 From the CUBE studios in Palo Alto in Boston, connecting with thought leaders all around the world, this is a CUBE conversation. You know, businesses have a staggering number of options today to support mission critical applications and much of the world's mission critical data happens to live on converged infrastructure. Converged infrastructure is really designed to support the most demanding workloads. Words like resilience, performance, scalability, recoverability, et cetera. Those are the attributes that define converged infrastructure. Now, with COVID-19, the digital transformation mandate, as we all know, has been accelerated and buyers are demanding more from their infrastructure in particular converged infrastructure. Hi everybody, this is Dave Vellante and welcome to this power panel where we're going to explore converged infrastructure and look at its past, its present and its future. And we're going to explore several things, the origins of converged infrastructure, why CI even come about, and what its historic role been in terms of supporting mission critical applications. We're going to look at modernizing workloads. What are the opportunities and the risks in what's converged infrastructure's role in that regard? How has converged infrastructure evolved and how will it support cloud and multi-cloud and ultimately what's the future of converged infrastructure look like? And to examine these issues, we have three great guests, Trey Layton this year, he's the senior vice president for converged infrastructure and software engineering and architecture at Dell Technologies. And he's joined by Ewe Kim Zetterblad, who's the director of SAP, the SAP practice for AMIA at Dell Technologies and our very own Stu Miniman, Stu, it's a senior analyst at Wikibon. Guys, great to see you all. Welcome to theCUBE, thanks for coming on. Thanks for having us. Trey, I'm going to start with you. Take us back to the early days of converged infrastructure. Why was it even formed? Why was it created? Well, if you look back just over a decade ago, a lot of organizations were deploying virtualized environments. Everyone was consolidated on virtualization. A lot of technologies were emerging to enhance that virtualization outcome, meaning acceleration capabilities and storage arrays, networking. And there was a lot of complexity in integrating all of those underlying infrastructure technologies into a solution that would work reliably. You almost had to have a PhD in all of the best practices of many different companies' integrations. And so we decided as Dell EMC Dell Technologies to invest heavily in this area of manufacturing best practices and packaging them so that customers could acquire those technologies in an already integrated, fully regression tested architecture that could sustain virtually any type of workload that a company would run. And candidly, that packaging, that rigor around testing produced a highly reliable product that customers now rely on heavily to operationalize greater efficiencies and run their most critical applications that power their business and ultimately the world economy. Now, Stu, of course, you were there. I was as well at the early days of the original announcement of CIWO. Looking back and sort of bringing it forward, Stu, what was the business impact of converged infrastructure? Yeah, well, David Davis, Trey was talking about, it was that wave of virtualization that had gone from just supporting many applications to being able to support all of your applications. And especially if you talk about those high value business mission critical applications, you wanna make sure that you've got a reliable foundation. What the Dell tech team has done for years is make sure that they fully understand the life cycle of testing that needs to happen and you don't need to worry about what integration testing you need to do, looking at support matrices and doing a lot of your own sandbox testing, which for the most part was what enterprises needed to do. You said, okay, I get the gear, I load the virtualization and then I have to tweak everything to figure out how my application works. The business impact, David, you wanna spend more time focusing on the business not having to turn all the dials and worry about, do I get the performance I need? Does it have the reliability and uptime that we need? And especially if we're talking about those business critical applications, of course, these are the ones that are running 24 by seven. And if they go down, my business goes down with. Yeah, and of course, one of the other major themes we saw with converged infrastructure was really attacking the IT labor problem. You had separate compute or server teams, storage teams, networking teams. They oftentimes weren't talking together. So there was a lot of inefficiency that converged infrastructure was designed to attack. But I want to come to the SAP expert. Youa Kim, that's really your wheelhouse. What is it about converged infrastructure that makes it suitable for SAP applications specifically? Well, you know, if you look at a classic SAP client today, there's really three major transformational waves that all SAP customers are faced with today. It's the move to S4HANA, the introduction of this new platform which needs to happen before 2027. It's the introduction of a multi-cloud or cloud operating model. And last but not least, it is the introduction of new digitization or intelligent technologies such as IoT, machine learning or artificial intelligence. And that drove to the need of a platform that could address all these three transformational waves. It