hire, and, you know, we've since, you know, doubled the team many times over and are a much larger organization now, but going to be going into an even bigger organization. So, you know, I think the way that I spend my time is a combination of, you know, talking to customers, understanding, you know, how they're approaching digital transformation within their organization broadly, you know, how we translate that into, you know, requirements and solutions for solving the data problem that comes along with application development and application operations, and then making sure that when someone comes to, you know, our website or visits at a trade show or attends one of our webinars, that kind of how we solve the specific pain that they have is really, really clear. And, you know, in a startup, you know, you end up wearing a lot of hats to do that, but that's kind of my mission. And you were at another start. So, you've been there for about three years now. You said where were you before that? You were at another startup. Were they also purchased? I know. I know. And unfortunately not. So, I'm not batting a hundred on this one, but, you know, it's, you know, if you don't mind, I can kind of take a little bit of detour, because I think it'll put a lot of the conversation in context. So, yeah, so I was at another startup prior to Portworx. I was a company by the name of Cluster HQ. And I spent about three years at that company as well. In fact, kind of as part of the acquisition by Pure, I recently wrote kind of a blog post reflection on the last six years in that journey and kind of what I learned along the way. And to give the kind of the high level review, I would say that first of all, when I joined the previous company, Cluster HQ, I had been working in kind of a, you know, a publicly traded company for the last five years and really enjoyed it, but kind of got the itch to join a startup. And for me, joining a startup, it's really about building something from scratch, right? Not just being on a winning team, but helping kind of get the team going so that it can win and really build things from scratch. And as I was looking around at kind of what started to join, you know, do I want to do something around, you know, a SaaS company or do I want to do something around, you know, I don't know, you know, neural networks that we're calling at the time, now we'd call AI. I kind of, you know, just did a broad search and realized or noticed that everybody in San Francisco at the time was just like Docker, Docker, Docker, Docker. Docker was the conversation at every infrastructure meetup anywhere. And as I learned more about it, I realized that, you know, containers, yeah, that's kind of virtualization 2.0, but in, with that building block, that virtualization base building block, it was really changing how applications were being built, deployed and managed. And I saw that as kind of a long-term trend. And so I said, okay, if this really becomes like if the data center, and even the public cloud becomes predominantly containerized, what does that mean for the other parts of the application stack? What does it mean for infrastructure? What does it mean for security? And I realized that, you know, kind of storage and data management writ large would need to be completely redone, reimagined, rebuilt, if containerization was going to become the primary operating model. And go ahead. I was just going to say something, you know, you brought up Docker and, you know, I've been here at Red Hat for 19 years now and pretty happy about that, a great company to work for. But it seems like every three years, there's a new, there's a new golden child out there. Remember when OpenStack first came out and it was OpenStack, OpenStack, OpenStack. And then there was, then there was the whole Docker craze. And then there was, what was it, you know, DockerCon was the new, you know, place to be at. And then that kind of dwindled. And then there was containers and containers, containers, containers. Now Kubernetes, Kubernetes is now the, you know, orchestration of containers. So how have you, you know, seen those changes as it relates to storage over the last, you know, just X number of years? Yeah, yeah. Excellent question. And, you know, I, so not to make this all about kind of, you know, my recent past, but, you know, six years ago when I joined this startup, I, I believe that if containers were going to continue to be important, someone would have to figure out the storage and data management. And it turns out that maybe like five or six people agreed with me and everybody else said, you know what, you know, Michael, we love you. But if you think that containers are going to run databases and, you know, analytics and other things like that, then you've completely missed the point of all of these Docker meetups, right? Containers are for stateless applications. And to answer your very pointed question, the evolution of storage and data management for containers was one of pretty much no one caring about it and not even beyond not caring being antagonistic towards the idea that we needed a data layer for containers to begin with through to realization that, you know, and we have lots of open ship customers that are in this boat. You know, the agility that they gain from being able to run applications on, on OpenShift is so monumental for them. And I'm kind of, I'm not even really being hyperbolic when I, when I say that it's really a game changer when you have a team that used to take six months to get a new application into, into production. And now with OpenShift, they can squeeze that down to, you know, weeks or days or in