 Hello everyone and welcome back to day two of AWS re-invent 2021 theCUBE's continuous coverage. My name is Dave Vellante. I'm here with my co-host, David Nicholson. We've got two sets. We had two remote sets prior to the show. We're running all kinds of activities and we've got AWS executives, partners, ecosystem technologists. Scott Weber is here as the director and an AWS partner ambassador from PwC. Scott, it's good to see you. Nice to meet you guys. Thanks for letting me be here. So your expertise is around application modernization. It's a hot theme these days. If you're a company with a lot of legacy debt, you've got a big complex application portfolio. I would think, especially with the four smarts to digital over the last year and a half, two years, now is really a time, probably too late, to really start thinking about rationalizing your portfolio. What are you seeing in the space? Yeah, definitely. We're seeing the customers that have reached that point. I view modernization as sort of the second wave of cloud that's coming. So you had your first wave, the early adopters that lifted and shifted into the cloud. We still have people looking at getting into the cloud. But for those that went early, now they're saying, how do I get more out of the cloud? How do I get closer to cloud native? And that's what we're starting to see around this modernization move, is I want to start to utilize those higher level services from AWS and the cloud providers. I want to get a better return. I want to stop worrying about running infrastructure and hardware. So, when you think about, I go back all the way back to Y2K, that was like a boondoggle for IT to spend a bunch of dough and do some cool stuff. And then of course, the dot com crashed. But today it's different. It's really about the business impact, the business outcome that you can drive and transforming your digital business. So, how do you, as a technology agnostic, you know, consultant, help a company understand what they should leave alone or sunset, what they should aggressively migrate. What's the process that you use to do that? You know, in some ways we go back, we can reuse sort of those six R's that maybe got a customer to the cloud or as they're on that cloud journey, right? And you really want to focus on where can you optimize ROI, right? And you're going to come across those things that you're going to be like, look, it's maybe it's a vendor-caught solution, right? There's not a lot we can do there. You're just going to have to continue down that path unless we can look to move that to a SaaS service. Maybe the vendor's gone to a SaaS offering, right? Or we get into looking at, they've done development in-house. But that development is still monolithic running on virtual machines, either in the data center or in AWS. But it's a critical system to that business. Maybe it's become fragile. How can we now modernize that? Because that's where there's going to be a great return on investment for that customer. And it's also going to allow business agility for those customers, right? As we can get them to microservices and Lambda and function as a service, the blast radius for changes becomes smaller. It allows the customer to move faster in what they're doing. So it's the rationalization becomes what's driving the business forward? What's critical to the business, but what's holding them back as well so that the customers can start to move faster? So it's a formula of, okay, what's the business value of those applications essentially? You can kind of rank that, but then it's a formula of, there's a cost equation. That's pretty straightforward to figure out what the AS is and the 2B. But then there's a speed. Like an ongoing time to value from a developer standpoint. And then I guess there's risk, right? Have you got your core jewels? You maybe don't want to touch those yet. Is that kind of your algorithm? It is, and on that sort of cost and value piece, that's where we can really see some interesting things happen where as we get customers away from license to OSs, proprietary databases, the return on investment can be huge. So we've helped customers migrate from running .NET applications on top of a typical Microsoft Windows stack and SQL server stack all the way taking those workloads all the way either to Linux containers or all the way to serverless. If we're going to take all the steps to rewrite, you can drive 60, 70, 80% of the cost of operating that platform out of it. Then you start this flywheel effect of reinvesting that money back into the next project to help the customer move forward. Quick follow-up, I know you want to jump in. Yeah, yeah, yeah. Why wouldn't a customer that's a Microsoft customer just run that on Azure? Why AWS? I mean, that's a good question and that sort of gets into a lot of philosophical like discussion we've talked about for a long time. The fact of the matter is, the majority of your Windows workloads still run on top of AWS to do today. I would argue AWS has some pretty superior things in their underlying architecture, their Nitro architectures and things like that. But I think it's also choice, right? And the whole move of .NET to Linux, Microsoft started that. They put the ability to, you can run SQL Server on top of Linux. Well, if I run SQL Server on top of Linux, I take out 20% of my cost right there. They put the support in for .NET Core to be able to run on Linux or on containers, but that's to help the developers move faster. That's to help us get to microservices. So that cloud provider choice, I think it becomes a bigger discussion, but a lot of people are choosing AWS because they're not just doing Microsoft workloads, right? And again, we could get very deep into trade-offs on why one over the other, but customers are choosing AWS for a lot