Featured Member
What are the main factors that affect EC2 pricing?
The main factors affecting EC2 pricing are;
- Compute Type & OS: The server size (vCPU/RAM) and operating system (Windows includes license fees, making it pricier than Linux) [amazon.com].
- Purchase Model: Your choice of plan—On-Demand (full price, highly flexible), Reserved (discounted for long contracts), or Spot (steep discount for excess capacity) [amazon.com].
- EBS Storage: Persistent virtual disk space provisioned for the server, which bills a flat monthly fee even when shut down.
- Data Transfer: Network fees charged exclusively for outbound traffic exiting the server to the internet or other AWS regions.
Why are On-Demand Instances suitable for short-term workloads?
- On-Demand instances require no long-term commitment or upfront costs, allowing you to pay strictly for the exact seconds your server runs.
- You can instantly spin up or terminate capacity on demand, making them perfectly adaptable to temporary or unpredictable traffic spikes.
- This model eliminates financial waste and penalties for short-lived tasks like development, testing, or one-off computational experiments.
What is the difference between Standard and Convertible Reserved Instances?
The main difference between the two lies in discount depth versus infrastructure flexibility.
- Standard RIs offer the highest savings (up to 72%) but cannot change instance families or operating systems once purchased.
- Convertible RIs offer lower savings (up to 54%) but let you exchange your instances for different types or OS options as your infrastructure evolves.
- Standard RIs can be sold on the AWS Marketplace if no longer needed, whereas Convertible RIs cannot be sold.
What are the payment options for Reserved Instances?
There are three payment options for Reserved Instances which are,
1. All Upfront
2. Partial Upfront
3. No Upfront
What is Capacity Reservation?
Capacity Reservation is an AWS feature that lets you reserve physical EC2 compute capacity in a specific Availability Zone for any duration.
Unlike Reserved Instances, its primary purpose is a hardware guarantee rather than a financial discount. It ensures that you will never receive an "Insufficient Instance Capacity" error when launching critical workloads, regardless of market demand.
You pay the standard On-Demand rate for the reserved slot whether your instance is actively running in it or sitting idle.