The Blog

DuckTools: AWS Cost Optimization

Calendar Icon 12.18.2020
aws-section-divider

Starting from the day The Duckbill Group opened its doors, we’ve been pretty down on SaaS tools that purport to save money on the AWS bill. We’ve had several reasons that were responsible for our distaste:

  • They all charge a staggering amount of money but don’t actually solve anything for you. “Yup, here’s a graph of the money you’re spending split on a new and exciting axis. Please pay me 2% of your bill in perpetuity!”
  • They all give the same unhelpful advice. If you’re using one of them, ask yourself sincerely: what has your AWS bill done since you rolled it out? We both know the answer.
  • They all are building something massive and all-encompassing, 80% of which you’ll never use.

We’ve long believed that the most impactful place for lowering and managing AWS costs is deep into the architecture–where software can’t help–and thus Duckbill has focused on a consulting services model rather than building yet another SaaS tool.

As we went about helping our customers on consulting engagements, our Cloud Economists were simultaneously writing a bunch of Python scripts to address specific challenges of AWS cost management for use with specific problems we kept encountering on our client projects.

The tools that we built became more and more fleshed out, and now it seems that this can help people who aren’t quite ready for our deep-dive consulting services yet.

DuckTools Apps

DuckTools is a suite of tools aimed to solve specific problems. We’re launching with two DuckTools Apps with many more planned. These apps are simple, because you shouldn’t need multiple days of training to understand what they do. They’re highly focused and easier to use. They’re not a panacea; they’re what we use internally on our consulting engagements. And we’re making them available to you.

Oh, and the best part: Unlike pretty much every other option out there, we charge a fixed monthly fee for these apps, regardless of how big or small your AWS bill is. We don’t feel we need to charge you a couple of million bucks a year just because you spun up more EC2 instances than someone else did.

Savings Plan Calculator

One of the biggest frustrations our clients have with Savings Plans is determining and modeling the right commitment level for a Savings Plan purchase. The AWS Savings Plan Recommendation system provides options to bootstrap a recommendation on the basis of the past 7 days, past 30 days, or past 60 days, and then provides an hourly commit to cover the compute for that time.

This leaves some very important questions unanswered:

  • What do the savings and total spend look like if the customer purchases half the recommended commit? No idea–the console can’t tell you. You’ll save “some money,” but how much is anyone’s guess.
  • What if the best bootstrap period was a specific seven day window last month? Sucks for you, because the console can’t handle arbitrary date ranges.

We thought those limitations were silly, so we completely reverse-engineered Savings Plans and built a proper calculator. Our calculator handles arbitrary data ranges, offers a full visualization, and features sliders to help customers do scenario modeling and find the commit they’re most comfortable with. And in case you’d rather just have the answer without any fiddling, we also calculate the best hourly commit based on your usage and appetite for spend. With this in front of you, suddenly answering “well what if we bought $X instead” takes seconds to answer instead of “let me get back to you after I rerun a bunch of reports.”

RI Migrator

As we wrote when the announcement came out, Savings Plans are a total game changer for AWS customers, but the migration story from Reserved Instances to Savings Plans leaves a lot to be desired. The official (and horrible) recommendation from AWS is to let your RIs expire, wait a week for the Savings Plan Recommendation to populate, and then buy the recommended commit.

That’s pretty silly for a number of reasons. First off, that’s seven days of running compute at full on-demand price–for some of our customers, that could be hundreds of thousands of dollars in wasted money. Second, the necessary information to recommend a comparable Savings Plan replacement is already available to AWS, which just makes this laziness on their part.

Our RI Migrator solves this by looking at all of the RIs set to expire in the near future and calculates the exact Savings Plan hourly commit needed to replace it, allowing customers to purchase the correct Savings Plan the same day the RI expires. That way you’re not “double paying” and wasting discounts, but you’re also not paying on-demand prices. Take a look:

Want to learn more?

We’re currently seeking beta customers to provide feedback on the existing DuckTools Apps and help us design the future ones. If that’s you, just enter your information on the form below and we’ll reach out with more details.

  • We send out occasional email updates on DuckTools. Would you like to be on the list?