of these work. Diversity and better cloud, better infrastructures. Yeah, and philosophical is an interesting way to look at it when it becomes a hostage negotiation. I'm not sure there was a lot of philosophy involved when Server and SQL 2008 were being end-of-support life and people were told, move it to Azure and we'll take care of you, don't move it to Azure, you're on your own. But something on the subject of ROI, ROI is typically measured over time. How do you rectify and address the sort of CIO dilemma, which is that if ROI is being delivered fantastically in four years, but the average tenure of a CIO is 2.7 years, how do you address that? What is the sweet spot for time frames that you're seeing for people to actually implement when you consider, as was mentioned today, the keynote, that somewhere around 15% of IT spend is in cloud today, which leaves 85% of it on premises. So what do we do about that? Yeah, no, that's a great question. So I think I like to get small wins. So find a very big pain point for that customer. How can we start to get them some small wins and start that flywheel effect going of like you save money here, now can we reinvest and start to show some wins? But we've engaged in projects where we've completely rewritten a whole application stack that was the core service for a business in a year and a half. And we took them from a run rate of somewhere between 40 and $60,000 a month. Had they been running that in AWS, they were running it in a data center today. So that was our estimate to less than $5,000 a month to run that application on a serverless platform inside of AWS. So when you talk about modernizing an application environment, that's typically not thought of as low hanging fruit. So does that mean that all the low hanging fruit has been consumed? Are all the net new things that are developed in a cloud native format? Have they already been done? Is this the only frontier for opportunity now? No, no, it's not the only frontier. I mean, there's a lot of customers that are still just trying to get into the cloud. A lot of traffic is out there. Yeah, yeah. And you look at things like mainframe as well, right? That's, I think, a coming area where customers are finally starting to say, you know, enough with the mainframe. We saw it in the keynote today of a new sort of service offering around helping customers rationalize how to start to do things with the mainframe. So, but sometimes you can get those easy wins. Like we find a scalability issue, right? You know, and we can inject scalability and pull back costs very rapidly, right? Because you run in that scenario. The provision for max capacity that may happen 10% of the year, now they're vastly overpaying, right? So we can still get some easy wins with slight tweaks to the platform while we help them rationalize those longer build times. I think the other thing we're starting to see is a shift in CIOs that are coming more from a software background too, right? That aren't from the peer infrastructure background. And as we see those software-based CIOs start to come in, we're starting to understand the gain that can be had of making the investment in the software and those upgrades to the software. And they're 10 years elongating, because, you know, CIO career is over was the joke. Now you're losing CIOs because they're going on to bigger and better. They're getting more options. I mean, they're becoming rock stars again. I want to ask you just as a side about that mainframe compatible runtime that they announced. Because it sounds like you've got some experience in converting the frame. Because I've always been a skeptic. And I, you know, we've seen this movie before where people have to freeze code. They've got to freeze code for 18 months. It takes 24 months or they, but now it's cloud. Adam Salipsky said we can cut migration time, which is critical here by two thirds. Right. Because that's the key. If you can reduce the time in which you have to freeze the code or maybe not even freeze the code. Again, I'm a skeptic, but what are you seeing with practical experience? At PWC we're seeing a lot of customers start down this path. And, you know, the ROI is pretty amazing when once you get in and you really start to dig in of what it can be if to go down this path. And there's a lot of tools out there. There's a gentleman on our team that's a real genius with this. And he's helped multiple customers go down this path. You know, there's tools that can start to do code conversion for you. I mean, we all get a little skeptical on those things because you never know what the machine's going to try to make the code look like. But it's a starting point, right? But there is more. Like a pre-wash. Yeah, there's more and more design patterns coming out to help us down those pathways. But it goes back to agility for the business because a lot of these customers running mainframes today are looking at a six month release cycle if they want to make any changes to their environment, right? If we can get them into an agile mindset to a microservice, they can get to two weeks or less for release cycles. So it's a big win for the company overall. Yes, there's a risk, right? But I think you can take, you can try to de-risk it as much as you can. You don't take the core, the absolute core critical piece of that mainframe. You start to pick away around the edges and you get comfortable with what you're doing. Right. Yeah, and going back to the concept of ROI, specifically in the mainframe space, there have been some not-so-subtle nudges from the marketplace that changed the dynamics associated with staying on your mainframe. Because if I tell you that the tax to stay on your mainframe is going to triple, or quadruple over the next several years, that changes the balance. Yeah. So