>> Woman: From theCUBE studios in Palo Alto and Boston, connecting with thought leaders all around the world, This is a CUBE conversation. >> Hello and welcome to today's session of the 2021 AWS Global Public Sector awards for the award of best migration solution. I'm your host Natalie Erlich and now we're joined by very special guests. We have Omer Enaam, application modernization leader at Deloitte Consulting and Bart Mason, technology lead for the Office of Recovery Services at the Utah Department of Human Services. Welcome, gentlemen. Good to have you on the show. >> Thank you. >> Thank you for having us. >> Well, terrific. I'd love to hear more about your migration from mainframe to AWS. Bart. Let's start with you. >> The state of Utah has a mainframe system and we have our child support application that was first developed in 1996 on the mainframe written in COBOL. The application served us well through the 24 years that we had it running on the mainframe. The issue was that the mainframe, it was getting difficult to find people who knew how to program in COBOL. But the biggest problems were any type of modernization. We were pretty much stuck to using what are called green screens, and there was no real easy way to do any type of modernization. And a lot of our applications that were public-facing or employee-facing, a lot of those web applications had to be written in a separate system and set up to connect and talk to the mainframe system. So it was a system that served us well but it was time to try and figure out what are we going to do about this? Because the mainframe was expensive and it was old technology that didn't let us advance to where we wanted to go in the future. So roughly about 2016, we started to investigate what are the possible ways that we can migrate our child support application off the mainframe. And we went through discussion such as a complete rewrite where we would start from the very beginning and rewrite our child support application. The child support application is a case management and an accounting system. And if we would have done a total rewrite we were told it would be upwards of $200 million to do a complete rewrite. We started looking at other possibilities and came across one possibility, and that is to do a migration off of the mainframe into the cloud. It would be a pre-session where we could do a lift and shift and basically take the code, change it into Java, and put it into the cloud running in EC2 instances. So it was an, we called it an intermediate step to modernization because it would get us one step to where we need to do, or where we need to go. And for modernization, it helps us to, since the program that it was, or the language it was migrated to was Java it made it so that we could do modernization. And we decided that if we did a lift and shift from the mainframe to AWS, that we could modernize at our own pace, we could modernize screen by screen or function by function. So it gave us the ability to control roll-outs and getting our application to where we needed to be. >> Terrific. And Omer, I'd love it if you could weigh in as well. What were, what was the support that you provided towards this migration? >> Yeah, of course. So as Bart pointed out, the state was looking for a approach that had high chance of success, high probability of user adoption with minimal impact to the organization. At the same time, have the ability to for the state to maintain and modernize at their own pace. So we work with Bart and explain to him a few options. And one of the options was using a automated coding data conversion approach where we take legacy programming languages like COBOL and convert them into Java. Just like translating the code from one language to another. And in the process, we guarantee that your your new system will work exactly. It will be functionally equal of what you do currently. And at the same time, it minimizes the risk. And it also allows the state to have no issues with their business continuity and additional training for their staff. So in a nutshell, we brought in a solution demonstrated to Bart and team and they bought into that, the idea that this is exactly what they want to do as a first step. And as we speak, we are working with the state to help them take that system in the cloud to the next level. Now we have unlocked the potential of digital transformation. Bart can build mobile apps in front of that application. That the state can. There are new analytics capabilities for that their employees can be more productive in providing services to the citizen. They can implement native capabilities from AWS to implement a process automation, implement some artificial intelligence-based tools to optimize the processes and make life easy and better for the employees, at the same time more importantly, serve the citizens in a better way. >> Mhm. And Bart I'd love it If you could share some further details on some of the considerations that you had such as risk and whether it could be used later in the future. >> The biggest thing, the biggest risk to us was that if we, as we migrated off the mainframe, there's a risk that we have to recertify our system with the Office of Child Support Enforcement in Washington, DC. When we build a system, the child support system, we're required to have them come in and do a assessment of our application and certify that it is an application that can be used for child support. If we would have done a rebuild from scratch, the risk would be that first a rebuild, from what we've seen can take anywhere from five to 10 years. I've already touched on how expensive it is, but it takes up to five or what we've seen, up to 10 years to do a complete