AWS Spot Instances: An Overview

Adhi Kesavalu
4 min readFeb 20, 2022



Amazon Web Services provides a number of buying alternatives that allow clients to maximize their cost expenditures based on real-time or dynamically changing user requirements. Customers can buy capacity, host infrastructure, and instances on demand for short-term use, subscribe to recurrent requests, or reserve instances for the long term. The price of each choice varies depending on the capacity and consumption statistics, the timetable, and the AWS services and solutions used.

AWS effectively has a pricing model for every sort of business and user demand. To optimize profits, the corporation is constantly extending its infrastructure and pursuing a broader market. This implies that at any one time, a variety of AWS infrastructure resources remain idle.

Failure to sell the available resource capacity results in a missed business opportunity and possible losses for the cloud provider. To entice additional customers, AWS provides the Spot Instances purchasing structure, which lets users take advantage of underutilized AWS capacity at a significantly reduced price.

What is AWS Spot Instance Billing and How Does It Work?

The AWS Spot Instance price structure provides a 90% savings on the On-Demand instance service. The service is identical to the standard EC2 instance offering, with the exception that it may be discontinued if the cost exceeds the amount requested by the client or if capacity becomes unavailable. The firm determines a price for Spot Instances based on supply and demand patterns and adjusts it based on instance availability.

For the Spot Instance, the customer places their highest offer and continues to get the market spot price as long as it does not exceed the customer’s bid. The market spot pricing is calculated at the start of the instance-hour and invoiced to the client on a per-second basis. As a consequence, pricing is often steady for at least one instance-hour, changing only when the demand-supply imbalance justifies the discounted price. Although AWS states that the average interruption frequency is less than 5%, the service is interrupted differently depending on the capacity availability of the particular instance type and Availability Zone. Consider the case where an instance is invoiced at 1 USD per hour under the standard On-Demand pricing structure.

AWS sets the rate for the identical instance at 0.3 USD per hour using the Spot Instance pricing structure, assuming low demand and strong supply availability. As part of the Spot Instance price structure, a client interested in purchasing the instance bids 0.6 USD per hour.

  • Diversify Capacity Pools: For each AWS Region and Availability Zone, AWS provides infrastructure capacity pools comprising instances of certain categories. Customers that use Spot Instance can customize their app’s resource allocation method by selecting launch requirements. To mitigate the effect of app outages, select various capacity pools from distinct Regions and Availability Zones. To fulfill particular Spot Instance bid criteria while accounting for instance availability in the Spot market, the AWS system may rebalance and distribute workload across capacity pools. AWS Spot Instances may be strategically deployed in conjunction with On-Demand and Reserved Instances to save costs while meeting performance requirements. Follow the steps in this AWS tutorial to create the best allocation plan for your Spot Instances.
  • Improve the system’s price awareness: Create price-aware systems that account for the distribution of infrastructure resources. Using the Spot Instance service with suitable storage and instance kinds, measuring and analyzing use, and designing a system to cope with disruptions can help you maximize cost savings.
  • Analyze Historical Price: AWS offers historical pricing information for capacity pools for the previous 90 days. This data may be used to assess price sensitivity and make better financial decisions. Because future spot market pricing is likewise controlled by previous supply-demand trends, older generations of occurrences with fewer price fluctuations and interruptions can be predicted to continue in the same way.
  • Choose Fewer, Larger Instances: While there is no definite solution to the several tiny instances vs. fewer large instances argument, the situation for AWS Spot Instance pricing may differ. In order to meet escalating demand in the spot market, the baseline price of instance classes is given based on demand, and the relative operating cost of smaller instances may be greater. In addition, historical evidence shows that lesser demand on larger occasions tends to lower the current market price. The pricing structure for the On-Demand service tends to fluctuate linearly across instance types, sizes, and categories. This isn’t always the case with AWS Spot Instances, and more research may be necessary to determine the optimum price circumstances for various instance alternatives.
  • Use Cases: Spot Instances are not appropriate for all workloads. In fact, AWS customers typically use the service for two types of workloads: time-insensitive workloads that aren’t bound by SLA requirements, and peak-time time-sensitive workloads that require additional supplement capacity to accommodate demand spikes for a predictable time period using On-Demand instances.

Hope you understand the concepts of EC2 Spot Instance.

Thank You!