came with a lot of complexity, increased cost, increased risk. And what CI did so uniquely was to provide that edge to core to cloud strategy, fully certified for both HANA, non-HANA workloads, for the classical, analytical and transactional workloads as well as the new modernization technologies such as IoT, machine learning, big data and analytics. And that created a huge momentum for converged in our SAP accounts. So, Tre, I want to go to you because you're sort of the deep technical expert here. You and Kim just mentioned uniqueness. So what are the unique characteristics of converged infrastructure that really make it suitable for handling the most demanding workloads? Well, it converges infrastructure by definition is the integration of an external storage array with a highly optimized compute platform. And when we build best practices around integrating those technologies together, we essentially package optimizations that allow a customer to increase the quantity of users that are accessing those workloads or the applications that are driving database access in such a way where you can predictably understand consumption and utilization in your environment. Those packaged integrations are kind of like, you know, I have a friend that owns a race car shop and he has all kinds of expertise to build cars, but he has a vehicle that he buys as his daily driver. The customizations that they've created to build race cars are great for the race cars that go on the track, but for running, building a car on his own, it didn't make any sense. And so what customers found was the ability to acquire a packaged infrastructure with all these infrastructure optimizations where we package these best practices that gave customers a reliable, predictable and fully supported integration. So they didn't have to spend 20 hour support calls trying to discover and figure out what particular customization that they had employed for their application that put some issue that they needed to troubleshoot and solve. This became a standard out of the box integration that the best and the brightest packaged so that customers could consume it at scale. So Yuhakim, I want to ask you, let's take the sort of application view. Let's sort of flip the picture a little bit and come at it from that prism. How, if you think about like core business applications, how have they evolved over the better part of the last decade and specifically with regard to the mission critical processes? So what we're seeing in the process industry and in the industry of mission critical applications is that they have gone from being very monolithical systems where we literally saw a single ERP components such as R3 or ECC, whereas today customers are faced with a landscape of multiple components. Many of them work in both on and off premise. There are multi-cloud strategies in place. And as we mentioned before, with the introduction of new IoT technologies, we see that there is a flow of information of data that requires a whole new set of infrastructure of components of tools to make these new processes happen. And of course, the focus in the end of the day is all on business outcomes. So what industries and companies doesn't want to do is to focus all their time in making sure that these new technologies are working together, but really focusing on how can I make an impact? How can I start to work in a better way with my clients? So the focus on business outcome, the focus on integrating multiple systems into a single consolidated approach has become so much more important, which is why the modernization of the underlying infrastructure is absolutely key without consolidation, without a simplification of the management and orchestration, and without a cloud-enabled platform, you won't get there. So Stu, that's key, what you and Kim just said in terms of modernizing the applications, being able to manage them, not as one big monolith, but integration with other key systems. So what are the options? Wikibon has done some research on this, but what are the options for modernizing workloads, whether it's on-prem or off-prem, and what are some of the trade-offs there? Yeah, so Dave, first of all, one of the biggest challenges out there is you don't just want to lift and shift. If anybody's read research from Wikibon, David Florida, Dave, for the 10 years I've been part of it, talks about the challenges if you just talk about migrating because while it sounds simple, we understand that there are individual customizations that every customer's made, so you might get part of the way there, but there's often the challenges that we'll get in the way that could cause failure. And as we talked about for, especially your mission critical applications, those are the ones that you can't have downtime. So absolutely, customers are reevaluating their application portfolio. There's a lot of things to look at. First of all, if you can, certain things can be moved to SaaS. You've seen certain segments of the market. Absolutely, SaaS can be a preferred methodology if you can go there. One of the biggest hurdles for SaaS of course is there's retraining of the workforce. Certain applications, they will be embracing of that because they can take advantage of new features, get to be able to use that wherever they are. But in other cases there are, the SaaS doesn't have the capability or it doesn't fit into the workflow of the business. The cloud operating model, something we've been talking about at Wikibon, Dave, for many years, we have seen rapid maturation of what originally was called private cloud, but really was just virtualization