Amazon EC2 – 15 Years of Optimizing and Saving Your IT Costs

Voiced by Amazon Polly

September 8, 2021: Amazon Elasticsearch Service has been renamed to Amazon OpenSearch Service. See details.

I still remember in the early years at AWS, we played around the terms “instances”, “virtual machine”, etc. when referring to Amazon Elastic Compute Cloud (Amazon EC2).  Looking back, since the moment AWS launched its very first Amazon EC2 instance in August 2006, the m1.small, AWS has redefined the way customers acquire and consume compute resources.  With the recent launch of GPU-optimized, EC2 G4 instances, Amazon EC2 has now 400+ instances that can provide you a wide selection of instance types for virtually any needs.  Jeff Barr kept a nice timeline of EC2 instance history in his blog post.

Whether you are looking for a general purpose instance that can balance your compute, memory and networking resources, or compute optimized instances that give you high performance processing and computing, or memory-optimized instances that are designed to process large data sets in memory, you will find the most appropriate mix of resources and instance sizes for your applications.

To help you navigate and discover the right instances for your workloads and business needs based on requirements, e.g. availability zone or region, compute, memory, networking, AWS offers you the “Instance Type Explorer”, a tool you can quickly narrow down the search by instance category, or by hardware configuration.  You can also use “Instance Types” function in your Amazon EC2 console, or access programmatically via APIs: DescribeInstanceTypes and DescribeInstanceTypeOfferings.

Pricing Reduction

Regular price cuts on all our services has been a standard way for AWS to pass on the economic efficiencies gained from our scale back to our customers.  As of April this year, AWS has reduced prices 107 times since it was launched in 2006. You can learn about all the previous price reductions effort, including EC2 instances, from this shortlisted price reduction related AWS blogs.

Purchase Options

Depending on your workload pattern, you can consider paying your EC2 resources as you go at the On-Demand rate, or select a one- or three-year term-based commitment programs, e.g. Reserved Instances or Savings Plans, and receive up to 72% discount.  You can also take advantage of EC2 Spot Instances to leverage unused EC2 capacity at a steep discount of up to 90%.

What is the difference between Reserved Instances and Savings Plans?

Both these two commitment-based purchase options can provide you a discount of up to 72% off On-Demand rates.  You commit to a specific hourly amount of usage for an instance family in a specific region (e.g. 0.5 normalized unit per hour of a t3.micro instance in US East) with Standard Reserved Instances, and a specific amount of hourly spend for an instance family in a specific region (e.g. $25/hour of c5 instance in US West) with EC2 Savings Plans. In spite of the same requirements of the specificity for instance family and region, EC2 Savings Plans are flexible in terms of OS and tenancies.

If you want more flexibility with your Reserved Instances, you can choose Convertible Reserved Instances, which allows you to exchange instances into the ones of different instance family, size, region, tenancy, and OS.  However, if you choose Compute Savings Plans, in addition to these flexibilities, your commitment will automatically be applied to all EC2 usage, regardless of instance family, region, instance size, tenancy, or operating system etc., and can also be applied to your AWS Fargate and AWS Lambda usage.

Besides the utilization versus spend commitment and various flexibility range between Reserved Instances and Savings Plans, the service coverage of these two purchase options is another factor to take into consideration.  Currently, Reserved Instances (launched in 2009) are available for Amazon EC2, Amazon RDS, Amazon ElastiCache, Amazon OpenSearch Service (successor to Amazon Elasticsearch Service), Amazon Redshift, and Amazon DynamoDB; while Savings Plans, as a newer purchase program (launched in 2019), has now three options: Compute, EC2, and Amazon SageMaker.

Cost Optimized Resources

You want to make sure your development and management teams who make the technology design and purchase decisions understand all the resources available to better match your EC2 capacity with demand.  The dynamic and predictive features of Amazon EC2 Auto Scaling helps add / remove EC2 instances based on the changing demand.  AWS Instance Scheduler allows you to configure the start and stop schedules for your EC2 instances.

Based on your historic spend, AWS curates recommendations to optimize your resources.  AWS Cost Explorer Rightsizing Recommendations detects your idle and underutilized instances and sends you recommended instance type(s) and size(s) along with the estimated savings. If your workload patterns fit for any of the commitment-based purchase options, e.g. Reserved Instances or Savings Plans, you will also receive guidance for the recommended instance family, size and commitment in your cost management console. AWS Compute Optimizer offers deep insights into the recommendations by identifying over- and under-provisioned resource dimensions and provides you a way to identify OS configuration changes needed to apply recommendations. Review and customize these recommendations according to your plan to gain the most savings.

Conclusions:

We are confident that as one of the foundational services of AWS, Amazon EC2 will grow stronger, excite you with more innovation, and continuously exemplify all the essential values of cloud computing:

• Scalability, allowing you to spin up and down compute capacity within minutes
• Security, enhanced by the Nitro System, supporting 80 and more security standards and compliance certifications
• Reliability, committing 99.99% availability for each Amazon EC2 region
• Cost-optimized, offering you the flexibility and tools to choose the right purchase and resource options that meet your workload needs
• Innovation, 400+ instances rolled out in the past 15 years to continuously optimize for compute, memory, storage.

We invite you to join our virtual celebration (Aug.23-24) of the 15 years birthday of Amazon EC2 and learn from AWS leaders and experts as they share lessons from building the world’s most comprehensive and broadly adopted cloud platform.

Also, be sure to check out Jeff’s News blog channel for his birthday wishes to EC2 on Aug.23.  As our Chief Evangelist and a tenured Amazonian (19 years and counting), he will for sure have more interesting stories and milestones to share regarding the EC2 launch and its 15-year growth journey.