rewrite. And the risk for us was that if we did a complete rewrite, we would still be on the mainframe for quite a long time. And we would have to have our system recertified with OCSE. And that can take anywhere from five to 10 years for a recertification too, so the risk was that if we did anything with the complete rewrites it would be several, several years going through rewrites and recertifications to get our system up and running in AWS. And the other problem would be that taking that amount of time would also, it would bring us probably not up to date with the current technologies as we did our rewrite because we'd be focused on rewriting that application and not taking advantages of services and applications that come up and can help us with our rewrites. So one of the biggest risks was that we'd have to do recertification with OCSE, With the migration, coming off the migration because it is a one for one migration where it went from COBOL to Java, we did not have to do a recertification. This allowed us to move the application as is and it functioned the exact same way that recertification was not a problem for us. OCSE said that, told us that it was not a risk or an issue that we'd have to take on. So the biggest risk was recertification for us but with the migration and moving into the cloud we went through their security processes and we came out without any big issues coming out of that. >> Fantastic. Thank you. Omer. I'd love to go to you now. What are some of the unique benefits of working with AWS? >> Sure. I think the biggest benefit is there, the extensive services that are available and having the the proven platform where you cut down your operational costs drastically. So comparing the mainframe costs with the Amazon cloud costs. Clearly the state has benefited a lot from the from a savings standpoint, infrastructure savings standpoint, and at the same time now, as I said, the the system is in the cloud, running on open architecture in the Java programming language, The AWS cloud provides us several capabilities natively which allows the state to use, to digitally transform the experience for the citizens and employees by implementing modern DevOps practices for for managing the, operating the system providing new capabilities to workers and supervisors for analytics to business process automation, having better call center integration capabilities and so forth. So there are endless opportunities. And the state is in the process of executing on a prioritized list Just before the pandemic hit, we worked with the state to lay out the future for their system and for their organization in the form of a one day innovation lab, where major stakeholders from the state gathered with Deloitte and we worked through a prioritization process and determined how we can take this system to the next level and really digitally transform the system and in the process, provide new services and better services to state employees and the citizens. >> Yeah. Terrific insight there. Now Bart, I'd like to shift it to you, asking the same question. What are your thoughts on working with AWS? Why choose them for this? >> We always have been looking at moving a lot of our applications into the cloud. We've been looking at that for several years. The advantages of moving to AWS is, from my point of view, and state's point of view, is that AWS provides a lot of services and it provides the capability for us to do a lot more for our applications. So for example, when we were on the mainframe, one of the biggest problems that we had was disaster recovery. We had a disaster recovery site in the Southern end of our states with another mainframe that we would sync up with our application. The problem was that we have over a hundred data connections. We connect to banks, external entities, internal entities. We have different types of connections. We have to do printing. We have to print checks and several things. Disaster recovery on the mainframe was something that we were never really capable of doing. We could get our application up and running but it just sat on the mainframe. We had no data connections, all that was extremely difficult and extremely expensive to do for disaster recovery on a mainframe and on alternate sites. Moving to AWS, one of the biggest things for us was that disaster recovery requirement. Because now that we're in AWS, it makes it more easier for us to spin up servers once servers go down, restore servers when they go down. We have all of our data connections in one location, and as systems become unrecoverable or have issues, it's easy for us to spin up another one or several in their place, or even our data connection, because they're all located in one place and we're using them all of the time. So disaster recovery was one of the big key components for us. The other component was that, as we modernize our application, we're looking at what AWS services are out there to help us with modernization. We're looking at services such as AWS Batch to replace our batch system. We're looking at databases to replace the current database that we're using. We're looking at using containers to containerize our applications and our ORSIS application, and also microservices. So moving off the mainframe was the first step and putting it all into servers on an EC2 instance. But then we look and say, okay, how can we do this and make this more modern and run better and more efficient? And then we started looking at all the AWS services that are out there, that run outside of an EC2 instance, for example. And we see that there's an endless possibility, and endless capabilities that we have at our fingertips to say, okay, we're off the mainframe less modernize by moving to Batch or let's start looking at containers and things like that to help us with our applications. So disaster recovery and the available services that we can move to to help us with our applications, what we look at. >> Well, thank you both so much for your insights, Bart Mason, Utah Department of Human Services as well as Omer Enaam, Deloitte Consulting and LLP. I'm your host for theCUBE. Thanks so much for watching. (outro music)