sometimes hours. That's such a big change that you say, how can we apply this to our entire infrastructure? And if, if there's a limiting factor, which is to say, you can get that goodness as long as there's no data involved, then it's really a deal breaker because how many enterprise applications don't have, don't have data. And customers began to realize that. Then the vendors, you know, big and small realize that they needed a solution there. And, you know, I was lucky enough to, to work for two companies that were there kind of six years ago. Portworx was founded in parallel with this other company. But, you know, none, you know, the EMCs of the world were among the first, but it probably took them, you know, a good maybe 18 to 24 months to really understand that containers was something that could challenge their business and that app was even longer. And now we kind of come full circle and, you know, the biggest companies in the world who are using containers are also running analytics, AI, you know, big data workloads on, on containers in production. And wouldn't you say, though, that because things are changing so rapidly and because containers offer so much flexibility and choice that, that this poses new types of challenges for customers about orchestration and managing all of that in a production environment? Yeah, absolutely. You know, the playbook hasn't been written yet. You know, I think it's, it's not like everything is unknown at this point. But there are, you know, still lots of, lots of unknowns. And oftentimes it's maybe the, maybe there are fewer technical unknowns today, but there are a lot of cultural unknowns. For instance, and a lot has been written and said about this is, you know, you used to have very rigorous separation of, of duties between developers and operations. And, you know, I think, you know, Red Hat was kind of one of the companies that really championed the system administrator, right, your, your Linux admin. And that was an operational role. And now we kind of run this DevOps world in which, you know, the, the administrator oftentimes writes software themselves. Now it could just be software in order to automate the deployment of other software. But in some cases, it's, you know, it's everything from kind of lightweight coding to very, very heavy coding when you're building large scale kind of platforms. And that cultural change for how do you operate in a company that's, you know, you know, one of our customers, I'm just going to pick on them because it, I think it illustrates the point here, they're one of the largest retailers in the United States. You know, they've been in business for, you know, over 100 years. And how do you take a company that's been around for so long, kind of started out not as a digital business really as a brick and mortar, and then transform that into a company capable of delivering software to compete with true digital native companies. When you're also transforming your culture, I think you're exactly right. Lots of questions get posed. And to you, you know, as customers are starting to adopt multi cloud and moving everything out of the data center, we see, I think we tend to see a lot of customers aren't necessarily moving their workloads into production right away. And I think probably a lot of that is about concerns around, you know, data protection. So, you know, how, you know, how do you see protecting applications the way we used to do with like, say virtual machines, now that everything's running in containers? Yeah, that's a good question. And I agree with you that, you know, the, you know, if you're adopting containers or Kubernetes or cloud native or whatever you want to call it, you know, there's a there's a learning period before those applications move into production, as there would be with any on any technology stack. And I think if you look at the industry overall, lots of organizations like you imagine each organization is on on kind of an escalator. And everybody gets on at the same spot. But some people get on before other people. So, you know, at any given point in time, even within the same organization, you have some some customers or some teams that are are in production where you have other teams that are starting the beginning of the journey. And so that's kind of one of the things you have to deal with as well, which is that there's different levels of maturity, even within the same organization. That said, I think when we talk to customers about data protection, we, you know, a lot of a lot of things are new, but a lot of things are not new. And so customers understand concepts like, you know, RPO and RTO when it comes to data protection and just for, you know, clarity sake, RPO stands for recovery point objective. That's, you know, how much data loss are you willing to accept in the case of a disaster? Right? Does if if a right just happened in one data center, does that right need to be available in your in your disaster site? Or are you say, you know, it's okay if we have a backup from an hour ago, that's okay. That's acceptable amount of data loss and different types of applications have different requirements. In fact, you know, in the same organization, you can have some with very, very stringent requirements and others that don't. And RTO is recovery time objective. How long is it? How long should I recover? So we start with the understand customers understanding, right? They they have their RPO and their RTO for applications defined. And they've done that for VMs. And like, you know, the chief security officer, the chief compliance officer, you know, head of IT, like all of these things have