How to Simplify AWS RI Complexity to Maximize Savings

The simplest part of using AWS Reserved Instances (RIs) is clicking the button to purchase them. The complex part is just about everything else. Understanding how RI planning works (even as AWS continues to change) and how the right RI planning tool can help.
The simplest part of using AWS Reserved Instances (RIs) is clicking the button to purchase them. The complex part can be just about everything else. Get around that complexity by understanding how RI planning works (even as AWS continues to change) and how the right RI planning tool can help.

RIs are a great way to be cloud cost efficient, but they require some in-depth knowledge and a thorough plan for the continuous monitoring and optimization that follows purchasing.

Whether you’re a massive enterprise or an agile startup, having the right data engine-driven RI planning tool will help you demystify RIs while maximizing their potential savings. For instance, the Cloudability platform does much of the heavy lifting for you, while staying caught up with the latest AWS changes like Instance Size Flexibility (ISF) or new options.

But you still need to understand the operational best practices before heading down the path toward RI success. So, let’s begin!

A Brief History on RIs

There’s no doubt that AWS will continue to add new features and options to their RIs, giving users more flexibility to save on compute and database hours. But with these new options come layers of complexity. Here are some major RI changes released by AWS to offer a scope of how things have changed since 2014:

  • December 2014: New Standard Classification, adding all-upfront, partial up-front and no-upfront RI options into the mix
  • January 2016: AWS launched Scheduled Reserved Instances
  • September 2016: Introducing Convertible RIs and Regional Scoping
  • March 2017: AWS added Instance Size Flexibility
  • May 2017: AWS announced 3-Yr, No-Upfront Standard RI option
  • …and likely more to follow this year!

With each change comes more options for end users to save on cloud costs. Having a means of visualizing and managing RI purchases and utilization in order to make the best decisions for their infrastructure is key to navigating the complexities. We strongly advocate that organizations should not take on RI planning alone, but instead should have a cost optimization tool to assist in planning, purchasing and modification.

Demystifying the “Coupon Model”

Now that we have a sense of the long lineage of RI changes, let’s recap what they are: RIs are best described as hourly coupons that can be applied each hour of the month for the duration of the RI term. One essential piece of RI knowledge is that these coupons are highly specific: they require you to reserve hours within a specific AWS Region or Availability Zone (AZ) (e.g., US-West-1).

You need to reserve every single hour within either a one-year or three-year term and those reservations are only applicable to the specific instance combination you purchase against (e.g., one year of discounted t2.micro hours for Windows servers in us-west-1).

cldy emerge 2017 06 maze chart 03 2821 - How to Simplify AWS RI Complexity to Maximize Savings - Apptio
Figure 1: Every letter block represents an individual instance and the hour it ran. For example, you can see that instance A and B ran for every hour. All the other instances ran for only a part of this time period, so their coverage varies depending on the reservations available.

For the figure above, also note that all of these instances are the same. Let’s say that this ten hour block is repeated over and over again for one year. We can see that if we purchased four RIs, they would each have 100% utilization; a fifth RI would have 90% utilization, a sixth RI 80%, and so on. It’s unlikely that you’d purchase a seventh RI, as it would only have 50% utilization.

The exception here is if you purchase Regionally-Scoped Linux RIs, an announcement from AWS enables these RIs to apply across any size of instance within a family. This latest update greatly simplified the process of planning for RIs, and also makes the whole exercise a lot more flexible than before. That is, assuming you have a strategy and tool that does the number crunching in order to make this process practical!

Another Reminder: Your RIs Are a Commitment to the Full Term

There are lots of savings to achieve with RIs, but only if your team can keep RI waste as low as possible. Regardless of the RI type you purchase, know that you are committing to every hour of the entire term, whether that be one or three years. If you have an hour that goes by without a matching instance to your reservation you are still going to pay for that hour, and if there’s no instance using up that hour, that hour goes to waste.

Get all of the essential RI planning best practices with our free comprehensive e-book.

Maximize Savings by Constantly Validating, Monitoring and Optimizing RI Usage

Knowing that RIs are simply a coupon construct that can apply to any matching instance, it’s imperative to build and validate your purchases against a histogram which covers a representative time period for your organization. Yes, you could try and handle this manually with a fancy spreadsheet and likely way too many hours of work – or, you can leave most of the analysis to a cost optimization tool like Cloudability.

cldy emerge 2016 03 andrew histogram 2702 - How to Simplify AWS RI Complexity to Maximize Savings - Apptio
Figure 2: Visualizing the water line and usage gaps with Cloudability.

Any hour of unused reservation (see the divots in the image above) effectively bumps up the per hour low cost you acquired when you purchased the RI and works to negate any savings you calculated. A cloud optimization tool like Cloudability distills these usage histograms for you, and you can easily set your own “water line value” (a visualization of ideal RI coverage popular with enterprises like Atlassian) to create an RI portfolio which suits your specific needs.

Use the Right Tool for the Job

We highly recommend keeping an eye on the waterline and making sure that you have access to the histogram of your RI investment. Any RI planning tool needs to be able to validate recommendations against operational data and offer a sense of forecasting throughout the RI term. Make sure your RI planning tool makes the most of the anatomy and nuances of RIs. Otherwise, you’ll be prone to generating waste or missing clear savings opportunities.

Pro-tip: Cloudability is constantly vigilant about RI changes made by AWS and adapts to them quickly. This ensures that our customers don’t miss a beat when it comes to tracking their reservations and maximizing their utilization. For a sneak peek at what that looks like in action, we’re always welcoming teams to give our tool a try.

The Bottom Line

There are two pieces of advice we are always happy to give when it comes to RIs:

  • Always adapt to the latest advancements from AWS which make purchasing more practical and cost efficient, such as ISF.
  • Always rely on a histogram and waterline strategy to constantly drive the validation of the financial investment you made, as well as to keep an eye on the performance of the RI fleet.

Once again, RIs can be simple to buy, but it is challenging to attain every bit of their potential value and savings. It takes the right tools and know how to plan, which is worth the effort as that leads to a lower cloud bill at the end of the month. Not to mention, you’ll spend less time worrying about whether you’ll get a return on those one to three-year commitments!

If you need help, we’re always ready to show you the data-driven way toward saving with RIs.

目次

カテゴリ

タグ