>> Narrator: Live from Las Vegas. It's the Cube. Covering VMWorld 2017, brought to you by VMWare and it's ecosystem partners. >> Okay welcome back everyone live here at VMWorld 2017 behind us we got the stage here set on the VMVillage, a lot of people hanging out, I'm John Furrier with Dave Alante our next guest is Vish Muichand who's the Senior Director of Product Manager HPE, Cube alumni Eric Burgener, Research Director at IDC. Guys welcome to the Cube. >> Thanks very much John. >> Vish, lot of storage action going on VMWare, you see Vsan, the cloud's here, true private cloud report from Wikibonds off the charts, showing a huge growth on prem, cloud operations, storage is impacted. What's the dots that we're connecting here this week? What's the storage story this week? >> So clearly there's a lot of different things happening in the marketplace right, different modes of operation and that in itself is demanding different approaches to infrastructure. So I think what you are seeing in the industry a variety of different approaches in storage, right? Whether it's external storage, whether it's software-defined storage, whether it's hyperconversions, or that's all flash storage. All of these things are coming together and trying to respond to the needs of data and how you want to process that data. >> We've been talking with, we talk to you guys a lot on the Cube, HP Discover, and we always say software's eating the world, we just heard Sanjay Punin from VMWare talking about it, he likes to drop that soundbyte. We take it one step further. He's a Harvard MBA, we got the bapsen mojo here. We say if software's eating the world, then data's eating software. So you guys have had a software core competence and you mentioned data. What is the impact and compromise, more and more data comes in from the edge, this primary, this secondary storage, this backup this data protection, it seems to be like this melting pot of changing architectures. How are you guys handling that at HP? >> Filling software is a very key element because it provides you with those capabilities, right? To really deal with the logical instantiation of assets and in this very virtualized world, this very dynamic world right now, gone are the days where you can do hardware type desegregation. Software gives you that speed, that agility, it gives you that flexibility. Gives you the changeability to move quickly. >> Eric you're at IDC you guys, this is your job. You guys track the market share, you guys have the pulse it's like keeping track of the baseball game. What inning, how are the Red Sox doing? Are they in first place are the Yankees catching up? What is the current state of the server virtualization because you know certainly the game's changing a little bit the world's going to cloud. What are you guys seeing in your research? >> Well so obviously most mainstream computing is running on virtualization, whether that's in the cloud or that's on prem. There's very little physical infrastructure left. There is still some of that but clearly that is not the future, virtualization is the future. >> So I wonder if I may, so you're saying virtualization is the future, so I wonder if we can unpack that a little bit because the theme here is cloud and everything is cloud related. Is your feeling, Eric, that that's sort of over your skis marketing, getting ahead of where the customer really is, I wonder if you could sort of elaborate. >> I think what the customers are really looking for is an easier way to do their jobs for less cost. And cloud provides that flexibility that you don't necessarily get if you're managing your own on-premise infrastructure, that's not 100% true based on some scale issues, but by and large, I think that's really what cloud brings to the table is a different payment model, and a flexibility that you wouldn't necessarily have with on prem infrastructure. >> So what are you guys seeing, do you feel as though the on-prem infrastructure leaders like HP, there are others obviously, are going to be able to bring that cloud-like simplicity to what do you call private cloud or whatever on-prem, is that happening, how fast is it happening, is it viable? >> Yeah so I absolutely think that's happening, in fact that's one of the reasons why software-defined storage is growing so fast is those type of products give you the kind of agility that you would normally get from a cloud environment and if you're running that on prem and you've implemented the right infrastructure around it then you're getting many of those same kind of benefits. Now you're paying for that hardware and software in a different manner than you do for the cloud, but you're getting many of those IT agility benefits that you might otherwise get from the cloud. >> And Dave, you know HP's tagline is Making Hybrid IT Simple right and so our point of view is that there is both on premise and off premise, just depending on what the usage models are and what the problems you're trying to solve, right. And bringing that simplicity where you may be going from a 100% on premise to maybe 20% off, but we've also seen some people at 50% off premise trying to come back a little bit on premise, right? So both directions I think are very very key. >> Is your point of view and I want