been debated endlessly by the organization. And none of that needs to change just because you're moving in containers. What changes is how do you deliver upon those SLAs? And when it comes to Kubernetes and and OpenShift, one of the most important things to understand is that in an OpenShift application is made up of a number of different components. And all of those components need to be available in order for an application to run post disaster or to be recovered from a backup. And those components are defined at the application level, not at the infrastructure level. And that's the big difference between kind of traditional data protection, where we would say we would back up our VM, right? We would, you know, we would use, I don't know, Veeam and that would back up all of our all of our VM ware instances. And because we ran one application per VM, if we had a backup of the VM, we also had a backup of the application. That model completely breaks down when it comes to containerization, because now my OpenShift application is actually running across many VMs, or in the case of bare metal, it's running across many bare metal instances. I'm running distributed applications. And because OpenShift enables us to have kind of radical multi-tenancy, right, to get those efficiency gains, not only as a single application running across multiple hosts, but each of those hosts is hosting multiple applications. So I've kind of got this, you know, this scatter plot of pods, configuration, volumes spread across my data center. And if I want to take a backup or recover an application, I need to know kind of the needle in the hay stack for a specific tag or for a specific namespace. And the reality is that the biggest challenge with data protection when it comes to moving to containers is that your existing solutions that were optimized for VM workloads don't speak the language of Kubernetes and OpenShift. So you can't give them, for instance, a namespace and have them understand what that means and then go and protect it, or give them a tag and then they go and protect it. Now you folks have a Red Hat certified operator, right? Yes. So we've been working with you folks around, you know, getting a container up and running available for customers so they know that, you know, the internal components of that container are, you know, Red Hat blessed code as well as, you know, supportable code from your side. How does the operator help people manage their data in this cloud native distributed environment? Yeah. So the operator concept is basically software for managing other software. And so there are a lot of things that you need to do in order to kind of execute these operations. So for instance, I want to take a backup. I might need to snapshot a distributed database. Maybe it's Cassandra and I have 10 Cassandra nodes and I need to snapshot them, but I need to do it in a way that actually runs a command within the Cassandra container, each Cassandra container before that snapshot is taken to kind of queues the database, flush any rights to disk, then take my snapshot, then unlock everything, and then run a command within, Cassandra called node to repair, which kind of resyncs any rights that we're pending. Yeah. Basically it's a complex operation that's something like an operator can do for you. So we have a rule within our backup engine, which is executed via an operator, which is a set of pre, a pre hook, pre snapshot rule, and a post snapshot rule. So all you have to do is say for my custom database, you need to run this command from within the database container before the snapshot, and then you need to run this command after. It's the type of thing that a traditional database administrator would do manually. We've codified that into code, you give us the rule, and then our operator is going to execute that command, and it doesn't matter if it's, you know, if it's on one pod, if it's on 100 pods, if it's once a day or if it's every every five minutes, we're going to be able to execute that faithfully and without kind of the human error that you can get from using someone needing a second cup of coffee, you know, and you know, fat fingering a key or something like that. So the operator concept doesn't, the operator concept does not provide data protection. It enables software that does provide data protection to do so in an automated manner and speeds up efficiency and reduces errors. Hmm. So how about, what about customers? Like what are some of the coolest customer cases you've seen out there today? Yeah, it's, you know, there are a bunch and unfortunately the kind of, you know, the whole COVID thing has really put into stark contrast kind of, you know, companies that had been able to cope with just the unprecedented demand in those that have struggled, and we're lucky to work with some very kind of forward-thinking customers that have been down this path for a while and it just, you know, proved to be the prescient decision when kind of the, you know, when basically the world change starting in, you know, February or March. So a couple of examples come to mind. One is a platform called Roblox. So it's a gaming platform. I think they have like 150 million monthly active users, mainly kind of teens and preteens. And so you can build your own games. It's a community. You can make money off of your game, like by selling games to other people. Kind of a DIY gaming network. And, you know, they've been a longtime customer and they talk about being the business of keeping kids happy and with the whole COVID thing. They're just usage as skyrocketed as kind of families try to figure out, yeah, how to deal with, you