plus with a little bit of a manager player on top. Now, much of the automation that you build in, AI technologies, you know, Trey's got a whole team working on things that if you talk to his team, it sounds very similar to what you, the same conversation she would have to the cloud providers. So cloud as an operating model, not a destination is what we're going for and being able to take advantage of automation and the like. Where your application sits, absolutely some consideration and what we've talked about, Dave, the governance, the security, the reliability, the performance are all reasons why being able to keep things under my environment with an infrastructure that I have control over is absolutely one of the reasons why I might keep things more along a converged infrastructure rather than just saying to go through the challenge of migration and optimizing and changing to something in a more of a cloud-needed methodology. What about technical debt, Trey? What, people talk about technical debt as a bad thing. What is technical debt? Why do I want to avoid it? And how can I avoid it? And specifically, I know throwing a lot of questions at you, but what is it about converged infrastructure and its capabilities that help me avoid that technical debt? Well, it's the interesting thing. When you deploy an environment to support a mission-critical application, you have to make a lot of implementation decisions. Some of those decisions may take you down a path that may have a finite life and that once you reach the life of expectancy of that particular configuration, you now have debt that you have to reconcile. You have to change that architecture, that configuration. And so what we do with converged infrastructure is we dedicate an entire product management organization, a team of engineers that treat the integrations of the architecture as releases. And we think long range about how do we avoid not having to change the underlying architecture? And one of the greatest testaments to this is in our converged infrastructure products over the last 11 years, we've only saw two major architectural changes while supporting generational changes and underlying infrastructure capabilities well beyond when we first started. So the converged infrastructure approach is about how do we build an architecture that allows you to avoid those dead-end pathways in those integration decisions that you would normally have to make on your own. You, Kim, I wanted to ask you, you mentioned monolithic applications before, that sort of we're evolving beyond that without application architectures, but there's still a lot of monoliths out there. So, and a lot of customers want to modernize those applications and workloads. What in your view, what are you seeing as the best path and the best practice for modernizing some of those monolithic workloads? Yeah, so Dave, as clients today are trying to build the new intelligent enterprise, which is one of SAP's leading guidance today, they need to start to look at how to integrate all these different systems and applications that we talked about before into the common business process framework that they have. So consolidating workloads from big data to HANA, non-HANA systems, cloud, non-cloud applications into a single framework is an absolute key to that modernization strategy. The second thing which I also mentioned before is to take a new grip around orchestration and management. We know that as customers seek this intelligent approach with both analytical data as well as experience and transactional data, we must look for new ways to orchestrate and manage those application workloads and data flows. And this is where we slowly, slowly enter into the world of a enterprise data strategy, right? And that's, again, where converged as a very important play to play in order to build these next generation platforms that can both consolidate, simplify, and at the same time, enable us to work in a cloud enabled fashion with that cloud operating model that most of our clients seek today. So, Stu, why can't I just shove all this stuff into the public cloud and call it a day? Yeah, well, Dave, we've seen some people that have a cloud-first strategy and often those are the same companies that are quickly doing what we call repatriation. I bristle a little bit when I hear these because often it's I've gone to the cloud without understanding how I take advantage of it, not understanding the full financial ramifications, what I'm going to need to do, and therefore they quickly go back to a world that they understand. So, cloud is not a silver bullet. We understand in technology, Dave, things are complicated. There's all the organization operational pieces they do. There are excellent cloud services and it's really, it's innovation. How do I take advantage of the data that I have? How do I allow my application to move forward and respond to the business? And really, that is not something that only happens in the public clouds. If I can take advantage of infrastructure that gets me along that journey to more of a cloud model, I get the business results. So, automation and APIs and everything in the ops movement are not something that are only in the public clouds but something that we should be embracing holistically and absolutely that ties into where today and tomorrow's converged infrastructure are going. Yeah, and to me it comes down to the business case too. I mean, you have to look at the risk reward, the risk of changing something that's actually working for your business versus what the payback is going to be. You know, if it ain't broke, don't fix it, but