Eric if you could chime in as well, from HPE's perspective, is hybrid IT sort of horses for courses in other words, workloads on prem versus workloads off prem, or is it beyond that some kind of federation model? >> So we see three key use cases. The first is of course wholesale, applications running on the cloud. Office 365, the perfect example of that, Sharepoint, Dropbox right, that's one. Then there is what I would call disaster recovery as a service, where you may want to have your third site in the cloud even though you got two sites on premise. Then there's also the third use case or in archiving that says how do I archive a portion of my data maybe into the cloud so it is online, but I don't have to manage it and I don't have to maybe deal with some of the associated costs around it. So these are the three sort of cases I see. >> Dave: Okay, what are you seeing in the customer base, Eric? >> Well so I completely agree that hybrid cloud is the way data centers are going to be built going forward. There are reasons to keep certain workloads on prem, generally there's performance, security or some kind of regulatory requirements that might make you put workloads on prem versus putting them in the cloud. It also depends on how often you're using the data. So Vish mentioned archive use cases. So that's a case where you need a lot of storage capacity that you keep for a long time but you may not necessarily be accessing it that much. If you're going to be accessing data a lot, that's another reason why you might consider bringing it on prem, as opposed to leaving it off prem. And of course the access, the costing access models that you get from people like Amazon and Azure are going to impact where you draw the line on that. >> So is there a difference between multi-cloud, I got a bunch of different clouds in my organization, I'm going to choose where to put stuff and cross-cloud sometimes you call it inter-clouding was, I like that term. >> Vish: You could dual source your cloud. >> And either dual source or federate or actually split application work. >> So I have seen several different aspects of that. So a customer has said to me that they need to move 20% of their data off premise, to do that they need two cloud vendors, and to get to two cloud vendors they need to see four or five of them so they can narrow it down and they they says okay, HPE all of the data that I have today is in your premise or with your equipment, how are you helping us broker that kind of arrangement. What are you doing to help federate some of that data? And work with some of these cloud vendors. So I think that's an interesting customer ask. >> Okay, well there's also cost consideration because if you multi-source or you have the opportunity to multi-source, you've got a competitive environment that's going to drive lower costs for you. As opposed to if you just got one choice. The other issue there is data mobility. If I'm locked into cloud vendor one, and it's very difficult, there's major switching costs to move, then that's another reason that might offset the potential price advantage I get from being able to go to any vendor. So there's a lot of vendors out there now, infrastructure vendors that are talking about making it easier to move data on prem to off prem, into different clouds from cloud to cloud and I think that's something that creates a more level playing field that really is going to ultimately result in lower costs. >> That's a great point about the costs, we'll just double down a quick question on that. Where are customers tripping over themselves in terms of total cost of ownership because what you're getting at here is hidden costs, right in plain sight. What are those trip fault wires if you will? What's the pitfalls what should they be looking for? >> Well, so I'll give you a general answer to that, but I think that it's very specific to workload type and the regulatory requirements that you're in but I'll tell ya one of the cases where we see repatriation, workloads moving from the cloud back into on prem is when you get to a certain level of scale. And the largest enterprises. >> John: Scale in terms of when to bring it back? >> Well just in terms of how >> or when to leave >> So how much data do I need to basically maintain in this environment and use on a regular basis. And the larger scale environments are the one where larger enterprises are able to actually bring back, create their own cloud infrastructure on prem, with their own environments and actually manage that for less cost than what they could otherwise pay a public cloud provider. >> So just to take it one step further, connect the next dot, the CXO, the CIO has to try to get some stability and there's some uncontrollable things certainly in retail it's predictable that the holiday season needs bursting or whatever so you do some things in the cloud but that's a known pattern, so you're saying that they're starting to recognize some of these scale issues for predictability they bring them on prem. Is that kind of what I'm getting? >> Well so the scale from a cost point of view, so if you're creating your own private cloud infrastructure and you're using the same kind of highly agile software to find storage designs to build that environment, you somewhat have the same ability to burst. Now yeah, you have to buy the hardware and there's redeployment issues and hopefully when we move forward towards much more composable infrastructure that becomes a lot easier problem to solve but that's you know some years in the future. But what I'm really talking about it's the cost. If I'm going to be maintaining a five petabyte data set over a ten year period, and I know what my access patterns