you have the old guard in the software business who will remain nameless, jacking up the prices because they feel like, you know what? What are you going to do? Right. What are you going to do other than write me a check? And the answer is, we'll move. Right. Right? Yeah, it's reached the point that the companies are moving and what I think companies start to see too is when we talk about purpose-driven databases, Adam was talking about that in the keynote today, too. We've seen that with customers. When we've done builds, what's the right database for this data? And now you can start to get things moving even faster and you unleash new ways of thinking. And yeah, I mean, some of the vendors are doing things like that. Well. And the companies aren't happy about it. Well, yes. But look, we're talking about Oracle in particular. No, well, no. Okay. Yeah, so that's all right. That's one of them. But Oracle invests in its database and it's two different theories, right? Adam, today, I think it's over the right job. APIs and primitives and Oracle takes the kind of Swiss army knife approach, but they do invest. If you have hardcore mission critical, recovery is everything, there's a risk factor involved there. But if you want to go fast and you're a developer, you're not going to necessarily knock on Oracle's door. You're going to go to AWS. But so, but it gets to my question. Having done a lot of TCO analysis, it used to be labor was always two thirds of the cost. With automation, especially in Oracle environments, software license costs are the dominant component. It's maybe less true for SQL server, certainly true for DB2. I remember the early days of the flash, we used to tell customers, install flash, you're going to be able to consolidate, reduce your Oracle licenses when they come up. So that was a preferred strategy. But what are you seeing in terms of the ability, first of all, is that a correct premise that software licenses are still a big component or an increasingly large component? And how do you unshackle from that? Yeah, so definitely software licensing costs for the OSs and for the databases are huge. I mean, there's numbers out there that like for SQL server enterprise, if you can get somebody off the SQL server enterprise and get them to an open solution, like Aurora Postgres or something like, it's a 90% ROI. And the numbers are similar for Oracle, right? And I talked to a lot of customers are like, but we don't know Postgres. But it's not really that different, right? It's still data modeling. And when you get to these managed services platforms like RDS and Aurora, you free up those DBAs to do the higher value things, right? The ROI of a DBA is not managing memory and disk and babysitting the servers. It's helping the developers build better data models, right? And those sorts of things that are higher value. So it is a big thing. And we're seeing customers saying like, help us reduce this licensing cost, right? And help us be more efficient because the open platforms now, especially in the relational database area, are on par in a lot of ways with the Oracle and the SQL servers, right? So then you start to say, well, what am I gaining by paying and being sort of held hostage to these numbers, right? So we definitely see customers making this transition. Yeah, I mean, the point about Postgres is a good one because you're going to get enterprise class recoverability. Now, but even EDB would say, okay, don't start with your mission critical core, pick around the edges, there's what he's saying over and over time, you're going to become more cloud native and get to the point. You know, can you get to that point where everything's cloud native, everything is a service, maybe not a hundred percent, but a large part of your application portfolio can get there, right? Yeah, but you're going to find those, that goes back to doing that application tiering and evaluation and ROI. So, you know, we have a case study that we did with Constellation Brands where they really needed a B2B type ordering portal solution. And they looked at sort of the typical vendors and that, you know, packaged solution, if you will, a COTS type solution. And we proposed doing a full custom solution, soup the nuts and building it natively in AWS. And it was built completely on top of platform services. There was no servers in that environment when we were done. We were using AWS Fargate to run their containers on top of, we were using RDS Postgres, we were using Lambda in some places, we were using DynamoDB for holding in flight orders. And so, the whole environment is deployable from one CloudFormation template. You know, so it completely changed how we even went through the testing of the thing because you ran the same CloudFormation template to deploy to a different environment and you knew you were getting the same exact thing. And so, they went from, they no longer had to worry about securing the underlying compute, secure the containers, run on top of Fargate, use a platform service for your databases and it was a beautiful solution for them. Yeah, you got a taste of that and your eyes open up and say, wow, what's possible. We heard that from NASDAQ this morning. I mean, amazing story. She said our first Amazon bill was 20 bucks. I bet it's higher now, but first hits free kind of thing. But the point is, if people talk about the AWS bill, et cetera, no question, you should try to optimize that. But at the end of the day, it's about the business value, Scott, isn't it? Yeah, it is. Hey, thanks so much for coming to theCUBE. It was great, great perspectives. I always appreciate having you guys on. Yep, thank you very much. Thank you. All right, and keep it right there, Dave Nicholson and I will be right back. You're watching theCUBE's coverage of AWS re-invent 2021.