April 27, 2022 · 3 min read
When it comes to Amazon Web Services (AWS), one of the most important things to understand is reserved instances. In short, a reserved instance is a type of AWS instance that you typically pay for upfront, and in exchange, you get a discount on the hourly usage price. While the concept might sound simple enough, there are a few key facets to understand about reserved instances, including how they work on both a technical and economic level. There are also no upfront RIs, which Usage.AI uses, that bill hourly, but still come with a contractual obligation.
On a technical level, reserved instances give you access to a certain amount of compute resources (e.g. CPU or memory) for a set period of time. This means that you can count on having those resources available when you need them, without having to worry about other users consuming them. In addition, reserved instances can be used in conjunction with autoscaling to ensure that you have the right amount of resources available at all times.
Cloud providers like AWS are naturally incentivized to keep prices low in order to attract and retain customers. However, they also need to make a profit in order to stay in business. This is where reserved instances come in. By offering discounts on hourly usage rates, AWS is able to generate revenue from customers who are willing to pay upfront or otherwise commit to their compute resources.
On an economic level, reserved instances can be seen as a way to hedge against future price increases. If you anticipate needing a certain amount of compute resources in the future, it may make sense to purchase a reserved instance now at the current price. This way, you can lock in that price for the duration of the reserved instance, and avoid any potential price hikes down the road.
In addition, reserved instances can help prevent churn by giving customers a financial incentive to stick with AWS. If a customer knows they have already paid for a certain amount of compute resources, they are less likely to switch to another cloud provider that might offer a lower hourly rate. This helps AWS keep its customer base stable, which is important for long-term growth.
Cost is a key consideration when it comes to cloud computing, and reserved instances can provide significant savings compared to on-demand pricing. In general, reserved instances offer a discount of up to around 70% compared to on-demand pricing, making them an attractive option for organizations looking to reduce their AWS costs.
If you're not sure which type of reserved instance is right for your needs, Amazon offers a reserved instance pricing sheet that displays the pricing and exact savings of any given reserved instance.
Overall, reserved instances can be a great way to save money on your AWS costs. Be sure to consider all your options before committing to a particular type of reservation, and check the pricing of reserved instances to ensure you're getting the best possible deal.
There are two main types of EC2 instances: on-demand and reserved. On-demand instances are exactly what they sound like - you can launch them when you need them and pay by the hour. Reserved instances are cheaper, but you have to commit to using them for a one or three year term.
The biggest difference between the two is price. On-demand instances cost more per hour, but you don't have to commit to anything. Reserved instances are cheaper per hour, but you have to commit to paying for them for at least a year.
Further, reserved instances give you the option to use capacity reservations. These guarantee that you'll always have enough capacity available, even during a heavy traffic period.
There are two types of RIs: Standard Reserved Instances and Convertible Reserved Instances.
Standard RIs offer the biggest discount (up to 72% off On-Demand), and are best used for when you need a consistent amount of resources. Convertible RIs provide a discount (up to 54% off On-Demand), and also let you change the attributes of the RI as long as the new Reserved Instance is of equal or greater value.
So, which type of instance is right for you? It depends on your needs and budget. If you need the flexibility of on-demand instances and can afford the higher price, they're a good choice. If you're willing to commit to using an instance for at least a year and want to save money, reserved instances are a better choice
Amazon pioneered the use of reserved instances in 2006. In a blog post discussing the new feature, AWS evangelist Jeff Barr wrote: "some of our customers have needs which aren’t addressed by the spot pricing mode."
The first of these needs was lower pricing for customers who could commit to using a certain amount of compute capacity over a longer period of time. Another need was to have a certain number of instances available at any time in a DR (Disaster Recovery) scenario.
In order to meet these needs, Amazon created reserved instances. With a reserved instance, you pay for a certain amount of capacity upfront, or commit to long-term hourly billing, and in return, you get a lower price for that capacity. You also get the guarantee that your instances will be available when you need them.
While reserved instances were originally designed for use with EC2, they are now available for several other AWS services, including RDS, ElastiCache, and Redshift. Reserved instances can also be used with third-party software running on AWS, such as MongoDB and Cassandra.
Google Cloud also offers a form of reserved instances with their committed use program. With committed use, customers commit to paying for a certain amount of capacity over a set period of time. In return, they receive a discount on the price of that capacity.
China's leading cloud service provider, Alibaba Cloud, offers its own version of reserved instances. Similar to AWS and Azure RIs, AliCloud RIs allow customers to commit to using a certain amount of compute capacity over a 1 or 3-year period. These are up to 79% cheaper than pay-as-you-go instances.
Further, IBM Cloud, which is a top choice among Fortune 500 companies, offers reserved instances called IBM Cloud Virtual Servers reserved instances. These RIs offer guaranteed capacity, global availability, reliable provisioning, and cost savings.
As cloud computing continues to grow in popularity, it's likely that we'll see more providers offering their own versions of reserved instances. For customers, this means more choice and more options for finding the right RI provider for their needs.
With Usage.AI, customers can take advantage of the cost savings of RIs without the commitment, saving up to 57% in EC2 costs. To do this, Usage.AI buys and sells RIs in the RI marketplace on the customer’s behalf. Customers get the advantage of discounted reserved instance pricing, but they don't have to worry about the commitment or upfront cost of RIs.