you may want to update it, change the oil every now and then, you know, maybe prune some dead wood and modernize it. But, Tre, I want to come back to you. Let's take a look at some of the options that customers have. And there are a lot of options, as I said at the top. You got do-it-yourself, you got hyper-converged infrastructure, of course, converged infrastructure. What are you seeing as the use case for each of these deployment options? So, and build your own. We're really talking about an organization that has the expertise in-house to understand the integration standards that they need to deploy to support their environment. And candidly, there are a lot of customers that have very unique application requirements that are very much customized to their environment and they've invested in the expertise to be able to sustain that on an ongoing basis. And that build your own is great for those folks. The next in converged infrastructure, we're really talking about an external storage array with applications that need to use data services native to a storage array and self-select compute for scaling that compute for their particular need and owning that three-tiered architecture and its associated integration, but not having to sustain it because it's converged. There are enormous number of applications out there that benefit from that. I think the third one is you talk about hyper-converged. I'll go back to when we first introduced our hyper-converged product to the market, which is now leading the industry for quite some time in VxRail, we had always said that customers will consume hyper-converged and converge for different use cases and different applications. The maturity of hyper-converged has come to the point where you can run virtually any application that you would like on it. And this comes down to really two vectors of consideration. One, am I going to run hyper-converged versus converged based on my operational preference? Hyper-converged incorporates software-fine storage, predominantly a compute operating plane. Converged, as mentioned previously, uses that external storage array, has some type of systems fabric and dedicated compute resources with access into those. Though your operational preference is one aspect of it, and then having applications that need the data services of an external storage, primary storage array, are the other aspect of deciding whether those two things are there are needed in your particular environment. We find more and more customers out there that have an investment of both, not one versus the other. That's not to say that there aren't customers that only have one, they exist, but a majority of customers have both. So, you, Kim, I want to come back to the sort of attributes of the app from the application requirement perspective. When you think about mission critical, you think about availability, scale, recoverability, data protection. One of you could talk a little bit about those attributes. And again, what is it about converged infrastructure that is the best fit and the right strategic fit for supporting those demanding applications and workloads? When it comes to SAP, we're talking about clients and customers' most mission critical data and information and applications. And hence, the requirements on the underlying infrastructure is absolutely on the very top of what the IT organization needs to deliver. This is why when we talk about SAP, the requirements for high availability, protection, disaster recovery is very, very high. And it doesn't only involve a single system. As mentioned before, SAP is not a standalone application, but rather a landscape of systems that needs to be kept consistent. And that's what a CI platform does so well. It can consolidate workloads, whether it's big data or the transactional standard workloads of SAP or PECC, the converged platforms are able to put the very highest of availability, protection standards into this whole landscape and making a really unique platform for CI workloads. And at the same time, it enables our customers to accelerate those modernization journeys into things such as ML, AI, IoT, even blockchain scenarios where we build out our capabilities to accelerate these implementations with the help of the underlying CI platforms at the rest of the SAP environment. Got it. Stu, I want to go to you. You had mentioned before the cloud operating model and something that we've been talking about for a long time in Wikibon. So can converged infrastructure substantially mimic that cloud operating model and how so? What are the key ingredients of being able to create that experience on-prem? Yeah, well, Dave, as we've watched for more than the last decade, the cloud has looked more and more like some of the traditional enterprise things that we've looked for and the infrastructure in private clouds have gone more and more cloud-like and embrace that model. So I think back to the early days, Dave, we talked about how cloud was supposed to just be simple. If you look at deploying in the cloud today, it is not simple at all. There are so many choices out there, way more than I had in the physical data center. In the same way, I think the original converged infrastructure from Dell, if you look at the V block, the criticism was, oh, you can have it in any color you want as long as it's black, just like the Ford Model T, but it was that simplicity and consistency that helped build out most of what we were talking about of the cloud model is I wanted to know that I had a reliable substrate, a platform to build on top of it. But if you talk about Dave today and in the