FinOps on AWS: Accelerated outcomes to maximize ROI

Hello, everyone. Can you hear me? Ok. My name is David Wharton. I'm the Chief Architect for AWS at CDW. Today, I'll be talking about FOPS on AWS - Accelerated Outcomes to Maximize ROI.

I've been moving workloads to AWS for about 12 years. So, um, my focus has primarily been migrations, but in the past four years, I've had a specific focus on cloud cost management and in particular, uh, FInOps in the last couple of years.

So today I'm gonna talk about uh, transparency for cloud financial management as well as some pathways to accelerate your FInOps journey. And I'll do a brief, uh, dive into Committed Use Discounts and I'll circle back to ways you can show your progress with your cloud stakeholders.

So for those who are not familiar with FOPS maturity life cycle, uh, oftentimes you're in a different place and no, it's really about surfacing the data and finding actionable use cases to get that across to your, uh, cloud stakeholders. So you may be out crawl, you may be out walk, you may be out run. It all depends on how sophisticated you are. And there are particular things you need to do to increase your maturity. And as you do that, your cloud consumption will drive down.

So let's talk a little bit about the process. What that looks like you want to put together uh an action plan. First thing you can do is acquire the data from CSPs. In this case, AWS, you can get the data natively from internal and native dashboards. There's also external tooling that you can lean on. You want to perform uh optimizations based on uh the cliche known as low hanging fruit. So that's initial recommendations and you want to start executing those so that you can get quick wins and uh start making progress in terms of unlocking ROI.

So the next step is to view trends, budgets, set weekly reports. You wanna talk about ways that you can view delete redundant infrastructure, zombie boxes, storage optimization opportunities and ultimately through automations, you want to create guardrails that will leverage tooling and get you towards that continuous improvement.

So hands down probably the biggest place you can save uh is reserved instances and savings plans. So fundamentally pay as you go is gonna be a lot more expensive. Uh but if you leverage savings plans reserved instances, you're gonna unlock uh savings as much as 60 to 75% particularly for leveraging things like bring your own licensing. And so it's a balance. Um important thing is understanding what that balance is and what's right for you and your organization, you'll always want to have some element of pay as you go.

And the reason for that is lower environments. Uh in most cases, whether it's a QE environment, a PE environment, you really only need it to be on maybe three days a month, uh maybe one week a month. And it's always going to be cheaper to have mechanisms to power that off through automations through business hour scripting uh than to buy an RI for it. Because if you're buying an RI or you're buying a savings plan for something that's only gonna be on a week or three days a month, it's not gonna be worth it. You're, it's, it's just, that doesn't make sense. There's also tools out there that will help you with your reserved instance, management life cycle.

Uh so obviously, it's, if you buy too many RIs and you don't use them, that's once again, money lost. So you can have an over commitment and the, whether it's ProsperOps or Zesty, uh these are tools that actually can manage the life cycle of your RIs can even uh get you out of RIs that have been uh purchased for like three years. Uh you'll notice that um workloads can actually run cheaper when you upgrade them. And you, you could have bought something for like a three year RI and it only, it ends, it has an end of life in about two years, maybe a year and a half. So you wanna have think of creative ways to get out of those commitments. If you did make that, that big three year commitment tools like Zesty and ProsperOps can help in that way.

There's also a reliable spot. So spot is, it's a way to save, but it's, it's typically not reliable. If you just use it out of the box with AWS, these tools leverage the spot technology and will make it more reliable. You can start to shift not just, you know, lower environments but also with appropriate testing, you can move some of your production workloads over to these platforms as well. And this all works within the AWS ecosystem.

So the most important thing is understanding what those committed use discounts look like and then also how that plays into your IT strategy.

So the biggest savings you can get is implementing an IT roadmap over the course of a year, two years, three years, this is what uh an executed roadmap looks like. So whether it's a task project or initiative, uh you can see here as you go from crawl walk run how as that sophistication increases, the consumption goes down. And this case is a real world scenario. This is a net savings annually of $3 million.

And then as you, you know, go through all these use cases, all these scenarios, it's important to understand the personas across the organization, the stakeholders across the cloud within your organization. Assembling a FInOps team could be an internal team, maybe you're leaning on an external partner. Uh but the important thing is to leverage all your successes, share all your successes and to get all the cloud stakeholders aligned with how important it is to unlock this value. Because sometimes it's not always about saving money, it's about unlocking value for innovation uh especially in an economic downturn. Uh, you wanna find creative ways to unlock your potential to unlock savings so that you can still do the projects that might have been on the cutting room floor.

Uh some of the things that might also help you are rate negotiations with, uh, you know, with AWS. Um, you'll, you can easily do that if you are a proper IAC shop and, uh, I can't go into too much detail there if you have questions about that, please see me after. But, um, that's also a powerful way to unlock some value as well.

So, uh, at CDW, we can help you. Uh, we have three main tiers to, uh, help you in your journey. We kind of beat you where you are. So we have the FInOps Accelerator. That's a 6 to 8 week, uh accelerator. It takes you from crawl to walk or from walk to run. It really depends on where you are. And then if you have teams internally that feel that's, it's too much to take on right away. We have an engineering, a consultative advisory tier for FInOps Engineering and we'll stick, stick with you for a 6 to 12 to 24 month time period to essentially teach you to fish. Help you execute, take you from crawl to run. And at that point, you, you'll know how to do it on your own through the quarterly updates, monthly updates you we have with you, uh, you'll understand the process and how to empower and enable your teams.

And if that's something that's not of interest and it's still too much work, there's always Managed FOPS and that's a continuous engagement through our Manage Services division. And we'll just white glove FInOps for you and continuously save for you also for our CSP customers, our consumption customers. We do have a free cloud cost management tool. It's called InScape. It's got great dashboards, uh, unlocking visibility for you, budget tracking. Uh, we have a whole anomaly component that's very effective. It actually tracks your crawl walk, run progress within the application itself and gives you the cost optimization recommendations that you're looking to achieve.

Uh also at booth 305, we're launching the, uh, Storage Optimizer for AWS. Just wanted to mention that, uh, it's a tool that, uh, is working with, uh, the front end of Slack. It's essentially a Slack bot that looks at your environment and, uh, highlights areas where you can, uh, make improvements uh for your storage and ultimately save you money.

Uh in this case, uh it's, there's like 15 scripts that are running right now constantly uh across your environment. And uh for example, uh I'm not sure who's familiar with multipart uploads, but multipart uploads are uh when you have an automated job that's moving data into AWS, if that job fails, uh let's say 90% of it got into AWS, but then it failed and then you have to retry it. Hopefully it's successful later, but that failed job stays there. So that, you know, gigabytes, terabytes of data that failed, that's there and that needs to be cleared out.

So, uh Storage Optimizer for AWS uh will help find that and then it will also help you select a life cycle policy. Uh, so that those things don't happen again. After seven days, those multipart uploads are then are purged automatically. And that's based on a policy that you would set.