AWS’s flexible and low-cost services is one reason it has become a leader in the cloud computing space. But organizations who leap into AWS services like EC2 without fully understanding all of their pricing and discounting options can be surprised to see how quickly costs can add up. To help you choose the best pricing model for your organization and avoid unexpectedly high AWS EC2 costs, we listed some high-level pros and cons of each of the three most popular AWS EC2 pricing options here.
What is AWS EC2?
AWS Elastic Compute Cloud (EC2) is AWS’s main cloud computing platform that provides organizations with secure, scalable computing capacity and enables users to develop and deploy applications without the need to invest in a hardware infrastructure. It is a simple web service interface allowing organizations to “rent” virtual servers to run applications in the cloud.
While AWS EC2 Dedicated Hosts are available, the most popular pricing options are On-demand, Reserved Instances, and Spot Instances.
1. On-demand – The Most Flexible
As the name implies, on-demand is the most flexible, pay-as-you-go pricing option. With on-demand instances, you pay for compute capacity on an hourly or per second basis and only pay for the EC2 instances you actually use. The cost per hour or per second depends on the instance type and the region the instance is located.
In return for this high level of flexibility, you pay more than you would with the other pricing models. That being said, on-demand pricing can be a great option for some organizations. For instance, this might be a good option if you run your EC2 instances in a very limited amount such as a couple of hours, a few days per week. Some organizations may even start with on-demand pricing to get a better idea of what their usage will be before potentially moving to a lower-cost pricing model.
In general, On-demand is a good pricing option for organizations with:
- Users that prefer the flexibility of Amazon EC2 without any up-front payment or long-term commitment
- Applications with short-term, widely fluctuating, or unpredictable workloads that cannot be interrupted
- Applications being developed or tested on Amazon EC2 for the first time
On-demand Discounting with an Enterprise Discount Program (EDP)
If you require a high level of flexibility and believe that on-demand pricing is best for you, we recommend you request information and pricing for an AWS EDP. This private pricing program provides a small discount in exchange for a prepaid, annual consumption commitment. That prepaid amount is then run down based on your consumption. To get an EDP, you do need to purchase enterprise support. The level of discounting you achieve with an EDP will also depend on the term length and annual spend commitment.
For more information on how to achieve better-than-standard price transparency and predictability in your EDP agreement, see our blog: 3 Must-have Commercial Asks When Negotiating an AWS EDP.
Pros of On-demand for AWS EC2:
- No term or annual spend commitment (unless you have an EDP)
- No upfront costs
- Ability to terminate the instance once you no longer need it
Cons of On-demand for EC2:
- High cost compared to other pricing models
- List pricing unless you make a multiyear, prepaid financial commitment (i.e., EDP)
- EDP discounting is significantly lower than the discounting on other pricing models
2. Reserved Instances – Built for Predicted Scalability
If you are able to commit to a 1- to 3-year term, reserved instances can provide a significantly higher discount than on-demand pricing. Under this model, you are charged per hour or per month and the pricing varies based on instance type and region.
Reserved Instances are recommended for organizations that have consumption predictability and are running applications that require reserved capacity for steady state usage.
Discounting for Reserved Instances
The level of discounting you can achieve with Reserved Instances depends on your term length as well as your upfront payment. You can pay no upfront, partial upfront, or all upfront. All upfront leads to the highest discount. There is also volume-based discounting with this pricing model.
If you are considering Reserved instances for EC2, we recommend that you ask for visibility into ALL discounts for various upfront payment options and term commitments.
Pros of Reserved Instances for AWS EC2
- Best for customers that can commit for a 1- to 3-year term
- Significantly higher discounting compared to On-demand
Cons of Reserved Instances for AWS EC2
- Reserved instances are frequently underutilized, thereby negating the discounting to some extent
3. Spot Instances – The Deepest Discount
Again, the price you pay with Spot Instances depends on the instance type and your region but in this case, it also depends on supply and demand for Spot Instance capacity. You are essentially bidding on spare computing capacity and pricing fluctuates every 5 minutes depending on what is currently available. There is no upfront commitment. If there is available capacity at or below the price you select, you’ll be charged for the Spot Instance price that is in effect for the period your instance runs.
With Spot Instances, you have the option to set defined duration periods in increments of 1-hour up to 6-hours. Choosing duration periods in chunks of 6-hours will give you a bigger discount than choosing 1-hour increments but setting a duration period at all will lower your overall discount with Spot Instances. Selecting off-peak periods can provide a lower price but your instance will only run within AWS’ specified weekend hours.
Compared to the other AWS EC2 pricing models, Spot Instances provide the highest discount but offer the least predictability due to pricing volatility and fluctuating availability.
Spot Instances are recommended for organizations:
- Running applications with flexible start and end times
- With urgent computing needs for high amounts of additional compute capacity
Pros of Spot Instances for AWS EC2
- Highest discount compared to on-demand
- No upfront commitment
Cons of Spot Instances for EC2
- Aspects of the SLA in the spot instance market are not guaranteed since availability can fluctuate between instances and users can be unexpectedly forced to switch servers without advanced notice
- No predictable pricing structure since price fluctuates based on supply and demand
Though AWS offers flexible and low-cost pricing options, it is still necessary to identify your appetite for flexible run times and calculate your expected consumption in order to choose the best model for your requirements. It is also critical to ask AWS to provide full transparency into pricing and discounting into all scenarios you are considering (term length, spend commitments, etc.). It is simply not enough to go by the pricing listed on AWS’ website. Pricing, discounting, and commercial terms should all be negotiated and put into your written agreement.