future, what do we want? First of all, I need that operating model in a multi-cloud world. So we look at the environments that can spread beyond just a single cloud because customers today have multiple environments, absolutely hybrid is a big piece of that. You look at what VMware's doing, look at Microsoft, Red Hat, even Amazon are extended beyond just a cloud and going into hybrid and multi-cloud models. Automation, a critical piece of that and we've seen great leaves and bounds in the last couple of generations of what's happening in CI to take advantage of automation because we know we've gone beyond what humans can just manage themselves and therefore true automation is helping along this environment. So yes, absolutely Dave, the lines and blurred between what the private cloud and the public cloud, and it's just that overall cloud operating model and helping customers to deal with their data and their applications regardless of where it lives. Well, you know, Trey in the early days of cloud and converged infrastructure, that homogeneity that Stu was talking about any color as long as it's black, that was actually an advantage to removing labor costs, that consistency and that standardization. But I'm interested in how CI has evolved, it's added in optionality. I mean, you again was just talking about blockchain, so all kinds of new services, but how has CI evolved in the better part of the last decade and what are some of the most recent innovations that people should be thinking about or aware of? So I think the underlying experience of CI has remained relatively constant and we talk about the experience that customers get. So if you just look at the data that we've analyzed for over a decade now, you know, one of the data points that I love is 99% of our customers who buy CI say they have virtually no downtime anymore. And that's a great testament. 84% of our customers say that their IT operations run more efficiently. The reality around how we delivered that in the past was through services and humans performing these integrations and the upkeep associated with the sustaining of the architecture. What we've focused on at Dell Technologies is really bringing technologies that allow us to automate those human integrations and best practices in such a way where they can become more repeatable and consumable by more customers. We don't have to have as many services folks deploying these systems as we did in the past because we're using software intelligence to embed that human knowledge that we used to rely on individuals exclusively for. So that's one of the aspects of the architecture. And then just taking advantage of all the new technologies that we've seen introduce over the last several years from all flash architectures and NVMe, on the horizon, NVMe over fabric, all of these things will, as we orchestrate them in software, will enable them to be more consumable by the average everyday customer. Therefore it becomes more economical for them to deploy infrastructure on premises to support mission critical applications. So Stu, what about cloud and multi-cloud? How does CI support that? Where do those fit in? Are they relevant? Yeah, Dave. So absolutely, as I was talking about before, customers have hybrid and multi-cloud environments and managing across these environments are pretty important. If I look at the Dell family, obviously they're leveraging heavily VMware as the virtualization layer and VMware has been moving heavily as to how to support containerized and Kubernetes environments and extend their management to not only what's happening in the data center, but into the cloud environment with VMware cloud. So management in a multi-cloud world, Dave, is one of those areas that we definitely have some work to do. Something we've looked at Wikibon for the last few years is how will multi-cloud be different than multi-vendor because that was not something that the industry had done a great job of solving in the past. Customers are looking to take advantage of the innovation, where it is in the services and the data-first architecture is something that we see and therefore that will bring them to many services in many places. I was talking before about in the early days of CI and even a lot of organizations, some organizations anyway, there's still these sort of silos of storage, networking, compute resources. And you think about DevOps, where does DevOps fit into this whole equation? Maybe Stu, you could take a stab at it and anybody else who wants to chime in. Yeah, so Dave, great, great point there. So when we talk about those silos, DevOps is one of those movements to really help be a unifying force to help customers move faster. And so therefore the development team and the operation team are working together, things like security are not a bolt-in, but something that can happen along the entire path. More recent addition to the DevOps movement also is something like FinOps. So how do we make sure that we're not just having finance sign off on things and look back every quarter, but in real time understand how we're architecting things, especially in the cloud, so that we're being responsible for that model. So the speed is one of the most important pieces for business and therefore the DevOps movement, helping customers move faster and leverage and get value out of their infrastructure, their applications and their data. Yeah, I would add to this that I think the big transition for organizations, because I've seen it in developing my own organization, is getting IT operators to think programmatically instead of configuration based use a tool to