AWS constantly evolves EC2 instance types to meet market demands. Improved hardware, additional cloud services...
and application needs always change. Cloud admins must upgrade instances over time to stay on top of AWS kaizen improvements and keep overall cloud costs down.
Amazon Web Services (AWS) currently boasts more than 30 Elastic Compute Cloud (EC2) instance types spread across more than 30 operating systems. And AWS constantly rotates new processor architectures in and out -- swapping in more efficient storage, adding memory and improving network speeds into existing EC2 instances across 10 infrastructure regions spanning seven countries.
It's important to check EC2 specifications. At any point, you can upgrade your instance type to gain performance improvements and possibly lower costs. A simple reboot of your existing instance could also give you better underlying hardware with the same instance type.
To change the underlying hardware, stop and restart the EC2 instance. Be sure you have Elastic Block Storage volumes and a new internal IP. Many companies purposely interrogate provisioned instances to identify CPU specs and then ditch ones that aren't current. Other companies discover a new generation of EC2 instances can replace existing instances at a lower cost.
To identify the underlying hardware of your instance, type
cat /proc/cpuinfo in the prompt.
If you were running an older generation c1.medium instance type, you would discover the c3.large type is similar in price. However, the c3.large instance delivers 40% faster CPU speeds with more than double the memory. In addition to the right-sizing procedure, cloud admins perform when initially provisioning EC2 instances. It is also equally important to architect your infrastructure to frequently recycle instances to benefit from EC2 improvements.
Swapping out EC2 instance types
There are additional drivers that will influence your decision to switch to a different instance. If you aren't auto scaling EC2s, individual instances may fail because of an overloaded CPU, memory issues or a misconfiguration. Use CloudWatch to monitor your EC2 instance and identify triggers that will require a new instance type.
Know the demands of your application, the costs of the instances you need to support cloud services, as well as the current state of hardware available within the EC2 family. A nimble architecture that allows you to quickly refresh EC2s to take advantage of the latest hardware without disrupting your services demonstrates the beauty of infrastructure as a service.
About the author:
Russ Vanderpool, MSCS and MBA, is a technologist interested in using cloud technology to deliver solutions, help companies better serve customers and to identify new businesses. He has hands-on experience as an architect/developer and a business adviser across the finance, energy, education, technology and nonprofit sectors. Russ has architected and built a cloud infrastructure for a green tech company and while working for Japan's largest system's integration firm, he developed proprietary object-oriented database visualization software for that market.