know, kids in the house when parents are working on all these types of things. And so, you know, they really had to grow what they call their points of presence, which is where they stamp out these infrastructures all around the world in order to meet the demand that's popping up. And so that's a really cool use case. Being able to support that. Another customer that comes to mind is called Esri. So they're one of these companies that, like, you know, we all rely on every day, but we've never heard of. That's the reason mapping data, right? Exactly. And they're becoming more of a household name because some of their maps around COVID tracking are kind of, you know, often used by, you know, New York Times, Washington Post, et cetera. So just to keep track of infection rates and testing and things like that using their ArcGIS platform. So you've probably been to websites where it's ArcGIS.esri.com. And they're running all of that in the public cloud and, you know, data is a huge part of what they do. So it has to be reliable. It has to be performant. It has to be secure. And we're helping there. And then, you know, those are kind of the, not necessarily consumer, but things that consumers might be more familiar with. You know, then we have customers that are on the opposite side of the equation, you know, financial services, pharmaceutical, where they're, you know, adopting containers and cloud native technologies like OpenShift in order just to make their business processes more efficient. And there, what's cool is how you take a really heavily regulated financial services company like Royal Bank of Canada. We recently did a webinar with them and say, how do you both meet your kind of regulatory requirements of just being, you know, a financial services institution and be as agile as, you know, Silicon Valley startup when it comes to getting to market quickly. Like, that's a hard needle to thread. And so being able to help their, you know, solving things like disaster recovery, multi-site DR for them is really exciting and kind of points to the future, I think of, you know, there are fewer and fewer of applications that are really too sensitive, you know, too, too regulated, too critical to running containers. I think, you know, we're seeing every day more and more instances where that's proving not to be the case. So what about the cloud providers? It seems like, you know, when I started at Red Hat, again, it was 2000, 2002, there were 40 different flavors of Linux out there. Everybody was in the Linux space. There was Blue Dog Linux and Miracle Linux and China had Red Flag Linux and, you know, Red Hat was the company that got it right, right? We figured out how to slow down the release cycles of the kernel releases so that customers could actually standardize on one distribution of Linux and then software vendors could certify their software on a target that wasn't moving all the time. So we went from Red Hat Linux, which was, you know, version six, version seven, version eight, version nine, you know, and we came out then with, you know, a version of Linux that was released every three years before the ABI or KABI would change. Now with Kubernetes, it's kind of the similar thing, like everybody's in the Kubernetes space. Everybody's got a different story around Kubernetes, especially, you know, like the major cloud providers. So who do you think is going to get it right? Who do you think is going to be the red hat of the Kubernetes cloud space? Yeah, it's an excellent question. And I think if I categorically do the answer, I'd be on eTrade right now, you know, buying some stock because I think the person who gets it right is going to really create a very hard to compete with business for a long time. Our CEO, Murley Thermorelli, has actually been doing a series of blog posts on this topic because it's really, it's both intellectually interesting and then also important to get right from a partnership perspective. And so I don't know that, you know, he knows the exact answer or we know the exact answer, but I think I have a model for thinking through it. And basically, I think that whereas Amazon pretty much got as close to a winner-take-all market as you can in enterprise software, right, if they also didn't compete against a lot of their customers like retail and things like that, I think it probably could have been a winner-take-all market, right, because if Walmart had decided to just go ahead and use Amazon because Amazon wasn't, you know, the retailer, then that would have taken some really big users away from Google and Azure made it harder for them to get the scale that they needed. Anyway, Amazon was the runaway winner in the cloud 1.0 award. At the same time, VMware, I think is still smarting from their business being sucked away because of cloud. And so you have a market now where the Google and Azure and to a lesser extent IBM, but I would call IBM the fourth big cloud where they're trying to figure out how do we prevent Amazon from running away with cloud 2.0. And that's where you see, you know, Microsoft with Arc, Google with Anthos, IBM with Red Hat slash OpenShift as a infrastructure agnostic management layer on top of public cloud infrastructure. So if you're Google, you're thinking about how do we make it so that Amazon's strength, their infrastructure play actually not a liability, but how do we commoditize that by bringing Anthos in, running on top of Amazon infrastructure and in Amazon basically has to keep eating margin because as they complete at the infrastructure level but Amazon or Anthos captures or Google captures the higher value software revenue. At the same time, you