Cloudability logo
Listen to our AWS re:Invent 2018 Reaction
Optimization

Ask the Experts: Planning for Reserved Instances

By Huy Tran on September 21, 2016
Strategic AWS Reserved Instance planning allows users to capture the most savings possible with EC2, Redshift, ElastiCache and RDS. We answer RI planning questions from AWS users and business leaders on how to improve cloud management and efficiency.

Strategic AWS Reserved Instance planning allows users to capture the most savings possible with EC2, Redshift, ElastiCache, RDS, and whatever new instance-type services AWS comes up with down the road. We answer a few questions from AWS users and business leaders about planning for Reserved Instances.

Check out our answers below. If you have your own burning questions about cloud cost management, feel free to submit them with the form in the blog’s sidebar, and we’ll get back to you with an answer as soon as possible!

Question 1: Is tagging important to my AWS Reserved Instances planning?

Properly tagging all AWS resources is a foundational part of understanding where costs and usage stem from. It also aids in allocating costs back to the right departments, teams, and projects. Each AWS resource can contain up to 50 tags per resource to add detail.

screen-shot-2016-03-22-at-1-13-38-pm

Here’s an example of a tagging schema at work.

ebook_robot-2-01Tagging allows users to associate costs of services, like EC2 for example, back to key business dimensions. From a line of business perspective, users can know how much a particular project or department spends on EC2.

Tags help cloud cost management tools, like Cloudability, break down EC2 costs to improve planning capabilities for DevOps and IT. It helps teams report on the savings of well-provisioned RIs over time, too. The ability to prove to finance that an RI investment has reached its “break even point” (learn more about them here) and is now providing cost savings is always a welcome conversation.

Question 2: What kind of savings do users see between the 1 and 3-year RI commitments?

The three-year-term RIs have the best savings, especially when paid for all-upfront. But it requires the operational confidence to commit to three years (technically, every hour of either term) of provisioning the RI rate for a given instance. The one-year RI term has less savings, but there’s a relatively shorter commitment.

If you’re wondering about various upfront payments, here’s a breakdown from our Definitive Guide to Reserved Instances:

Let’s take a closer look at this using EC2 as an example once more. So, a hypothetical business wants to save on their m4.xlarge costs. Here are the options for RIs:

screen-shot-2016-09-21-at-1-26-24-pm

The three-year term might seem like the best choice for the most savings, but what if the operational history of this fictitious business shows an annual growth of service usage? This means there’s a chance to outgrow this instance as the product and service grows in use and cloud compute requirement.

They might be using an m4.xlarge today, but their compute requirements might force them to upgrade to an m4.4xlarge, or even a different family altogether. Once the proper instance is found, a new RI will need to be purchased to capture the most savings again.

ri_mods

Here are a few examples of RI modifications from the Cloudability app.

However, if engineers decide to downsize instances, Reserved Instances can be modified into smaller sizes to provide savings. These instances must be in the same family, though (e.g. downsizing m4.8xlarge into two m4.4xlarges, etc.).

The bottom line

Users can make the most of the three-year term if they know, via the right cost and usage data, that the instances won’t change over the course of that time. If users who monitor their cost and usage data can forecast growth or decline, working with a shorter term RI might be the best way to go.

Keep in mind though that RIs can be modified to adjust to the changing nature of cloud utilization and the instance adjustments that need to be made to maintain a high level of efficiency. For RIs that can’t be modified to fit new infrastructure changes, they can be sold on the AWS marketplace.

If you have reservations that aren’t being used, or can’t be modified into RIs that fit your instances, there are a couple of options available. Socializing existing, but unused, reservations with your organization can surface other teams or projects that might be using the same instances but with on-demand rates. Letting those teams use your “old” RIs lets them save during the course of their projects.

You can also recoup some—though not all—of the initial purchasing fee by selling them in the Amazon EC2 Reserved Instance Marketplace. There’s no guarantee that your reservations will be bought if you put them up for sale, but if you have a significant number of reservations that won’t save you any money if kept, giving it a shot can be your most cost-effective option.

For more on choosing the right EC2 instances, or more about Reserved Instances (pricing and modification), check out the respective eBooks.

Question 3: What happens to my Reserved Instances if my engineers decide to switch to newer, better AWS products?

It’s no surprise that our talented operations and engineering folks for wanting to use the latest and greatest from AWS. Also, we often get the best rates of service from the newest generation of AWS products.

Similar to the answer above, RIs that no longer apply to a given infrastructure can be sold on the AWS marketplace. This is common with larger enterprises resourcing hundreds to thousands of various instances and RIs. It’s a great way to recuperate costs from RIs to keep a lean, strategic Reserved Instance portfolio.

breakeven-point

In this fictitious example, the RI purchase “breaks even” and savings on the instance are realized around the six-month mark.

This doesn’t mean folks should switch their RIs on a whim. RIs only deliver their maximum benefit if organizations use them beyond their “break even” points. This is the amount of time where the savings of the RI are higher than their initial costs. Using cloud cost management tools, like Cloudability, to monitor costs and usage helps teams identify these break even points and plan to achieve maximum savings.

So, before engineers spin up the next cutting-edge AWS instances, there should be some up-front planning to ensure they properly provision and stick with instances to get the most RI savings possible. A cloud cost management tool assists with this planning and helps get the right cloud cost data to the right decision makers.

See this kind of RI planning at work

Seeing actual cost and usage data through a cloud cost management tool, like Cloudability, can provide answers to all kinds of additional RI planning-related questions. We invite everyone to check out a Free Trial of Cloudability. We also have Reserved Instance Optimization experts waiting to help.

Being in the know feels great