Saving Cost Intelligently – Unused AWS Reserved Instance Modifications

AWS Reserved Instances(RI) can significantly lower your computing costs by up to 75% and are very popular choice for many customers to reduce their total cost of ownership(TCO) in the cloud. However purchasing RI happens during business budgeting cycles in the enterprises and it also depends on the workload/new applications adoption along with current consumption changes.

One of the cool features of AWS RI is modification feature,  an ability for customers to modify their existing Reserved Instances from earlier active purchases to help in capacity re-allocation and easily scale-up/scale-down the EC2 instance sizes if their workload patterns change.

However, one of the bigger challenge with respect to RI planning in enterprises and heavy AWS users is the variation of workload patterns, changes in deployment architectures and new apps adoption etc that make it difficult to effectively monitor and manage the existing RI purchases for maximising ROI. This requires you to constantly keep track of your RI utilisation, unused RI and required RI.

Let’s say, your company may have purchased Reserved Instances for 1 or 3 years but that does not mean, there won’t be any modifications required as your workload and computing consumption varies. You may realize that there are a few unused Reserved Instances that cost you unnecessarily even if you are not using them anymore and simply by modifying them for re-allocation will keep your costs under control. This happen also because of Reserve Instance (RI) utilization consumption change over a period of time for most customers, an easier way to manage AWS Reserved Instances modifications is absolutely essential to save cost and required human effort in analysis.

Today, we are happy to announce an intelligent recommendation module to Botmetric RI Planner to help you with making smart decisions by using our – Unused AWS Reserved Instance Analyzer. Many of our beta customers love this recommendation engine and launching it in GA for everyone. The Botmetric EC2 Reserved Instance Planner now suggests many insightful recommendations to help you with unused RI modifications and save you a significant amount of time in manual analysis.

Botmetric Unused AWS RI Analyzer detects the existing reservations for which there are no corresponding EC2 instances mapped in the running state. By using the meta data of EC2 instances, RI purchases and hourly billing information from AWS APIs, it analyzes running instances in AWS cloud account to identify unused reservations and find the potential modifications so you can save money and quickly make a decision for re-allocation.

We help you detect unused RI and potential savings by evaluating:

  • If a reservation is not being used, that means you are paying extra money (or already paid in case of All Upfront) for a resource you are not using any more
  • You have instances running, for which you are paying the On-Demand price so you are spending twice

With this new feature, get recommendations for RI modification to quickly realise cost savings:

  • You are using the money that was already spent on the EC2 RI purchase
  • Your running instances would now be counted as a Reserved Instance and you will be paying less

In order to leverage this, just go to EC2 RI Planner > Unused RI and click on “Show Recommendation” as outlined in the below image.


Botmetric’s Unused AWS Reserved Instance Modification Analyzer

Following modifications can be performed on AWS Reserved Instance from Botmetric:

  1. Switch Availability Zones within the same Region

AWS allows to modify instance reservations within the same region to map across different availability zones.

Example: We have a m3.medium machine present in us-east-1a which is reserved but unused  as we are running any m3.medium machine in that AZ. However, we have one m3.medium machine that is running in on-demand mode in us-east-1b which costs more due to on-demand hourly billing.

Botmetric’s RI planner will provide a recommendation to switch the availability zones for this RI purchase by showing potential savings with this modification. There could be multiple modifications possible for single unused AWS Reserved Instance depending on your existing workloads. Botmetric will show the savings in a sorted order.

  1. Change between EC2-VPC and EC2-Classic

EC2 classic was a useful at early stages of cloud evaluation, but as you adopt cloud,  businesses need an isolated virtual private cloud and customised network controls so  EC2-VPC is an essential component.

However RI purchases can be for EC2-Classic or EC2-VPC. We help you smartly adopt the unused RIs in Classic vs VPC. The reservation discounted rates for instances can be applied to either platform, but the capacity reservation will be limited to the platform for which it was purchased for.

Note : Modifications are only applicable on reservations running on Linux OS.


Unused AWS reserved instance modification recommendations

  1. Change the Instance Type within the Same Instance Family

AWS provides a range of instances grouped into instance families. These instance families differ in their compute capacity, network throughput and memory. Any RI reservation can be modified within the same instance family, provided it lies within the same AWS region. With every instance in families there is a ‘normalization factor’ attached. The normalization factor is a number assigned to an instance, used to compare instance’s value relative to each other. The larger the instance size, the higher the normalization factor; the smaller, the lower.

For example: 1 m1.medium machine can be modified into 2 m1.small machines or vice-a-versa. We can combine 2 t2.small machines into 1 t2.medium machine.

AWS Reserved Instance pricing model

Let’s take an example where a customer had purchased the following On-Demand instances:

  • 4 x m1.small instances in Availability Zone us-east-1a
  • 4 x c1.medium instances in Availability Zone us-east-1b
  • 2 x c1.xlarge instances in Availability Zone us-east-1b

The customer then purchases the following Reserved Instances:

  • 2 x m1.small instances in Availability Zone us-east-1b
  • 3 x c1.medium instances in Availability Zone us-east-1a
  • 1 x c1.xlarge instance in Availability Zone us-east-1b

The pricing is applied in the following ways:

  1. The two m1.small reserved instances is applied to two of the four running m1.small instances in Availability Zone us-east-1a by modifying Availability Zone to us-east-1b.
  2. The other two instances in Availability Zone us-east-1a will continue to be charged at the current On- Demand rate.
  3. The three c1.medium Reserved Instances don’t match the Availability Zone of the running c1.medium instances.
  4. The four running c1.medium instances continue to be charged at the current On-Demand rate.
  5. If the customer launches a new c1.medium instance in Availability Zone us-east-1a then the discounted usage fee will be applied to that instance.
  6. The discounted usage fee rate for one c1.xlarge reserved instance is applied to one of the two running c1.xlarge instances in Availability Zone us-east-1b.
  7. The other c1.xlarge instance in Availability Zone us-east-1b will continue to be charged at the current On-Demand rate.
  8. Instances can be modified within normalization factor equality.

For example: A medium machine with normalization factor 2, can be modified with 2 small machine with normalization factor 1 or can be modified with 4 micro machine with normalization factor 0.5. So long as your instance size footprint remains the same, AWS Reserved Instance modifications won’t impact your bill, as billing is proportionate with normalization factor.

Botmetric’s RI planner provides you all possible modifications that can be done for your Unused RI’s along with the savings per month. It also provides the possible reservation IDs for that reserved instance. There might be multiple reservations IDs present. This is because reservation IDs are mutually exclusive and can’t be particularly linked with same instance-id while reserving it. If no possible modification is available for your AWS Reserved Instance then Botmetric’s RI Planner will simply suggest you to sell it via AWS.

However, here are a few key points you should keep in mind:

  1. Switching availability zones and network type modifications are applicable for all product platform types while changing instance type is only available for Linux platform.
  1. Linux Reserved Instances cannot be modified to RedHat, SUSE Linux Reserved Instances or vice versa.
  1. After modification, the pricing benefit of the AWS Reserved Instances is applied only to those instances that match the new parameters. However, instances that no longer match the new parameters are charged at the On-Demand rate unless your account has other applicable reservations.


To simplify this task and intelligently identify RI modification opportunities based on your day to day utilization, we recommend you to try Botmetric’s Unused RI Analyzer.

You can start a free 14-day Botmetric trial here.