have got the independent cloud platforms. They're also like VMware with Tanzu and then probably more relevant IBM and Red Hat OpenShift where VMware can do the same thing. Tanzu gives them a credible way to really offer something of high value in the public cloud run Tanzu on a federated cloud architecture of Amazon, Google and Microsoft and used by least lowest com denominator infrastructure, storage network and compute from the cloud provider. But then your kind of your ELA is with Tanzu and I think Red Hat can offer something similar. So I think ultimately there is not going to be only one winner in this market the way that there was a very predominant winner in the public cloud. I don't know if VMware long-term is going to be at a compete. I think each of the clouds will be successful and I think actually, I'm not just saying this because I'm on the OpenShift Commons webinar but I think IBM strategy is really smart and Red Hat is like the entire key to it. OpenShift is the entire key to it. And so if nothing else it's going to be really interesting to watch over the next five years and customers have a lot of choice which is a great thing. So what's next? Meaning like there was open there was the OpenStack shiny object docker sheen already talked about that. Is there going to be another shiny object 24 months from now or do you think we're kind of settled into Kubernetes and container orchestration for the foreseeable future? Well, I think there's always a shiny object and that's a good thing because smart people are getting born every day and they figure out how to solve problems differently. I don't think that at a high... Well, let me say that I don't think there's going to be something in the next couple of years that fundamentally changes application delivery the way Kubernetes slash containers have changed application delivery over the next five years. I think there's going to be innovation within that space but the idea that there's something called X that two years from now everybody's like run away from Kubernetes and start doing that. I don't think that's going to happen in kind of the three to five year timeline. I think the next big thing though is going to be Kubernetes as the control plane for the entire data center not just for containers. And I because I think that that's been one of the things that's been holding back adoption of Kubernetes which is that like you have to buy into containerization or you have to commit is probably a better word to containerization if you want to get those benefits but if you look at someone like Netflix they made microservices popular before Docker was popular and all of that for the most part was on Amazon and all of that for the most part was running in EC2 instances and they're a highly agile organization lots and lots of microservices and yet they weren't a containerized business. So if the idea behind Kubernetes is often said how can we make any business like Netflix or like Google well why are we still saying Kubernetes has to it can only manage container and we're starting to see things like OpenShift virtualization we actually just did a blog on our website about how you can use OpenShift to manage VMs in the same agile manner. And I think once that happens Kuvert is a project that's being popularized a lot of people are looking at ways to kind of get VMware out of their infrastructure because it's not that we don't want virtualization it's that we don't necessarily need people call it the VMware tax like I think that's you know we just don't we don't need the value that VMware is providing for these set of applications and so we want to do it with virtualization but not VMware. I think that happens. I think that might disagree with you though. What's that? I think I think VMware might disagree with you on that. Yeah I think they have they absolutely would and you know it's it's their their argument to lose but you know customers you know customers vote with their with their checkbooks and you can't take your customers for granted not that that's what VMware is doing but I think some customers say hey you know Kuvert plus Kubernetes plus kind of Docker containers enables me to build an agile infrastructure that looks quite similar to what I was getting from vSphere but I can do it in a multi-cloud I can do it in the public cloud and you know they're they see some value there so So what about these challenging times you know like I've been living here at my summer house now since March 20th and when I first got here I left Massachusetts because of the challenging times and I was snowmobiling into my cabin here because the road's not maintained or anything for me my business of partner marketing here at Red Hat has just absolutely exploded like through the like we are so busy how have you seen these data where we are right now affect your business in a in a in a positive or a negative or in an indifferent way Yeah it's um I I I I joke with my wife all the time is like you know whoever you think of being quarantined as like this is like how are you gonna make the time pass it's gonna be so boring you can't do anything and as a family we've never been busier between you know two kids in school both of my my wife and I working virtually and you just you multiply that by every Portworx employee each Portworx employee interacts with you know 100 different customers and you just you keep multiplying it out and everybody's busier and if you look at like the the the things that are making modern life liveable in this very unusual times it all comes back to software and so if you kind of say well what does it mean for software businesses it means