are, is it cheaper to put that in Amazon or is it cheaper for me to build an infrastructure in house and maintain that myself. >> That's a great point. That's huge and Vish what's your reaction, is this basically validates all the action going on on the private cloud right now, on prem activity is setting up the cloud models. They can't do that unless you have the operating model. >> I'll talk about two things right, one called Cloud Bank and another one called Nimble Cloud Volumes and soon to be called HPE Cloud Volumes. So Cloud Bank allows you to take on premise data running on a three part array, and actually take a portion of that data onto either an on premise object store or an off premise object store. And we call that Cloud Bank working together with something called Recovery Managed Central and store once bringing that cloud picture together. Now the HPE cloud volumes on Nimble Cloud Volumes, it's another interesting concept where you have a cloud service that's block storage service, but it gives you the six nines SLA, it gives you the ability to do snapshots and transform data without a lot of charges that Eric talked about. It gives you the ability to move the data to different clouds because it's disagregated from the major cloud providers, it's connected via a close proximity connection so these are just two examples I think that show you how putting these used cases into action. >> Hey can we geek out a little bit here? (laughter) >> Aren't we geeking out now? You want to go deeper? >> So people want simplicity, we know that, we're talking about bringing cloud on prem. How do they get there? Well one of the ways is VVOLs, we sort of been talking about this, they haven't really taken off. Eric you've written some content around this. Like you said off camera, customers don't wake up in the morning and say I got to get me some VVOLs. But they do want simplicity. >> Absolutely, yeah. >> What are VVOLs, why do they matter, and how does it relate to simplicity. >> So yeah, let's talk a little bit about that. So what everybody no matter whether they're putting storage in the cloud, they're building on prem, they're building a private cloud, everybody wants to be able to manage their environments more easily, more intuitively, and one of the things that we've seen as a trend over the last five years is in general across the industry, storage mangement tasks are migrating away from dedicated storage admin teams, more towards IT generalists. In many cases, those are the virtual administrators. To enable that kind of a move, you need to make storage much easier to manage. So the whole idea behind VVOLs is to basically allow a non-storage person who maybe thinks about things in terms of I'd like to do this operation to an application for example, I've got Oracle running or I've got this file system here and I want to create a snapshot of it or I want to do some other task on it. To be able to just select it at the application level and perform that operation, that's very intuitive, it's easy for a non-storage person to understand and VVOLs effectively enables that kind of an ease of use management in block based environments. >> An application view of the storage? >> That's right, and I mean it's effectively it ties storage operations to a single virtual machine, and basically you're running an app on a virtual machine and so that's how you get that tie in in that way. But one other thing I'll say about VVOLs is that so it's not just what VMWare provides, there's some work that needs to be done on the storage array side to integrate with that management framework. And then how that vendor has chosen to integrate with that framework is going to determine the functionality that you have access to when you're using that VVOLs API. >> And how have you chosen to integrate with that framework? >> Yeah so Dave if you look at VVOLs, both HPE and HPE 3Par nimble have bene very very strongly focused on VVOLs in fact we've been working with VMWare gosh over the last five years now, on the reference architecture for VVOLs. Most recently we've now introduced replication support for both 3Parand nimble platforms with VVOLs and I think that capability now within VVOLs is a very important watershed capability because everybody needs resilience, disaster recovery. >> Automation's right around the corner, orchestration all big topics here at VMWorld. >> Correct and so that's a very key piece. And I think if you look at to Eric's point around simplicity, VVOLs is one key area. Two layers maybe I'd like to highlight as well. Number one is the visibility to what the application sees and within the Nimble community, they've talked about this app data gap, which is the applications not knowing why they can't get access to data and so this notion of bringing that level of understanding visibility to that gap saying is it in your computer infrastructure, is it in storage, is it in the network? So this notion of VMVision, Infosight, the Nimble (inaudible) because you're going to bring out the rest of the HPE portfolio I think is very key around simplicity. The third thing let's not forget, VMWare's built a whole ecosystem of management platforms around V-Center, V-Realize operations, all the orchestration and operation pieces and so continuing to integrate and offer customers that view is very key, right, so three prong vector I would say on making things simple. >> Also it gives HPE discovers coming up in Madrid shortly. Congratulations good to see you, Eric thanks so much for stopping by and sharing the IDC perspective. Great job, live coverage here at VMWorld 2017, I'm John Furrier, Dave Alante we'll be right back with more live coverage after this short break. >> Thank you.