How to keep cloud costs under control (and what to do if they get away from you)

Once the flywheel of change gets going, your movement to the public cloud will start to accelerate. Your teams will get excited with all the things they can do with the cloud, and early on it will generate a lot of activity as everyone starts to learn and explore with public cloud tools. But there’s another thing that’s going to be new to your organization:

Your monthly cloud bill (cue the scary music!).

Something you’ve never really had to think about before is now going to get the CFO knocking on your door to find out what the heck is going on.

To illustrate, I’ll tell you a true story. A telco we’re working with moved more than 2,000 workloads to the public cloud after making a successful pitch to its board that the company would save an estimated 75% on total cost of ownership (TCO) compared to its on-premise data centers. Once the board approved it, its hyperscaler came in and helped to quickly “lift and shift” the applications to the public cloud. Sounds great, right? But when the work was done, the run rate cost was MORE expensive than the old data center.

As they say in Spain: no bueno.

While the IT group looked around for tools that could help them figure out how to reconfigure their instances to save money, it was the FINANCE TEAM that ended up selecting a cost optimization tool that identifies problems and recommends fixes. However, there was one problem when they put the tool into action: the engineers had to evaluate every single recommendation the tool made to reduce costs and then make a decision on each one—work that was time consuming and took them away from their other tasks.

Sadly, what this telco experienced is an all-too-common problem, but also easily remedied. Luckily, the team came to TelcoDR, and we’re psyched to be helping them out of their jam.

FinOps to the rescue

When you move to the public cloud, costs shift from fixed to variable. And with that, power shifts from finance to individual technologists.

In the old days, with an on-prem data center, when IT team members needed more capacity, they had to write a business case, get budget approvals, work with Finance and go through Procurement to buy the additional machines—a very slow, but highly controlled process. With the public cloud, that’s all changed. Now anyone can access more compute and capacity in seconds. The upside: it’s so easy to experiment! The downside is… no control. When your engineers have access to the public cloud, your technologists will be excited to play with these new tools. If they have an idea, it’s so easy to provision some compute and test it out. But it’s also easy to forget to turn it off when they leave for the weekend, or even vacation.

In the before times, when there was no variable bill coming in, the cost was there as pre-purchased, unused capacity sitting in your on-prem data center. Your organization had already paid for it in the form of excess capacity and spent CapEx. With the public cloud, suddenly you can see exactly when and how much you’re paying. You’re no longer paying for extra capacity when you don’t need it. That’s a win. But you’re going to have to get used to fluctuating costs as the needs change from month to month: a pre-planned product launch or an unexpected event flooding usage. Finance is going to have to learn how to roll with the lower, but variable, cost of the public cloud.

These shifts—from fixed to variable costs, from central to distributed control—are part of the culture change of moving to the public cloud. Finance has to get used to the idea of fluctuating costs. Technologists have to start thinking about the bottom line. IT needs to work with Finance to project and manage budgets, and to do that, it needs to understand the cost of what it’s designing technically. What’s the best solution? What’s the most cost-effective? Sometimes one big machine is cheaper than three small machines, or vice versa. In a recent Telco in 20?podcast with Nanda Kumar, executive director of cloud engineering at Verizon , we talk about this very thing, starting at about eight minutes in.

The solution: recommendation systems vs. action systems

So that’s the challenge. What can a telco do about it? You’re going to have to build a team that focuses on cloud cost optimization. This team needs to work closely with Finance to project spend while also keeping tabs on the hyperscaler’s technology AND pricing announcements to understand how the changes affect the environment—both technically and financially. You’ll need to arm your team with tools to help them scan the cloud environment for fixes to deploy.

If you’re someone who enjoys the cost optimization process, AWS publishes blogs regularly to help customers help themselves. There are a ton of them, which makes it hard to keep up. For example, here’s one on automatic?cost optimization?for Amazon Simple Storage Service ?(S3), another on?migrating your Amazon Elastic Block Store (ESB) volumes , and another on?provisioning performance separately from capacity with Amazon EBS . You can also check out this?new podcast ?from AWS-superfan Rahul Subramaniam. It just launched, and each episode will feature an Amazon product manager sharing tips for using their products and services.

There are consultants ready to help, like?Corey Quinn , chief cloud economist at?The Duckbill Group , a consultancy that specializes in helping companies not only shrink their Amazon Web Services (AWS) bill, but also be smart about their infrastructure. There are other consultants like Accenture, TechMahindra, and Tata that offer services to come in and help.

At TelcoDR, we focus on using great software tools to help us optimize costs for our clients. There are two main categories of software you can purchase: systems of recommendation, which identify cost savings and suggest a fix, and systems of action, which identify cost savings and then implement the fix automatically with your approval. At TelcoDR, we like the latter. Our tool of choice for quick AWS optimizations is?CloudFix , which will deliver 10-20% savings on your AWS bill and takes minutes to set up. And as I mentioned, there are lots of other options, too. They’re more complicated and more expensive, but can deliver 30-50% savings on your cloud bill. TelcoDR can help with AWS, GCP and Azure tuning.

So, when you’re ready to move to the public cloud (or if you’re already there), make sure to include cloud optimizations as part of your cost-management strategy. If you get stuck and need help, don’t hesitate to give me a call. I’ll help you in a jiffy.