configure a device. Think about how do we create programmatic instruction that interacts with all of the devices that creates that cloud like adaptation feeds in application level signaling to adapt and change the underlying configuration of that infrastructure to better run the application without relying upon an IT operator, a human to make a change. That this sort of thinking programmatically is I think one of the biggest obstacles that the industry face. And I feel really good about how we've attacked it, but there is a transformation within that dialogue that every organization is going to navigate through at their own pace. Yeah, infrastructure is code, automation or this is a fundamental to digital transformation. Joachim, I wonder if you could give us some insight as you talk to SAP customers in Europe, across Amia. How does the pandemic change this? I think that pandemic has accelerated some of the movements that we already saw in the SAP world. There is obviously a force for making sure that we get our financial budgets in shape and that we don't overspend on our cost levels. And therefore it's going to be very important to see how we can manage all these new revenue generating projects that IT organizations and business organizations have planned around new customer experience initiatives, new supply chain optimization. They know that they need to invest in these projects to stay competitive and to gain new competitive edge. And where CI plays an important part is in order to, first of all, keep costs down in all these projects, make sure to deliver a standardized common platform upon which all these projects can be introduced and then of course making sure that availability and risks are kept high versus at a minimum, right? Risk low and availability at a record high because we need to stay on par with our clients and their demands. So I think, again, CI is going to play a very important role as we see customers go through this pandemic situation and needing to put pressure on both innovation and cost control at the same time. And this is where also our new upcoming data strategies will play a really important part as we need to leverage the data we have in a better, smarter and more efficient way. Got it. Okay, guys, we're running out of time, but Tre, I wonder if you could break out your telescope or your crystal ball, give us some visibility into the futures of converged infrastructure. What should we be expecting? So if you look at the last release of this last technology that we released in Power One, it was all about automation. We'll build on that platform to integrate other converged capabilities. So if you look at the converged systems market, hyper-converged is very much an element of that. And I think that where we're trending to is recognizing that we can deliver an architecture that has hyper-converged and converged attributes all in a single architecture and then dial up the degrees of automation to create more adaptations for different type of application workloads, not your traditional three-tier application workloads, but also those microservices-based applications that one may historically think, maybe it's best to run that off-premises. We feel very confident that we're delivering platforms out there today that can run more economically on-premises, provide better security, better data governance, and a lot of the adaptations, the enhancements, the optimizations that we'll deliver in our converged platforms of the future about colliding new infrastructure models together and introducing more levels of automation to have greater adaptations for applications that are running on them. Got it. Trey, we're going to give you the last word. If you're an architect, a large organization, you've got some mission-critical workloads that you're really trying to protect. What's the takeaway? What's the really the advice that you would give those folks thinking about the sort of near and mid-term and even long-term? My advice is to understand that there are many options. We sell a lot of independent component technologies and data centers that run. Every organization's environment around the world. We sell packaged outcomes in hyper-converged and converged. And a lot of companies buy a little bit of build-your-own. They buy some converged and they buy some hyper-converged. I would employ everyone, especially in this climate, to really evaluate the packaged offerings and understand how they can benefit their environment. And we recognize that there's not one hammer and everything is a nail. That's why we have this broad portfolio of products that are designed to be utilized in the most efficient manners for those customers who are consuming our technologies. And converged and hyper-converged are merely another way to simplify the ongoing challenges that organizations have in managing their data estate and all of the technologies that they're consuming at a rapid pace in concert with the investments that they're also making off-premises. So this is very much, the technologies that we talked today are very much things that organizations should research and investigate and utilize where they best fit in their organization. Awesome, guys. And of course, there's a lot of information on dell.com about that. Wikibon.com has written a lot about this and many, many sources of information out there. Trey, Uakim, Stu, thanks so much for the conversation. Really, meaty, a lot of substance, really appreciate your time. Thank you. Thank you, guys. Thank you, David. And thank you everybody for watching. This is Dave Vellante for theCUBE and we'll see you next time.