that there are a lot more customers they're buying a lot more they're they're running at scale a lot faster um there's a kind of a I saw it on I think it was on Twitter maybe there's some other social media platform but it was or maybe as LinkedIn doesn't matter um like a little um uh kind of mock survey and it said who in your organization has been responsible for digital driving digital transformation and like choice A CTO choice B CIO choice C COVID-19 and like that one's checked because the reality is like COVID it just it was an adapter die moment and so you know customers um like like Esri um like Roblox experiencing growth that they hadn't planned for and yet they had they had to um adapt to or just you know be surpassed zoom is another example not not a poor customer but a clear example of a company that's just unprecedented and um I think all those like zoom we use uh we use blue jeans here but there's there's probably five or six major mainstream ones I can't imagine the the infrastructure demand that was put on those as a result of these last seven or eight months it must be amazing yep yep absolutely and you know you know my kids are learning about role-based access controls as they go through school because they're you know they're authenticated into their school account using um google um using google often all of these things and like they're at you know the tender ages of eight and ten are getting a level of computer literacy that just I think portends of what the future is where every major um kind of life decision and life activity is going to be mediated by computers that doesn't mean that we can't get together anymore but that your things that we used to do with paper you know face to face are increasingly going to go online and all that infrastructure needs to run somewhere it needs to be managed it needs to be secure um and you know to bring it back to your earlier question you know who wins I think it's the companies that can make that easy and reliable for their customers because that's ultimately what what the end user needs is they need it to work right at at any level of scale and you know it's too complicated even for people like you and me we do this for a living it still can be really really complicated to understand how all the pieces fit together and so you take a customer that also has a day job you know doing banking or doing gaming or whatever their their specific area of expertise is they can't understand you know every single piece of minutia and so that's where you know I think red hat really um executed extremely well in the operating system revolution by taking something as complex as Linux and just making it so that it just you plug it in and it works within the enterprise I think application infrastructure um is is another example of that that you know where that need is equally high well I'm looking at the clock we're coming up here at the end of at the end of our our content period here we don't want you to go back after being here today and have your boss or your CEO say Michael you were there you were streaming on the open shift commons briefing show addressing you know people all over the world why didn't you why didn't you say XYZ so let's let's let's make that not happen what is it that you want to make sure that you deliver as a message here while we're while we're here today yeah it's a good question I mean I I would say the the thing that I'm really excited about is um you know by joining pure storage I don't know if the you know the audience knows anything about pure um I'm not going to do justice to their to their story history um but I think the opportunity that pure import works um have in in the context that we've just described is that applications are more important than ever data is more important than ever um and you know applications are running in multiple clouds and lots of different environments and security is an ever-present threat availability is an ever-present threat or requirement and someone has to provide that management layer the data management layer to go to live alongside kind of the the application management layer provided by open shift or you know whatever container platform you're using and I I think the best solution on the market for that is um is pure and and pure works and you know ultimately you know the people listening it doesn't decide but we believe that you know if you are running serious applications on open shift and you care about data protection you care about performance and you care about you know multi cloud operations that you know what works is the the most complete platform for Kubernetes data management services and um I think you know getting a chance to work with innovative customers is what really makes us tick really really passionate passionate about it and so you know looking forward to the next six years to to complement the previous six I think there's a lot of cool things in store going on to the resources tab we took the liberty of putting up a resources link here Michael on how to get started with Portworx you know the the operator that you folks have that you built with us is made available and published and distributed via the red hat catalog we have the link up there and of course open shift commons probably one of the one of the the most exciting things that I've had an opportunity to work with here at red hat we also have a link to it there if anyone would like to join the the commons community here at red hat please feel free to do so Michael thank you so much for joining us today it was really good my pleasure and we'll we'll talk to you soon and thanks for coming great thank you