Elastic Compute Cloud was created for the sole purpose of making web cloud computing services scalability a reality for many enterprises and individuals. Its auto-scaling capability allows you to easily maintain multiple instances that share similar characteristics based on the set criteria.
EC2 Auto Scaling Groups
Auto-scaling groups are a collection of Elastic Compute Cloud instances. These instances share close similarity hence treated as a logical unit. Maintenance and scaling of the logical grouping are done as a whole.
An autoscaling group is initiated by first setting the number of elastic compute cloud instances required to fit its capacity. The auto-scaling group maintains the set instances by performing periodic performance diagnostics. If an instance is malfunctioning or unhealthy, the group automatically terminates it. The instance is then replaced with another case.
Auto-scaling groups have scaling policies that guard the number of instances included in the group at the set capacity. When in effect the scaling systems automatically adjust the number of cases for optimal use and reduced costs. This may mean rendering certain dormant or malfunctioned instances off as well as including additional situation if the capacity is increased.
You can create an auto-scaling group using the API or CLI.
What's an EC2 Instance?
An instance is a virtual server used to run applications on the AWS services infrastructure. Its created from (AMI) Amazon Machine images. Machine images act like templates of configured software and operating systems used to determine your operating environment.
Amazon Web Services provides businesses with different types of instances and a variety of configurations such as CPU, memory, networking resources and storage to suit their workload needs.
Instances types are categorised into groups based on the target application profile. They include compute-optimised, memory-optimised, GPU instance, accelerated computing, storage-optimised, general purposes, and micro-instances.
AWS services Instance Purchase Options
Amazon Elastic Computing Cloud supports a variety of instance purchase options for customised costs. They include:
- default on-demand instances
- reserved instances
- scheduled instances
- dedicated hosts
- dedicated instances
- capacity reservations
- spot instances
On-demand instance requires you to pay per second exactly when you launch. Reserved instance purchase option allows you to outsource available cases at a significant discount for a period not less than one to three years. Capacity reservations, on the other hand, will enable you to secure capacity for your instances for a set duration of time at any specific availability zone.
Schedule instance allows you to purchase Amazon web services on the specified recurring schedule. They are usually available for a one year term. Spot instance purchase option grant you access to unused Elastic Compute Cloud instances which lower your charges significantly.
The dedicated host option allows you to pay for a physical host that is entirely devoted to the running of your instances. Dedicated hosts lower the cost of your per-socket, per-VM software, and per-core cost licenses. As for specialised cases, you are charged hourly for every instance that runs separately on single-tenant hardware.
Maybe you'll find this white paper intersting: