Instance Purchasing Model in EC2

Abimuktheeswaran Chidambaram
3 min readNov 14, 2023

Instance Purchasing Models:

1.1 On-Demand Instance:

pricing is pay per second that should be in running state with a minimum of 60 seconds. No advance payment (upfront cost) is needed. It is suitable for short-term commitment, unpredictable workloads, and irregular workloads. you have full control access over the instance life cycle, so you can decide when to start,stop,restart, hibernate, and terminate the instance. you can change the capacity of instances as per you needs like storage, compute, RAM, and network etc. It is well suited for developing/testing the applications.

1.2 Spot Instance:

It provides less cost than on-demand instances. Used to clean up the unused EC2 instances. pricing is paid by hours. you can save the cost up to 90% discount. It is well suited for data analysis, batch jobs, and background processing jobs. A Spot Fleet is a collection of Spot Instances and optionally On-Demand Instances. A Spot Capacity pool is a set of unused EC2 instances with the same instance type, operating system, Availability Zone, and network platform.

1.3 Reserved Instance:

It provides less cost than on-demand instances. It is suitable for long-term commitment (or) predictable workloads. It does not renew automatically, if it expires means you can continue to use you will be charged on-demand rates. once you purchase a reserved instance, you can’t cancel your purchase but you may modify, exchange, or sell your instance.

payment options:

All upfront — full payment must be paid initially at the start term and receive a large discount.

partial upfront — a partial amount must be paid initially at the start term and receive a medium discount.

no upfront — no payment must be paid initially at the start term and receive a low discount.

classes in Reserved Instances:

Standard RI: It provides more discount than convertible RI. you can be sold and modify the instance. But cannot exchange the instance.

Convertible RI: It provides a less significant discount than standard RI. you cannot be sold. But you can modify and exchange the instance.

Scheduled RI: you can schedule your instance on a scheduled basis (daily, weekly, monthly) with a specified time and duration for 1year.

1.4 Dedicated Instance

Dedicated Instances are dedicated to a single customer. It runs on a shared physical server. But you are isolated (separated) from other users. You do not have visibility into the hardware of the underlying host. Dedicated instances are cost-effective, scalable, and flexible, making them suitable for a wide range of workloads. You can re-size the resources like CPU, memory, storage, etc.

1.5 Dedicated Host

Dedicated Hosts are dedicated to a single customer. It runs on a physical server. You have visibility into the hardware of the underlying host. Choose dedicated hosts when you need performance, security, and compliance, they offer complete control over resources and hardware. dedicated hosts are more expensive than dedicated instances. You cannot re-size the resources like CPU, memory, storage, etc.

1.6 Capacity Reservations

Capacity Reservations allow you to reserve compute capacity for Amazon EC2 instances in a specific Availability Zone. There is a data transfer charge when copying AMI from one region to another. Capacity Reservations can be created in placement groups. Capacity Reservations can’t be used with Dedicated Hosts. Capacity Reservation usage metrics can be monitored in Amazon Cloudwatch.

Thanks for Reading!

Last updated: 07-Jan-2024

--

--